Based on the ranging principle of satellite laser ranging (SLR) system, the distribution of echo photon reflected back from the angle reflector is simulated numerically, and the relationship between the shape effect of the satellite angle reflector and the ranging precision of SLR system is analyzed and discussed. In addition, in order to verify the applicability of the shape effect of the satellite angle reflector, the satellites at different orbits are observed by the SLR system in Changchun Observatory, National Astronomical Observatories, Chinese Academy of Sciences. The results show that the shape effect of the satellite angle reflector has significantly changed the distribution of the echo photons reflected from the satellites, and the laser pulse width is broadened. For Starlette, the laser pulse width is extended from 50 ps to 72 ps, and the ranging error caused by the shape effect of the satellite angle reflector is about 5.4 mm. For the high-orbit Etalon-1 satellite, the middle-orbit Lageos-1 satellite and the low-orbit LARES satellite, the ranging errors are about 15.4, 7.4, 5.0 mm, respectively. Therefore, for the development of the measurement accuracy of SLR system to the mm level, the influence of the shape effect of the satellite angle reflector on the ranging precision should be considered, which also provides new ideas and approaches for the evaluation performance and structure design of SLR system.

The tropospheric aerosols in the northern suburb of Nanjing is observed and analyzed with the Rayleigh-Raman-Mie lidar in Nanjing comprehensive observation base of the China Meteorological Administration. The Raman scattering signals with the wavelength of 607 nm are smoothed by the wavelet denoising method on the basis of soft threshold filtering. The extinction coefficient of upper tropospheric aerosols is retrieved based on the data from Raman channel of the lidar. The backscattering coefficient and the lidar ratio of upper tropospheric aerosols are retrieved with the Mie scattering signal with the wavelength of 532 nm. The experimental results show that the wavelet denoising method on the basis of soft threshold filtering can remove the signal noise well and improve the accuracy of the retrieval results. In sunny and cloud-free day, the extinction coefficient of upper tropospheric aerosols is from 0.03 km-1 to 0.07 km-1, the aerosol backscattering coefficient is from 0.011 km-1·sr-1 to 0.024 km-1·sr-1, and the lidar ratio is from 22 sr to 52 sr. It is indicated that there are still some aerosol particles in upper tropospheric of the northern suburb of Nanjing. Cirrus clouds have been measured many times during the experiment. The lidar ratio cirrus clouds is 17 sr±10 sr.

In order to explore the extinction characteristics of different types of aerosols and effect of multiple scattering on the transmission performance of visible light, four wavelengths of 400, 488, 550, 694 nm and four common aerosols of oceanic, dust-like, water-soluble and soot are selected. Based on the Mie scattering theory and the steady-state Monte Carlo model, the transmission attenuation characteristics of visible light in atmosphere are studied. The results show that the scattering intensity of single particle decreases with the increase of incident wavelength. The extinction efficiency factors of oceanic, water-soluble and soot particles decrease with the increase of wavelength, whereas dust-like particles have opposite effect. The extinction characteristics of oceanic, dust-like and water-soluble particles are dominated by scattering, while soot particles are dominated by absorption. The results of simulation of multiple scattering by steady-state Monte Carlo method show that the transmission rates of light wave in oceanic aerosols, dust-like aerosols, water-soluble aerosols and soot aerosols decrease in turn. The transmission rate increases gradually with the increase of visibility. The absorption effect can be neglected in multiple scattering processes when the visibility reaches a certain degree. This conclusion is helpful to construct a more accurate image degradation model in a imaging path of specific aerosol optical thick.

Spatial laser collimation system often uses the position sensitive device (PSD) as a photoelectric receiver unit, and a semiconductor laser as a light source. Because of the nonlinearity of PSD, the diffraction characteristics of laser light will result in different responses of collimation system at different spatial distances. To reduce the repeated calibration at different spatial distances during the operation of the spatial laser collimation system, we propose a simulation method of detecting the semiconductor laser spot position by PSD. In this method, the non-linear data of PSD target surface and the laser intensity distribution data are obtained with the design of the relevant optical systems, and the response of the PSD to the spot intensity center of gravity is simulated. Then, the simulation results are verified by experiment. The experimental results show that the simulation system designed in this paper can reproduce the actual PSD detection process, and the maximum error is only 0.026 mm in the relative displacement range of 0-4.5 mm.

The echo signals have a large dynamic range in the applications of photon counting mode lidar. The far-field echo signals of lidar are within the linear output range of detectors and thus no calibration is required. However, because the near-field signals are usually within the non-linear response area of detectors,it is necessary to calibrate the near-field signals. In order to improve the signal-to-noise ratio of far-field signals, a novel calibration method is proposed. According to the pulsed lidar equation, the corrected near-field signals are obtained by the retrieval of the far-field signals, and simultaneously are compared with the received echo signals to obtain a calibration factor. The data before and after correction by this calibration factor are further used to retrieve the visibility, respectively, which is compared with the detection results obtained by the visibility sensor. The results show that, for the uncorrected data, the average deviation and standard deviation between the visibility obtained by the far-field signals and that by the visibility sensor are 0.57 km and 1.89 km, respectively, while the average deviation exceeds 10 km for the near-field signals. The visibility obtained from the corrected data is in a good agreement with that by the visibility sensor. The average deviation and standard deviation are 0.43 km and 0.76 km, respectively.

The packaging technology of fiber Bragg gratings (FBG) has a significant impact on sensor performance. So far, two packaging methods most commonly used in the design of FBG sensors are the packaging after all grating pasting and the two-end pasting-packaging after grating pretension. As for these two packaging ways, the key parameters determining the performance of a sensor, such as sensitivity, linearity, repeatability, creep and temperature-compensation, are investigated experimentally. For each packaging way, three FBGs with the same parameters are arranged on the same equal-strength cantilever. A free and bare grating is used as a temperature reference for a contrast experiment of strain sensitivity and temperature-compensation characteristics. The experimental results show that the six FBGs under the two packaging ways have good consistency in sensitivity, linearity and repeatability, and also good creep resistance. In the test of temperature-compensation, the grating self-differential compensation effect under the same packaging way is superior to the differential compensation from the reference bare FBG. The self-differential compensation effect under the two-end pasting packaging way is the best with the wavelength shift difference less than 9 pm, better than that under the packaging way after all grating pasting (within 20 pm). In contrast, the differential compensation effect under the two-end pasting-packaging way from the reference bare grating is the worst with the maximum value up to 53 pm.

Aiming at the slow measurement speed problem for the conventional Brillouin optical time domain analysis (BOTDA) technology, a fast BOTDA sensing system based on the coherent detecting digital optical frequency comb (DOFC) is proposed and designed. With the coherent detecting DOFC, the Brillouin gain spectrum (BGS) and the Brillouin phase spectrum (BPS) needed for sensing can be reconstructed without any frequency sweeping. Owe to the BGS and BPS, the Brillouin frequency shift distribution of the sensing fiber can be obtained without any averaging process, and thus the response time of the BOTDA sensing system can be greatly shortened. Through the experimental tests, it is can been seen that the response time of the sensing system over 10 km sensing fiber is 0.1 ms, and the detection accuracies are 1.6 ℃ for temperature and 44 με for strain, respectively. The corresponding maximum measurement deviations of this sensing system are about 0.3 ℃ for temperature and less than 10 με for strain. The experimental results show that the BOTDA sensing system based on the coherent detecting DOFC can be used to realize a fast, long distance and high accuracy sensing of temperature and strain.

In order to accurately extract the peak positions of correlation interference signals and provide a theoretical reference for the demodulation mechanism of the correlation interference of an optical wedge type fiber Febry-Perot sensor, the correlation interference signals of three typical spectral distributions of light source are numerically simulated and the obtained correlation interference signals are analyzed in detail. Moreover, the influence of spectral bandwidth on the width, amplitude and contrast ratio of the correlation interference signals is discussed under the condition of different spectral distributions. The analysis results show that the correlation interference signals have relatively small noises, sparse stripes and high signal contrasts when the light source has a Gaussian spectral distribution and a relatively large bandwidth. It is beneficial to search peaks accurately and realize a high precision demodulation.

The end-bonding fiber Bragg grating (FBG) sensors have many advantages such as no multiple peaks, no direct force in grating area and all points suffering the same stress, which are widely used to the packaging of substrate-type, clamp-type and other types of sensors. However, the shear deformation in the adhesive layer results in the difference between FBG strain and matrix strain, and thus the strain measurement error is introduced. In the practical applications, it is necessary to precisely obtain the function relationship between the FBG strain and the matrix strain under the influence of the shear deformation in the adhesive layer to improve the measurement accuracy of strain. For this purpose, the strain transfer equation of end-bonding FBGs based on linear viscoelasticity is derived and the average strain transfer models between fiber grating and matrix under instantaneous response and quasi-static response are developed. The parameters that influence the average strain transfer rate are discussed and analyzed. The influence law of the adhesive layer parameters with which the grating is obviously superior to that of a grating-bonding FBG sensor is demonstrated. The validity of the theoretical equation is verified by the simulation by the finite element method, and the proposed model provides a theoretical basis for the design and application of end-bonding FBG strain sensors.

The detecting method of measuring axial force of double-walled model pipe pile by fiber Bragg grating (FBG) is proposed. The relationship between the strain change of the model pipe pile and the stability and drift of the central wavelength of the fiber grating is discussed. FBG sensor is packaged in the external tube of the mode pipe pile, and is attached on the internal tube of the mode pipe pile. The strain gauges are stuck on the same position of fiber grating of external tube and internal tube of mode pipe pile, respectively. After the vertical stress is exerted on the model pipe pile, the strain of the model pipe pile is tested, and the test results of the FBG sensor and the strain gauge are compared with the theoretical formula. It is found that there are errors between them, but within test requirements. When the vertical load of 5 kN is applied to the external tube, the maximum strain measured by the FBG is 59.2 με and the relative error is 4.9%. The maximum strain of internal tube measured by the FBG is 122.3 με with a relative load of 5 kN, and the relative error is 7.2%. These errors are allowed for the engineering test requirements.

A Mach-Zehnder interferometer is employed to study the effect of magnetic field on the local corrosion process of a ferromagnetic metal. The local corrosion of a pure iron electrode in the 0.1 mol/L NaCl solution is investigated with the combination of traditional electrochemical method and potentiodynamic polarization measurement. The video images of interference fringes in the electrode/solution interface are first recorded by CCD, then these fringes are analyzed by Fourier transform and the adaptive bandpass filter, and finally the dynamic process of local corrosion in the electrode/solution interface is reproduced by the dynamic phase reconstruction of interference fringes. The results show that, without the external magnetic field, the local corrosion appears only in the middle area of the electrode/solution interface at first, but the local corrosion area of the electrode/solution interface soon expands with the increase of voltage, and finally the entire electrode/solution interface is activated and dissolved. With the external magnetic field, the corrosion degree of the electrode surface is positively correlated with the magnetic induction, and an obvious edge corrosion phenomenon occurs. In other words, the external magnetic field can accelerate the electrochemical corrosion of ferromagnetic metals.

The presence of the speckle noise greatly reduces the signal-to-noise ratio and spatial resolution of the digital holographic reconstructed image. In the lensless Fourier transform digital holography, the object light field can be reconstructed by only one step of Fourier transform on the interferogram recorded by the charge coupled device (CCD). A new speckle reduction method is proposed based on the principle that each pixel in the hologram contributes differently to the distribution of the speckle noise intensity in the reconstructed image in this paper. With this method, we divide the interferogram outputted by the CCD into several regions and exchange the positions of the regions. Many different new reconstruction interferograms are obtained, and then all new interferograms are reconstructed. Then, the reconstructed image with lower speckle noise intensity can be obtained by superimposing and averaging all of the reconstructed images, and the equivalent number of looks improves significantly. Compared with the spatial domain mask method, the proposed method reduces the processing time by 80%, and increases the edge preservation index of the reconstructed image by one times.

We propose a max margin based semi-supervised reranking method for multimedia information retrieval. We use hypergraph regularization to preserve the neighborhood of the sample in the original space and introduce the labeled and unlabeled sample information to construct the objective function, so as to achieve full and efficient use of data information for ranking. By using a small amount of annotation samples to construct the priority relationship pairs, the priority information between samples is introduced into the objective function to construct a ranking learning model. This method can show users in priority the results that meet their demand better, and improve the retrieval accuracy. The experimental results on MSRA-MM 1.0 dataset suggest the proposed method provides superior performance compared with several state-of-the-art methods.

The visual background extraction (Vibe) algorithm cannot effectively remove the shadow of the target, and cannot quickly remove the ghost phenomenon. To address the shortcomings of Vibe, an improved YUV_Vibe fusion algorithm is proposed. The algorithm expands the value range of the sample field, which effectively avoids the repetitive selection of the same samples. The updating factor is adjusted from 16 to 4, and the number of sample updates is set at 2, which accelerates the update rate of the background to eliminate the rate of ghost detection. The fusion of the YUV color information features with the Vibe algorithm eliminates the influence of shadows. By constructing a double fusion model, the false detection rate of shadows is effectively reduced. The algorithm is experimentally applied to video datasets. The test results reveal that the improved YUV_Vibe fusion algorithm has improved the accuracy and recognition rate, and the experimental detection results are more accurate.

A new algorithm is proposed to extract the edge of the closed point cloud through the edge seed point. Using the algorithm can extract the sharp edge and intersecting edge by setting an appropriate search angle. The algorithm determines whether a point has the characteristic of edge point by simulating the aggregation degree of pull between points. First, build the kd-tree to find each point quickly. Second, define the neighbor radius of the point cloud, and select the new edge points by calculating the edge coefficient of the seed point located in the neighbor radius. Use the search angle of the seed point to control the search direction. Finally, extract the closed point cloud's edge. Choosing different seed points and search angles can extract different edges. The effects of the neighbor radius and the search angle were analyzed by the experiment. The appropriate neighbor radius and search angle were used to extract the sharp edge and the intersecting edge, and the accurate edge of the point cloud was obtained, which proves the effectiveness of the algorithm.

A phase unwrapping algorithm is proposed based on the coefficient solution of wavefront aberration mode. The algorithm is implemented on the basis that wavefront can be accurately represented by Zernike polynomials. By solving the coefficients of the corresponding Zernike polynomials, we realized a fast and accurate phase unwrapping. Taking the phase unwrapping in the wavefront sensing based on binary aberration mode measurement as the research target, we verify the feasibility of the algorithm by numerical simulation. Numerical simulation results show that compared with the least square phase unwrapping algorithm, the proposed phase unwrapping algorithm realizes a great improvement in accuracy and speed of unwrapping.

Scanning electron microscopy (SEM) imaging can visually reveal the microscopic world. In SEM imaging, the device parameters must be repeatedly adjusted to ensure the optimum image contrast. This process is often time-consuming and labor-intensive. We propose a novel no-reference quality assessment method for evaluating the SEM image contrast distortion based on multi-scale characteristics, which can be used as a guide to select imaging parameters. Firstly, a SEM image database is established, and the corresponding subjective mean opinion score (MOS) is obtained via subjective experiments. According to the multi-scale characteristics of the human visual system, 10 features are extracted, including singular value decomposition similarity with different scales, frequency domain features, and entropy. The MOS values and 10 features are then used to train a regression model via support vector regression. Finally, this model is used to predict the image quality score. The experimental results reveal that the proposed method can maintain a high level of consistency with subjective evaluation results, and its performance is superior to the mainstream full-reference and no-reference quality assessment methods.

Fine feature matching in image stitching and blurred panorama areas is time consuming due to inaccurate image registration. To mitigate this issue, this study proposes a new image registration model based on improved grid motion statistics and weighted projection transformation. The method uses the oriented fast and rotated BRIEF (ORB) algorithm to extract and describe the image features. The brute-force matching algorithm is used for rough image matching. The image is divided into multiple square grids. Then, these grid features are counted and five grids feature scores are calculated to eliminate error matching and obtain the refined matching feature set. Lastly, image registration is achieved by adding a distance weighting coefficient to construct the weighted projection transformation model. Comparing the proposed algorithm with other methods used in the stitching sequence set, the experimental results revealed that the accuracy of the proposed algorithm was improved by an average of 28.7% for image registration and the feature matching speed was improved by 43.6%. The stitched panoramic image did not show any obvious geometrical dislocation or distortion, and the overall imaging appears quite natural.

Pedestrian detection is the key technology to realize intelligent traffic and passenger flow monitoring. Currently, the training model of deep learning method has achieved good results in pedestrian detection. However, when the training samples are poor, the training model often fails to achieve good results. In order to improve the effect of pedestrian detection under hazy weather and strong exposure environment, the dark channel defogging algorithm is applied to pretreat deep learning samples. And pedestrian detection model is trained with fast deep convolutional neural network. In this experiment, the dark channel defogging algorithm is applied to preprocess the 10,000 sample images. After that, the sample images preprocessed by the defogging algorithm with and without dark channel are used to train model, respectively. Finally, detection accuracy of these two models under different scenarios are compared. The experimental results show that the depth model obtained by using the dark channel defogging pretreatment sample has a better detection effect and the detection rate increases under many scenarios.

Aiming at the problems in the face recognition algorithm based on Weber features that the directional information is not made full use and the extracted information is also insufficient, we propose a novel face recognition method based on multi-directional Weber gradient histograms. On the basis of original differential excitation, the neighborhood pixel gradient is increased, and the improved differential excitation and Weber gradient features are extracted. The improved differential excitation and Weber direction are quantized, and the two-dimensional histograms are extracted in blocks, which are further converted into one-dimensional histogram features. The histogram features are extracted along the Weber direction. Two features are connected to form a compound feature and simultaneously the nearest neighbor classifier is used for classifying. The experiments on different face databases show that the proposed method has not only a good recognition effect, but also a relatively strong robustness to illumination, expression and partial occlusion.

A foggy image sharpening algorithm based on nonsubsampled Contourlet transformation (NSCT) is proposed. The foggy image is mapped to the HIS color space, and the luminance component H and the saturation component S are processed respectively. The NSCT is used to process the luminance component H. The low-frequency component containing most energy is negated and then processed by the improved single-scale Retinex algorithm. The image is negated again and is superposed linearly with the low-frequency components processed by the improved single-scale Retinex algorithm directly. A fast bilateral-filter is applied to the high-frequency components that contain most of the linear details of the image. Then the two processed components are inversely transformed by NSCT, and the processed luminance components are obtained. Finally, the saturation component S is linearly stretched to achieve color compensation. The processed image of each component is mapped backward to the RGB color space to get a clear foggy image. The experimental results show that the proposed algorithm achieves good results of the details and color fidelity for the foggy image. Compared with other algorithms, the standard deviation, information entropy and peak signal to noise ratio are improved.

The traditional algorithms for the extraction of sign language features only rely on the low-level features to realize recognition, which makes it difficult to obtain the high-level semantic features and the misunderstanding of sign language is further induced. To solve this problem, the idea of image semantic analysis is introduced into the study of sign language recognition and then an optimized fully convolutional neural network algorithm is proposed. The fully convolutional neural network is used to extract the semantic features of sign language images and the discriminative random fields for semantic annotation is used for the post-smoothing to recover the detailed information among pixels and thus the sign language recognition is completed. The experimental results show that the proposed algorithm has strong robustness and can be used to obtain the semantic features effectively. Compared with the traditional algorithms, this method can be used to identify sign language accurately with an average recognition rate of 97.41%.

In order to solve the problem of insufficient segmentation of brain tumors in magnetic resonance imaging (MRI) caused by noise, poor contrast, and diffused boundaries of tumors, a three-dimensional (3D) segmentation algorithm for brain tumor MRI images based on the improved continuous max-flow is proposed in this paper. Firstly, three types of MIR images, Flair, T1C and T2, are pre-processed with median filtering and fast fuzzy C means clustering. Then, the pre-processed images are linearly fused in the ratio of 5∶1∶4 (Flair, T1C, and T2) which is statistically observed from a large amount of experiments. Next, the 3D fused image is clustered by the fast fuzzy C-means algorithm to obtain the 3D under-segmented image. Finally, the proposed improved continuous max-flow algorithm acts on the 3D under-segmented image to obtain the final segmentation result with scattering points removed according to the analysis of the structural features and statistical characteristics of the 3D under-segmented image. The experimental results show that the average Dice coefficient, precision, and recall of the proposed method relative to the gold standard is up to 0.90, 0.94, and 0.86, respectively. The proposed algorithm can realize the 3D segmentation of the target regions precisely and automatically to meet the clinical medicine requirement.

It is one of the main aims for the most astronomical observations to obtain the diffraction-limited images by the large ground-borne optical telescope. The limited resolution induced by the atmospheric fluctuation can be overcome by the addition of non-redundant aperture masks on the telescope pupils and the closure phase technique. The progress, principle and applications of non-redundant aperture masking technique are reviewed, and the mathematical model for the application of this technique into the binary-star detection is established. The computer simulation and the corresponding experimental investigation are conducted. The research results show that one can obtain the closure phase by processing the interferogram obtained from the non-redundant aperture masking and thus the binary-star contrast ratio and angular separation can be obtained by fitting algorithms.

The sensitivity of schlieren equipment is the key issue of the experimental results in the density measurements of fluid. A sensitivity measuring system was designed, taking a hot air gun as the disturbing region. The effect of flow velocities, temperatures of the hot air, and equipment parameters on the sensitivity of the system was investigated. The images were obtained under various temperatures and air densities. The sensitivity of the schlieren system could be characterized by the temperature variation, so called fringe number on the image. The more obtained fringe number of the image in the hot air gun disturbed schlieren imaging system, the higher sensitivity and the wider range of temperature variation. The optical paths of the schlieren imaging system were simulated by Zemax, and the experimental results agree with the theoretical analysis. The results show that the hot air disturbing method could characterize the sensitivity of schlieren equipment effectively, which provides a convenient and sensitive approach for sensitivity measurement of the schlieren system.

Regarding to the total generalized variational model cannot fully utilize the self-similarity information of the image structure when reconstructing images, an improved generalized variational image reconstruction model under non-local constraints is established to improve the quality of image reconstruction in the sparse sampling situation. This model introduces a non-local self-similarity of the transform domain as a priori information for image reconstruction. And the multi-directional total generalized variational regularization constraint is calculated in the eight-neighborhood space to protect the structural characteristics of the image. Further, the augmented Lagrangian theory is used to remove the constraint and solve the model, and an image reconstruction algorithm based on the improved total generalized variation is proposed. Simulation experimental results show that the proposed reconstruction model and image reconstruction algorithm can effectively remove the artifacts and noise in the image and meet the requirements of image reconstruction quality under sparse sampling condition. Compared with the famous reconstruction algorithms, the images reconstructed by proposed algorithm has significant improvement in both subjective visual effects and all objective evaluation indicators.

The phase modulator is a key component of the beam synthesis unit in the coherent synthetic aperture telescope, and the accurate measurement of the motion information of the phase modulator is one of the key technologies for phase modulator control. The three-degree-of-freedom (pitch, azimuth, and piston displacement) measurement method based on the plane normal vector is proposed. The angle and displacement information are solved according to the change of the normal vector. The proposed method is described in detail and all the factors influencing measurement accuracy are analyzed. The experimental results show that the displacement range is 0-3 μm when the rotation ranges of the rotation angles α and γ are 0-0.44 mrad and 0-0.28 mrad, respectively. Within the measurement range, the angle error is less than 3.3 μrad and the displacement error is less than 50 nm. The proposed method has the advantages of simple structure, high measurement accuracy, and so on, which can be widely applied in the three-degree-of-freedom measurement of phase modulators and others.

Digital image correlation method (DIC) has higher requirement for search speed and measurement accuracy in displacement measurement. Some good algorithms in sub-pixel level measurement are difficult to take account of speed and accuracy. Therefore, a combinatorial DIC algorithm is proposed in this paper. It is based on an adaptive genetic algorithm to rapidly obtain the maximum correlation coefficient point at the sub-pixel level. Then,this point is used as a fitting window, and a sub-pixel displacement value is obtained by spline interpolation and quadric surface fitting. Taking the simulated speckle image as the research object, we analyze the measurement accuracy, measurement stability, and the search speed of the combinatorial DIC algorithm. The results show that the combinatorial DIC algorithm has obvious comprehensive advantages over a single algorithm in search speed, measurement accuracy, and measurement stability. It can meet the speed and accuracy requirements of DIC in sub-pixel micro displacement measurement.

To overcome the drawbacks of low accuracy and flexibility of the traditional camera calibration methods, a camera calibration method based on the phase target is proposed. First, the liquid crystal display (LCD) is used as the phase target to generate the orthogonal sinusoidal fringe patterns in red and blue channels. Only one fringe pattern is used to obtain the unwrapped phase map in both the horizontal and vertical directions via the Fourier transform and the spatial phase unwrapping algorithm. Then, the phase target is arbitrarily placed at several different positions within the depth field of this camera to establish an accurate mapping relationship between the world coordinate and the pixel coordinate. Finally, the intrinsic parameters of this camera are calibrated and the accuracy of calibration results is evaluated by means of the re-projection errors. The experimental results show that we can obtain the unwrapping phase map of two directions at each phase target position by collecting only one fringe pattern with the proposed method. Compared with those in the existing methods, not only the calibration steps in the proposed method are simpler, but also the re-projection error is only 0.042 pixel.

The TiC+TiB2 reinforced Ni-based self-lubricant composite coatings are prepared on the TC4 matrix surface by the laser cladding technique, and the effect of CeO2 content on the phase compositions, microstructures, microhardness and friction and wear performances of these coatings is investigated. The results show that the phase compositions of these coatings with different CeO2 contents are mainly composed of matrix γ-Ni, ceramic-reinforced phases of TiC and TiB2, intermetallic compound of Ti2 Ni, and lubricant phase of Ti3SiC2. The addition of suitable content of CeO2 can make the reinforced phase size decreased and grains refined. When the CeO2 content is 1.5%, the microhardness is around 1400 HV, the lowest friction coefficient is between 0.22 and 0.24, and the lowest wear mass loss is 1.7 mg, which indicate that these coatings possess nice self-lubricant and wear-resistant performances.

In order to investigate the influence factors of the fiber welding modes for 304 stainless steel, a group of experiments with the laser power, welding speed and defocus distance as the variables are designed in accordance with the idea of an uniform experimental design. By the analysis of the formed welds, the characteristic forms of the heat-conduction welding and the deep-penetration welding are obtained in the fiber laser welding process. The relationships among the weld mode, each welding parameter and their interactive action are calculated quantitatively, and the influence factors of the welding modes are determined. The results show that, there are four typical weld formation forms determined by the heat input density for the 304 stainless steel fiber laser welding. The weld has an arc shape when the heat input density is smaller than 43.04 J·mm-3 and the corresponding welding mode is the heat-conduction welding. When the heat input density is larger than 66.82 J·mm-3, the weld has a nail-head or a nail shape and the corresponding welding mode is the deep-penetration welding. Meanwhile, there exists an intermediate mode when the heat input density is between the two values for the above welding modes.

The laser cladding technique is adopted for the fabrication of 2205 dual-phase stainless steel/TiC composite coating on the 16Mn steel surface. The effect of TiC content on the microstructure, microhardness and frictional wear properties of cladding layer is investigated and the relationship between the melting and precipitation behaviors of TiC in the cladding layer and the properties of cladding layer is also discussed. The results show that the dilution rate of cladding layer increases gradually with the increase of TiC content. Moreover, the TiC particles melt and precipitate during the laser cladding process. The microhardness of cladding layer increases gradually as the TiC content increases, and the highest microhardness of cladding layer is up to 612 HV when the TiC mass fraction is 15%. Simultaneously, the wear weight loss of cladding layer is the lowest.

Based on an artificial neural network, the training function fitting of the laser-cutted Ni-based alloy samples is conducted. With the current, pulse width, cutting speed and defocusing amount as the input vectors and the comprehensive score of the slag width, kerf width, and cutting integrity as the output vector, the hidden layer node with the minimum error is found. Based on this model, the laser cutting quality is predicted. The maximum error is 7.66% and the minimum error is -0.32%. With the comprehensive score as the fitness value of the genetic algorithm, 50 species within the range of the practical process parameters are randomly selected as the initial optimal group. The treatments such as crossover, mutation and iteration are then made and the optimal fitness value and its corresponding process parameters are searched. The optimal fitness value which is predicted theoretically is 98.41, but the actual value is 89.53, and the error is 9.03%. The quality of this verification sample is obviously higher than those of 25 experimental samples. The average slag width is 81.5 μm and the kerf width is 164 μm.

The physical content of smoothing by spectral dispersion (SSD) technology is theoretically explained that the focal spot smoothing effect of SSD results from the temporal change of the far-field primitive electric field phase, and there is no physical image of "moving" focal spot in the practical application of SSD. On this basis, the theoretical model of "intensity sweep" laser beam, which has a far-field intensity distribution with constant form but temporal changed spatial position, is established. The smoothing characteristics of the beam are studied by numerical simulation compared with that of SSD.

A continuous fiber laser is used for the tailored blank laser welding of 800 MPa and 1000 MPa dual-phase (DP) steels and the effects of martensite content on the microstructure and properties of laser welded DP steel joints are investigated. The results show that the laser welded DP steel joints with different martensite contents are all composed of the full-martensite area and the non-full-phase-transformation area. The widths and microhardness of two laser welded joints are relatively reduced, and their tensile fracture positions are both in the base materials (BMs). The 1000 MPa welded DP steel joint directly cracks and expands along the non-full transformation area during the erichson test and its erichson value is 85.0% that of the BM. In contrast, the 800 MPa welded DP steel joint cracks perpendicular to the full-martensite area and its erichson value is up to 91.7% that of the BM.

A visual background extracting (Vibe) algorithm based on multiscale space is proposed to solve the problem that there exist ghosts and dynamic background disturbance in the target detection process for the conventional Vibe algorithms. Before the background model is established, the input video sequence is pyramid transformed to obtain three different resolution images at the top, middle, and bottom layers. Then the Vibe foreground detection is performed at different resolutions, and the detection results are fused to reduce the influence of dynamic background. At the same time, a ghost elimination strategy is proposed which enhances the ghost elimination by using the information between frames and by adding a second judgement. Finally, in order to well adapt to the dynamic environment, a metrology under a complex background is proposed, in which the threshold can be adaptively adjusted according to the complexity of backgrounds. The experimental results show that the improved algorithm accelerates ghost elimination and has strong robustness to dynamic environmental noise disturbances.

Three-dimensional reconstruction of the topography of objects is extensively studied in the field of vision measurement. Current industrial cameras use the binocular vision measurement method; however, this method requires expensive equipment, specialized software, and trained technical personnel as well as involves complicated operational processes. To mitigate these shortcomings, this study proposes a three-dimensional reconstruction method based on a smartphone camera. Multiple pictures are captured from different angles using a smartphone, and their three-dimensional models are reconstructed using smartphone camera calibration, image preprocessing, feature point detection and matching, basic and essential matrix calculation, and singular value decomposition. Experimental results show that the proposed method can achieve the measurement accuracy of the general industrial camera, meet basic requirements of three-dimensional topography modeling, and considerably expand the scope of vision measurement applications.

The proportion of traffic accidents caused by driver factors is high, therefore, it is of great significance to study a recognition method for the correct identification of abnormal driving behavior by analyzing the driver activity state. We propose a recognition method of abnormal driving behavior based on the covariance manifold and two classification of multi-class LogitBoost classifier. First, we extract the basic features, such as texture, color and gradient direction, to overcome the shortage of recognition of driving behavior based on a single feature. Then, we use the covariance manifolds for the multi-feature fusion to eliminate the feature redundancy and reduce the impact of image processing and recognition due to excessive differences in numerical values of different features. Finally, the classification and identification are performed using a multi-class LogitBoost classifier based on two classification. The experimental results show that compared with the traditional multi-class LogitBoost method, the proposed method greatly improves the correct rate of multi-classification, and the correct recognition rate for different targets can reach 81.08%.

An aurora sequence classification method based on deep learning is proposed. It combines the rich spatial domain information and the sequence information corresponding to the advantages of convolutional neural network (CNN) features and long short-term memory (LSTM) network. In addition, aurora attributes employed as feedback constraints to the CNN make features more suitable for aurora images. Supervised aurora sequence classification and unsupervised aurora event detection are performed on the Chinese Yellow River Station All-Sky Imager (ASI) dataset. The experiment shows that our method can characterize aurora sequences effectively and can be able to implement automatic classification for massive aurora sequences.

Fast recognizing, positioning and surface detection of multi-shaped objects in complex environment are studied to satisfy the requirement of smart machines, which is expected to grab the objects or inspect surface defection in complex environment in real time. Fast recognition, positioning, stereo matching and post-processing algorithm of point clouds are discussed. At first, new targets in the scene are recognized by robust principal component analysis, and the image location of the targets is accurately acquired by improved k-means clustering algorithm. Then, the region of interest is screened out by support vector machine, and one-dimensional searching is carried out by epipolar restriction to obtain the regions to be matched in binocular images, and local three-dimensional point clouds are quickly obtained. Finally, special denoising operation of point clouds is conducted to reduce the error. The experiment results indicate that the proposed algorithm effectively reduces the running time of the process and effectively reduces all the noises caused by complex backgrounds, and improves the accuracy and adaptability of point clouds acquisition in complex environment, and it is a robust, effective and fast algorithm for three-dimensional point clouds acquisition.

Using machine vision technology to solve the problem of textile edge detection is the primary method in the field of textile edge detection. Edge detection algorithms are the key to improve the speed and precision of textile edge detection. In this paper, according to the requirements of the edge detection system in practical production, the problems existing in various image processing algorithms applied to the textile edge detection are studied, and the grayscale images of the textile with various characteristics are analyzed. Then an adaptive threshold edge detection algorithm is proposed. The algorithm can obtain the edge position of different textile quickly and accurately. The experiments show that the accuracy of the algorithm can reach 0.1 mm and the algorithm has a small running scale, which meets the requirements of high speed and high precision for the edge detection algorithm, and provides support for the following overall design of the edge detection system.

In the cases of appearance changes and viewpoint changes, the accuracy and robustness of traditional visual loop closure detection algorithms become very poor.To overcome this problem, we propose a loop closure detection algorithm, which utilizes the features of multi-level convolutional neural networks. The high-level convolution features contain much semantic information and can cope with viewpoint changes. The medium-level convolutional features contain more geometry and spatial information, which is more robust to lighting changes. Therefore, the accuracy and robustness of loop closure detection is improved by taking full advantage of the characteristics of the middle and high levels convolutional features and modular similarity measures. However, the convolutional feature vectors have a particularly large dimension, so the convolutional feature vectors are firstly dimension-reduced. The experimental results on the Gardens Point dataset show that the image matching detection effect is better by using multi-level convolutional features than by other single layers. In addition, for the dynamic interference factors in the images captured at different moments, a dynamic interference semantic filtering mechanism is further proposed. The filtered images are used to perform the matching. The experiments on the Tokyo 24/7 dataset prove the feasibility and effectiveness of this method.

The laser welding head is integrated with the CCD video tracking module, and a method for automatic seam detection based on a word line laser is proposed. This method uses laser triangulation to obtain the height, width, and other shape information of the weld. In the aspect of image processing, we perform median filtering and binarization on the obtained weld area image to obtain the two-dimensional coordinates of the weld feature points, achieve target position and calculate its three-dimensional coordinates in the world coordinate system. In the tracking algorithm, we use high accuracy and high speed kernel correlation filters target tracking algorithm to track the positions of common straight and curved welds. The verification experimental results show that the fitting curve and weld shape error are within 5%, the degree of coincidence is high, and the real-time tracking effect is good.

A numerical simulation model of the temperature field in the laser phase-transformation hardening of the highly-enhanced diesel engine valve seats is established and the influences of different process parameters on this temperature field are investigated in the laser phase-transformation hardening process. The results show that, with the increase of the laser power, the peak temperature of the phase-transformation hardening of the valve seats increases, the temperature field presents a cometary shape with a tail, and the depth and the width of the hardened layer increase synchronously. In contrast, with the increase of the scanning speed, the peak temperature of the phase-transformation hardening of the valve seats decreases, the cometary tail of the temperature field gets smaller, and the depth and the width of the hardened layer decrease synchronously. Moreover, with the increase of the laser spot radius, the peak temperature of the phase-transformation hardening of the valve seats decreases and its corresponding position shifts, the cometary tail of the temperature field gets larger, the depth of the hardened layer decreases.

The yttrium aluminum garnet(YAG) ceramics are prepared by using the solid-phase sintering with YAG∶Ce3+ phosphor powder as the raw material. In order to improve the sintering performances of YAG ceramics, the active SiO2 is coated on the YAG∶Ce3+ phosphor powder surface by the sol-gel method and the CuO-TiO2 composite phase is added as sintering aids. The particle size, the optical performances before and after coating, and the micro-morphologies of the phosphor powder are analyzed and the mechanical properties and microstructures of YAG ceramics are investigated. The research results show that the particle size can be effectively reduced by grinding and ball-milling. The sintering and densification process of ceramics is promoted by the coating of the active SiO2. When the coating amount (mass fraction) is 2%, the sintering temperature of YAG ceramics is decreased to 1575 ℃, the relative density reaches 96.3%, the rock-well hardness is 87.6 HRA, and the fracture toughness is 1.8 MPa·m1/2. Moreover, the sintering effect of the CuO-TiO2 composite sintering aids is superior to that of the CuO single sintering aids. When the mass fraction of CuO-TiO2 is 2% and the mass ratio between CuO and TiO2 is 1∶2, the sintering temperature of the YAG ceramics is reduced to 1450 ℃, the rock-well hardness is 88.5 HRA, the fracture toughness is 1.7 MPa·m1/2.

In order to improve the efficacy of phototherapy for neonatal jaundice, we design and develop a neonatal jaundice phototherapy light source by spectral matching technology. In order to study the degradation of bilirubin, we use the designed special blue LED source, the ordinary blue LED source, the traditional blue source, and the white LED source to irradiate standard bilirubin solution in vitro with concentration of 130.932 μmol/L for 1 h, 2 h, 4 h and 8 h, respectively. The results show that the blue LED light source and the traditional blue light source have obvious degradation effect on the bilirubin solution, and the designed special blue LED has the best degradation effect. The key factor of phototherapy for neonatal jaundice is the spectrum of the light source,and the designed special blue LED can effectively reduce the concentration of bilirubin in vitro.

The head-up display (HUD) system can help drivers to check the speed, navigation, and other related information conveniently, shorten the visual blind spot, thus reduce traffic accidents, and improve the driving experience. In order to speed up the popularization of the vehicle-mounted HUDs, a mini-projector system based on the digital micro-mirror device (DMD) is proposed, which focuses on the reflective optical system conforming to the vehicle structures, and the simulation model is provided. The image enters into human eyes through multiple reflections by the projection screen, thus forming a virtual image in front of the windshield. The design result shows the virtual image distance can reach 2.5 m so that the driver can view the image without changing sight distance. The horizontal and vertical viewing ranges reach 140 mm and 60 mm, respectively. The optical transfer functions at different positions are all close to the diffraction limit, which indicates the clear image and good projection effect. The horizontal and vertical fields of view are 12° and 6.75°, respectively, which can provide sufficient display area without disturbing the drivers' visual field.

In the confined space, heat dissipation problem is one of the factors that restrict the development of high power LED in automotive headlamps. A flat panel micro heat pipe group with copper powder as sinter core, is proposed as a radiator used in automobile LED headlamp. Under different working conditions of vehicle lights, the effects on the radiator performance of thermal resistance, ambient temperature, and convective heat transfer coefficient of the fin group are studied. Finally, the heat dissipation experiment of the micro heat pipe for vehicle light is conducted. The results show that, under normal working conditions, when the power of the LED lamp is 60 W, the junction temperature of the LED is 67 ℃, and the thermal resistance of the radiator is stable at 0.61 K/W. Under complex and high power conditions, the junction temperature of LED is below 120 ℃,and when the convective heat transfer coefficient of the fin is greater than 80 W/(m2·K), the junction temperature of LED tends to be stable. Meanwhile, the performance of the radiator is verified by experiments. When the output power is 50 W, the LED junction temperature of the heat pipe radiator is 58 ℃, indicating that the new heat pipe radiator can meet the heat dissipation requirements of the LED headlamps.

Based on the double refraction and double reflection phenomena and on the electro-optic effect of crystals, a 2×4 free-space 90° optical hybrid with an integrated structure is designed. This optical hybrid is composed of an electrically controlled birefringence crystal and a birefringence crystal which is used as an analyzer. With the electrically controlled birefringence crystal, the splitting and coupling between the signal light and the local oscillation light are realized by means of four internal double reflections and the electro-optic modulation. By adjusting the control voltage of the electrically controlled birefringence crystal and the double refraction effect of the analyzer, four-route signal-local mixed lights with a 90° phase shift are obtained. This hybrid exhibits the advantages of simple and compact structure, low loss and stable performance, which is applicable to the spatial coherence communication system.

In order to improve target location accuracy of the sky-wave over-the-horizon radar, a target localization model was proposed based on the improved dragonfly algorithm to optimize the extreme learning machine for the multi-static sky-wave over-the-horizon radar system. Firstly, in order to avoid dragonfly algorithm falling into local optimum, the Logistic chaotic mapping, reverse learning strategy and mutation process are introduced into the dragonfly algorithm to create an improved dragonfly algorithm. Then, the improved dragonfly algorithm is used to optimize the weight and hidden layer bias of the extreme learning machine. Finally, the optimized extreme learning machine is applied to multi-static sky-wave over-the-horizon radar location. Theoretical research and simulation results show that the method can achieve high locating precision of target, and its location accuracy and reliability are better than those of current sky-wave over-the-horizon radar location methods and target location methods based on back propagation neural network and radial basis function neural network. A new target location method is provided for the multi-static sky-wave over-the-horizon radar system.

Full waveform data decomposition is critical to obtain the effective information of tested object from the large-spot laser altimetry data. Gaussian decomposition and wavelet decomposition are two universal methods to achieve the full waveform data decomposition in the large-spot laser altimetry currently. However, the decomposition effect and accuracy are ambiguity for different echo signals of ground objects. In this paper, the two methods are applied to the full waveform data of the Geoscience Laser Altimeter System (GLAS), and the waveforms of several typical ground objects in flat and slop areas are decomposed. The results are analyzed and compared qualitatively and quantitatively according to the index of goodness-of-fit and the iteration times of achieving the best goodness-of-fit. The results show that the Gaussian decomposition and Guassian wavelet decomposition are nearly same for the accuracy of best goodness-of-fit. With the increase of the complexity of the ground objects, the iteration times of achieving the best goodness-of-fit with the Gaussian decomposition is less than that with the Gaussian wavelet decomposition.

Airborne laser radar (LiDAR) system can directly and effectively obtain three-dimensional point cloud information of ground features, to provide powerful data guarantee for the generation of digital elevation model, building detection and three-dimensional reconstruction. However, the original point cloud data will inevitably produce noise points. A method of noise removal for airborne LiDAR point cloud based on the multi-window top-hat transformation is proposed. The grid interpolation is performed on the point cloud according to the interval of point cloud to obtain the maximum and minimum grid data, respectively. The grid data is clustered, and the original noise areas are detected by setting the area size threshold. The maximum and minimum grids are processed using the white and black top-hat transformation theory respectively to detect the final grid area where noise points are located. The method is compared and analyzed with other methods based on the ISPRS data. The results show that the proposed method can remove the noise points, and completely preserve the details of the original point cloud.

The femtosecond laser direct writing technique is widely used in the micro-fabrication field, and it has been achieved important progresses in the field of bioinspired materials with special wettability. The recent application progresses at home and abroad of the femtosecond laser micromachining technique in the bioinspired superhydrophobic field are summarized, and the analysis from the three aspects of the basic superhydrophobic surface materials, the superhydrophobicity-related functional wettability and the applications of the superhydrophobic materials is completed. The future challenges and progresses in this field are prospected.

The technical methods of improving the anti-laser damage performance in the world are discussed. The advantages of different methods to improve the laser damage resistance are pointed out. Microstructure, material system, coating fabrication method, damage threshold, reflectivity, absorptivity, etc. are mainly illustrated. We summarize the research progress in manufacturing different materials coatings to improve the anti-laser damage performance. From the perspectives of microstructure, particles, and new research methods, the future research directions are prospected as follows: the effect of particle diameter on laser damage performance will be studied, and such programs as MATLAB and ANSYS will be used in the anti-laser damage performance research.

The spatial resolution, sensing distance, measurement accuracy and measurement time of the Brillouin optical time-domain analyzer (BOTDA) are interrelated to each other. How to improve the performance of the BOTDA system has been a hot topic in the field of distributed optical fiber sensing. Because the sensing distance and measurement accuracy of the BOTDA system are closely related to the signal-to-noise ratio (SNR), the performance enhancement of the system focuses on the improvement of the SNR and the spatial resolution. The technical methods to improve the performance of the BOTDA system are reviewed. These technologies have extended the sensing distance and improved the measurement accuracy of the system, which make the BOTDA system more suitable for the practical engineering. After analyzing the existing problems, the future research directions are prospected as well.

Three-dimensional (3D) scanning based on optical principle is a technology of scanning the object's spatial shape by optical system, which can acquire 3D information of the objects. The technology has the advantages of non-contact, high precision and high resolution. To our knowledge, the structured-light 3D scanning accuracy is up to 0.01 mm, and the point cloud contains millions of points with the working distance less than 1 m. Texture reconstruction can further present the color, material and other information of the scanned objects, and improve the verisimilitude of reconstructed objects. Due to the influence of camera error and illumination environment, it is easy to produce seam, blurring and ghosting in texture images after texture mapping. By introducing the camera model, we derive the relationship between 3D space point and two-dimensional (2D) image, and then analyze the causes of texture artifacts. The methods of eliminating residual artifacts in texture construction are reviewed, and their advantages and limits are summarized. At last, in view of the shortcomings of texture reconstruction, the development trend of the texture reconstruction method for colored 3D model is prospected.

A multi-denoising method based on the empirical mode decomposition (EMD) is proposed to minimize the interference noises in the multi-pass absorption spectra. The denoising number is theoretically analyzed and determined. The CO2 spectral lines with different volume fractions detected by the direct absorption technology and wavelength modulation spectroscopy at second harmonic are denoised by the multi-EMD, and the results are compared with those by the denoising methods with multi-averaging and low-pass filtering. The results show that the interference and random noises can be effectively filtered by the multi-EMD denoising method. Moreover, the denoised signal amplitude has a good linear relationship with the volume fraction of the gas to detect, and the detection sensitivity of the whole system is up to 3.5 × 10-5. The spectral signals of CO2 standard gas at different pressures and temperatures are detected. After the multi-EMD denoising, the interference amplitudes of the spectral signals are reduced by about two orders of magnitude, but the signal to noise ratios are enhanced by about two orders of magnitude.

In order to more accurately analyze the importance of spectral absorption characteristic parameters, which in different soil moisture absorption bands in soil spectra, in soil moisture content estimation, we collect 38 soil samples in Ugan-Kuqa river oasis in Xinjiang to measure soil spectral reflectance and soil moisture content. The characteristic parameters of spectral water absorption are extracted with the continuum-removal method, the features include the maximum absorption depth D, the absorption peak right area Ra, the absorption peak left area La, the absorption peak total area A, area normalization maximum absorption depth DA, and symmetry S. With the correlation analysis of the features and soil moisture content, we use random forest method to classify the characteristic parameters of spectral water absorption, and obtain the importance of each parameter to soil moisture content. Multiple stepwise regression model is used to establish soil moisture content inversion model. The results are as follows: D and A have the strongest correlation with the soil moisture content, the correlation between spectral absorption parameters in the band of 2200 nm or 1400 nm and SMC is better than that of 1900 nm band; the top five parameters that are important for soil moisture content are obtained, they are D2200, La2200, A2200, D1900 and Ra2200, respectively; the best prediction model of SMC is the multiple stepwise regression model with A2200 and D2200, the decision coefficient of the modelling set is 0.88, root mean square error of modeling set is 2.08, decision coefficient of the test set is 0.89, prediction root mean square error is 2.21, and the relative analysis error is 2.80. Random forest classification can obtain the important spectral water characteristic parameters which have great influence on soil moisture content, and it provides a new method for accurate and rapid estimation of soil moisture content in arid areas.

A method for detecting chemical oxygen demand (COD) in water based on ultraviolet and fluorescence multi-spectral fusion is proposed. The experimental samples are 53 actual water samples, including coastal seawater and surface water. The physicochemical values of the experimental samples are calculated by the standard chemical method, and the ultraviolet absorption spectra of the samples are collected by the ultraviolet-visible spectrometer, and the three-dimensional fluorescence spectra are collected by fluorescence spectrophotometer, then the processed spectral data are used to build model. Using the ant colony-interval partial least squares(ACO-iPLS) as feature extraction algorithm and the particle swarm optimization least squares support vector machine(PSO-LSSVM) as modeling method, we establish the prediction model based on ultraviolet absorption spectra and fluorescence emission spectra at single excitation wavelength, the data level fusion model and the feature level fusion (mid-level data fusion, MLDF) model based on ultraviolet and fluorescence multi-spectral information, respectively. And the prediction results of various models are compared. The results show that the prediction effect of the MLDF model based on ultraviolet and fluorescence multi-spectral information is optimal, and the prediction accuracy of COD in water is relatively high. The determination coefficient of calibration set is 0.9999, the prediction determination coefficient of validation set is 0.9912, and the root mean square error in prediction set is 1.1297 mg/L. It provides a new research idea and solution for the rapid detection of COD in water.

To solve the problem that the traditional spectral matching algorithms have low accuracy in matching spectral data of different pigment materials in the same color system, we propose an adaptive threshold edit distance spectral matching algorithm. The edit distance is researched to improve the matching accuracy by using its characteristics of being sensitive to the spectral reflectance difference. At the same time, by adaptively setting the judging conditions of the edit distance, we reduce the error of this algorithm in matching the spectral data of the same pigment materials under different conditions. The results show that the matching accuracy of the adaptive edit distance algorithm is higher than that of the traditional spectral matching algorithms, and the recognition results of the adaptive edit distance algorithm for the pigment is better than that of the traditional algorithms.