To measure multi-waveband laser atmospheric transmittance, a novel ISP02-style infrared solar radiometer was developed. The hardware composition, optical system design, and workflow of the equipment were described in detail. The equipment was calibrated and the performance indicators achieved by the equipment were given. Atmospheric transmittances of 1.064, 1.315, and 1.54 m were measured using the established method based on direct solar radiation measurements. The actual results and comparison experiments show that the errors are less than 6% compared with the POM02 solar radiometer whose results are obtained by an extrapolation method. The total water vapor measured is less than 7% compared with the POM02. Compared to the extrapolated 3.78 m transmittance, the error is less than 5%. The 1.315 m transmittance is less than 2% compared to that of laser atmospheric transmission assessment software in autumn and winter. The results of the equipment are reliable and stable, providing an effective means to obtain multi-waveband laser atmospheric transmittance at the same time.李建玉|lijianyu@aiofm.ac.cn

To visually detect a target ball for a laser tracker in a complex scene, a method of target ball detection improved by deep learning was proposed. Firstly, the characteristics of the target ball, its application environment, and its effect in tracking recovery were analyzed. Subsequently, Hypernet and shallow high-resolution features were adopted, New feature maps and an optimized region proposal were added to the original network, improving the network sensitivity to enable the detection of multi-scale and small targets. Hard example mining with strong background interference was used to reduce the ratio of error recognition, which resulted from similar objects. Finally, the dataset was established and a comparative experiment was carried out. The experiment results show that the improved method proposed in this study and hard example mining with strong background interference can increase the correct recognition rate obtained by Faster R-CNN, yielding a value of 90.11% in the test and meeting the tracking recovery requirement.

To accurately measure the buckling deformation of high-strength steel sheets under high-temperature welding conditions, a visual measurement technique for the full-field dynamic welding deformation of high-strength steel sheets was proposed based on digital image processing technology, the stereo vision principle, and the digital image correlation method. This process was applied to a welding image based on the study of the high-precision matching method of a plaque image. For the discoloration or shedding of the speckle texture caused by high-temperature welding, high-temperature paint and high-temperature glue were mixed and coated to obtain a low-cost, high-stability speckle texture. The measurement of the back side of the thin plate and the addition of a filter set in front of the camera lens suppressed the interference of glare, sparks, and smoke during the soldering process. Gaussian smooth digital image processing technology was used to find the optimal filtering parameters and apply them to the welding image, which improved the image matching success rate and matching accuracy under the premise of ensuring measurement accuracy. The experimental results show that the proposed method can effectively reduce the mean error, root mean square error, and Z-direction static mean error of the welded image. The maximum value of the mean error is 78.6% and the mean square error is 47.7%. The Z-direction static mean error was 37.9%. The method and system of this study can meet the full-field strain measurement requirements of the buckling deformation of thin plates under high-temperature welding conditions. The speckle texture is stable and the speckle image matching success rate is high, demonstrating adequate stability. This technique is an effective way to detect the whole field dynamic welding deformation of thin plates.

To solve increasingly serious food safety problems, especially the rapid detection of food-borne viruses, an E. coli sensor based on the surface-modified graphene enlarged-cone Mach-Zehnder interference structure was proposed in this study. First, a 4 cm solid photonic crystal fiber was intercepted and coarsely fused with two single-mode fibers to form an interference structure based on the Mach-Zehnder principle. Next, surface-modified graphene sensitive material was prepared. The modified graphene, coated on the surface of the solid photonic crystal fiber, allowed the sensor to have a higher sensitivity to E. coli solutions. Finally, the above sensor is placed in a water tank to detect the concentration of the E. coli. The experimental results show that the interference spectrum of the sensor has a significant blue shift with a sensitivity of 3.43 pm/(cfu·mL-1) in the range of 50-600 cfu/mL of the E. coli solution. The linearity between the bacterial concentration and wavelength shift is 0.956 49, detection limit is 67.18 cfu/ml, and response time is 15 s. The sensor has low cost, small volume, fast response time, and is suitable for the rapid detection of low concentration E. coli.

The sun is the main source of Earth's energy. As such solar activity and changes have a great influence on the Earth. To meet the solar observation and research needs of astronomers, a new type of Lyman alpha and visible dual-band internal occulting coronagraph was designed to meet the needs of high-resolution imaging observation of the corona at 121.6 nm and 700 nm for space. Firstly, the parameters of the instrument were determined according to the radiation characteristics of the solar disc and corona in the 121.6 nm and 700 nm bands and the initial structure of the coronagraph was given. The initial structure was optimized by an evaluation function. The imaging performance of the coronagraph optics was calculated and the influence of various optical elements on stray light was analyzed. The main optical components and mechanical structure affecting stray light were determined and the position and surface treatment method of a Lyot stop were designed. The coating film structure of the optical components and optical chromoscope of the coronagraph were designed to achieve simultaneous imaging in the 121.6 nm and 700 nm bands. The design and analysis results show that the field of view (FOV) of the coronagraph covers a 1.1 (Rs=0.267 degrees) to 2.5 solar radius, the effective focal length of the system is 945 mm, the spot radius root mean square (RMS) of the entire field of view is less than 10 um, and the angular resolution is better than 4.8 arcsec. The stray light suppression level of the system is better than 10-6 at 1.1 Rs and 10-8 at 2.5 Rs.

Surface damages of optical materials are inevitably introduced during machining processes. These damages heavily degrade the working performance and lifetime of optical materials-based production. Compared to the machining process of grinding, the art of polishing smoothly removes the damage layer of the sample; however, it is often difficult to ensure the complete removal of the damage layer. Therefore, to determinethe sample's surface quality during polishing periods is very necessary despite difficulties in the measurement technique.Here, we analyzed the typical surface structure of a polished sample and studied the measurement principle of the quasi-Brewster angle method. Simulations show the advantages of this approach to characterize the surface quality of a polished optical plate. As a demonstration, Nd∶GGG optical plates were polished with different processing time and tested carefully with a spectroscopic ellipsometer.Compared with the surface topography measured by white light interferometry, the quasi-Brewster angle offset and the slope of the phase-angle change curve accurately reflect the changes of surface quality during polishing, indicating that the quasi-Brewster angle method has a high sensitivity to the surface roughness and subsurface damages of the sample. This also implies the possibility of the proposed method in characterizing the changes of the surface quality under polishing progress. Finally, the future development trend of subsurface damage detection technology for polishing process.

To realize high-precision and high-efficiency measurements of the geometric dimensions of inertial confinement fusion (ICF) capsules, an X-ray digital imaging system applied to characterize the geometric dimensions of capsules was designed and developed. First, the X-ray direct-projection imaging and X-ray lens coupled microimaging were theoretically analyzed and a technical approach based on X-ray lens coupled microimaging was chosen due to its small size and the weak-absorption contrast of ICF capsules. Then, the key factors affecting the imaging resolution, image contrast, and measurement efficiency of the system were analyzed. A technical scheme of low geometric magnification imaging, a low voltage X-ray source, a small focus, and high power and high-resolution charge-coupled device (CCD) detection was determined. Utilizing this scheme, the phase-contrast effect and penumbra error can be suppressed effectively and the problems of serious edge expansion and large measurement error, which exist in X-ray digital imaging equipment used to measure the capsules, can be solved. Finally, the system performances were tested and the experimental results illustrate that the system has an excellent imaging contrast, high efficiency, and high resolution better than 0.5 μm. The measurement uncertainty of the geometric dimensions of ICF capsules is approximately 0.9 μm (k=2), meeting the high-precision and high-efficiency measurement needs of ICF capsules.

Current theoretical polarization imaging models are based mostly on the assumption of ideal polarizers whose extinction ratio is significantly large and whose main direction is already known. However, the actual non-ideality of polarizers will have a significant impact on the accuracy of polarization imaging measurements. A correction model of visible light polarization imaging considering the non-ideality of polarizers was investigated to decrease this impact. Based on the Stokes vector polarization theory, a correction model of visible light polarization imaging considering the non-ideality of the polarizer was proposed by analyzing the change of the polarization state of the incident light caused by the actual polarizer. The corresponding calculation formulas for the degree of linear polarization (DoLP) and angle of polarization (AoP) were given. An experiment measuring the DoLP of the incident linearly polarized light via a time-division polarization imaging system was conducted. When the extinction ratio of the polarizer is 100∶1, the average relative error of the DoLP based on an ideal model is 5.53%, while the average relative error of the DoLP based on the correction model is reduced to 3.62%. Results show that the proposed correction model can effectively improve the accuracy of polarization information measurements. Using the proposed correction model can achieve a high DoLP measurement accuracy when the measurement system is constructed by a low extinction ratio polarizer. Furthermore, a larger field of view can be obtained at a lower cost.

Coupling moments, nonlinear friction, and unmodeled dynamics between inner and outer gimbals are the main factors that influence the precise control of a double gimbal control moment gyro (DGCMG). To improve the interference suppression ability and ensure the output angular speed accuracy of gimbal systems, this study presents a composite decoupling control method based on a nonlinear cascade extended state observer (NCESO) and sliding mode control. All disturbances in a gimbal system are regarded as the lumped disturbance, which is estimated by the designed NCESO. By combining a sliding mode controller, these disturbances can be eliminated from the output channel of the system. Finally, simulations and experiments of the control method proposed in this study and a linear cascade extended state observer (LCESO) combined with state feedback were carried out. The simulation and experimental results show that the proposed method has better interference suppression performance and dynamic response performance. The angular speed fluctuation of the inner gimbal is reduced from 0.5 (°)/s to 0.2 (°)/s and the angular speed fluctuation of the outer gimbal is reduced from 0.45 (°)/s to 0.15 (°)/s. Moreover, the speed tracking error is reduced from 1.8 (°)/s to 1.2 (°)/s and the phase delay is reduced from 8° to 1.3° when tracking the sinusoidal reference signal.

A cascade sliding mode control method based on a variable gain reaching law was proposed to improve the target acquisition and tracking performance of a photoelectric tracking system. First, to reduce sliding mode chattering and improve the dynamic response speed, a novel sliding mode reaching law with variable gains was designed. The variable gains were designed based on the inverse sine and exponential functions. Then, based on the novel reaching law with variable gain, the speed and position sliding mode controllers of the system were presented. The speed and position sliding mode controllers include cascade sliding mode control, which was designed to enhance the dynamic and robust performance of the system and further improve the system target acquisition and tracking accuracy. Finally, a comparison was made between the control performance of a traditional cascade PI control and the proposed cascade sliding mode control, based on the azimuth axis of a photoelectric tracking system. The experimental results show that compared to traditional cascade PI control, cascade sliding mode control can reduce the acquisition time by 32% when detecting a target with a speed of 1 (°)/s. Moreover, it can reduce the RMS of the tracking error by 31% when tracking the position sine signal with a maximum speed of 4 (°)/s and a maximum acceleration of 2 (°)/s2. These results indicate that cascade sliding mode control can improve the control performance of the tracking system.

To solve the balancing problems of high speed and high precision, a drive method based on conduction angle adjustment was proposed. The electromechanical coupling model, drive method, and motion mechanism of a V-shaped double-foot-worm motor were studied. Firstly, the electromechanical coupling model of a V-shaped biped inchworm motor was established according to the designed structure of the V-shaped biped inchworm motor. Then, the driving method based on the conduction angle adjustment was determined. Then, the motion mechanism of the V-shaped biped inchworm motor was analyzed. Finally, an experimental platform was built for experimental verification. The experimental results adequately agree with the theoretical derivation within a certain allowable error range, verifying the rationality and feasibility of the driving method. With the introduction of the conduction angle, the minimum step distance is reduced from 500 nm to 300 nm, a reduction of approximately 34%. By adjusting the frequency f and the conduction angle concurrently, the motor achieves a driving speed of 0.5 mm/s and a 330 nm positioning accuracy at the same time. By introducing a variable conduction angle, the motor step distance is successfully realized and the driving speed is adjusted independently.

During the process of fixed abrasive (FA) lapping, the micro-fracture of abrasives is the main approach to realize the self-conditioning process of the FA pad; the lapping load is one of the key parameters that affects it.The FA pads were prepared by using single diamond (SD) and agglomerated diamond (AD) abrasives. Marathon lapping tests were carried out, using quartz glass as the workpiece under a lapping load of 15 kPa. The material removal rate (MRR) and machining stability of the two FA pads were comparedtogether. Four kinds of AD abrasives with different concentrations of the ceramic binder were prepared and chosen as the abrasives to prepare the corresponding fixed AD abrasive (FADA)pads; their self-conditioning performance was evaluated under different lapping loads. The surface roughness(Ra) and morphology of the lapped quartz glass were observed. The results show that: the Ra of quartz glass lapped using the FADA pad is lower than of that lapped using fixed SD abrasive(FSDA) pad and that the former has a greater machining stability. Under a higher lapping load of 15-21 kPa, the FADA pad with the second highest ceramic binder concentration has the highest lapping efficiency; its MRR is 8.94-12.43 μm/min and Ra is approximately 60 nm. Under a lower load of 3.5-7 kPa, the FADA pad with the second lowest ceramic binder concentration is the most stable; its MRR is 2.67-3.12 μm/min; the Ra is approximately 40 nm. AD abrasive with a high ceramic binder concentration is more suitable as the abrasive for a FA pad in high-load applications, while that with a low concentration is more suitable for low-load applications.

HF-based etching is used in this paper to remove laser damage precursors to improve the laser-damage resistance of fused silica optics. Several CeO2-polished samples were post-treated by various HF-based etchants and the responses of etching rate, surface cleanliness, roughness, transmission, and laser damage performance of optics to the etching state, etchant composition, and etched depth were investigated. No solid evidence shows that adding acoustic power ~0.6W/cm2 to 6%wt. HF contributes to either the etching rate or the damage performance of samples. The appearance of etching-induced deposits, which deteriorate both the surface roughness and transmission significantly, is highly dependent on the composition and concentration of the etchant. The damage testing results reveal that 6%wt. and 12%wt. HF can efficiently improve LIDT optics ~1.9 times the value of an un-etched sample with a (5±1) μm material removal value. Moreover, the damage performance of the optics is found to be somewhat isolated with surface root mean square (RMS)roughness and light transmissionandthat surfaces with a high laser-induced damage threshold (LIDT>20 J/cm2@3ns) are usually smooth with an RMS roughness<2 nm. Therefore, the wet chemical etching condition that yields surfaces with the desired damage performance can be obtained and the control criterion of surface roughness can be determined.

Piezo-actuated Fast Tool Servo(FTS) technology is apromising method for processing complex surfaces such as optical free-form surfaces. To break through the stroke limitation of the piezo-actuated FTS mechanisms, a piezo-actuated long-stroke FTS mechanism was designed with a two-stage lever amplifying mechanism in the present paper; this effectively improved the long-stroke output performances of FTS mechanisms and eliminated parasitic displacement. The dynamic model of the mechanism was discussed based on the pseudo-rigid body model and Lagrangian principle. The mechanism parameters were optimized by synthesizing the stroke as well as the natural frequency performance; Finite Element Analysis(FEA) and experimental verification were also carried out. The errors of equivalent stiffness and the natural frequency of the mechanism between the theoretical model and FEA are 6.4% and 1.6%, respectively, validating the theoretical model. The experimental results show that the designed FTS mechanism can achieve a 100-μm output stroke and 730 Hz natural frequency, which agrees well with the analysis and simulations. The closed-loop tracking experiments of the designed system are also given, demonstrating the design's adequatetracking performance.

A flexible micro-gripper system was developed to realize the precision clamping of micro-or Nano-sized objects. The design, kinematics, dynamics, and control methods of the flexible micro-gripper were studied. First, a flexible micro-gripper was designed by using a flexure hinge design method, and a Pseudo Rigid Body (PRB) model was established by using a PRB method. Moreover, the kinematic models of the system, for example, the mechanical amplification ratio and input stiffness, were established using the PRB model. Furthermore, the dynamic equation of the system was established by the Lagrange method, and the natural vibration frequencies of the system were obtained. Finally, the ANSYS finite element method was used to simulate and verify the model of the system. Additionally, a PID control algorithm was used to control the micro-gripping system. The experimental results show that the tracking control error is 2.4%, and the magnification ratio is 9.12. The system performs the micro-clamping function on the micro-or Nano-scale, and its total precision can reach the micron level or Nano-level. The micro-gripper meets the design requirements in several studies.

In order to improve the fabricating efficiency ofn micro-and nano-structures by laser-induced backward transfer technology, a three-beams laser-interference-induced backward transfer (LIIBT) method was established to fabricate micro-nano structure arrays, which was based on laser interference and laser-induced backward transfer technology. In this article, Indium Tin Oxide (ITO) glass was used as a substrate to receive an Au nanoparticles jet from Au thin film. The LIIBT was carried out by three-beams laser interference. The Scanning Electron Microscopy(SEM) images indicate that the a perfect 5-μm period structure could be fabricated under the laser fluence of 25 mJ/cm2 for the Au 50 nm thin film 50 nm. Meanwhile, uniform distribution of the Au nanoparticles could be observed on the surface of the period structures, and the over 80% of Au nanoparticles over 80% were smaller than 100 nm. The Energy-Dispersive X-ray spectroscopy (EDX) analysis shows that the microscale period structures were composed of a lot of Inand Au elements, which prove validates the fabrication of theisemicroscale period structures due to the laser pulse ablation on the ITO layer, and then, Au nanoparticles were deposited on it, simultaneously. Finally, the SERS was studied by the Raman spectrum for the Au nanostructures after coating them with 1.0×10-5, 1.0×10-7, and 1.0×10-9 M rhodamine 6G. The results show that these peaks are located at 612 cm-1, 773 cm-1, 1 190 cm-1, 1 319 cm-1, and 1 511 cm-1, which indicate that the Au nanostructures had obvious SERS effect for the trace of the R6G. The LIIBT will be used to improve the efficiency of the laser-induced backward transfer technique on the micro and nanostructure. This technique, which will have the potential applications for the ultra-sensitive detection, optoelectronic devices, and microfluidic flow control systems.

To solve the problems of large-scale and high-precision measurement as well as geometrical parameter characterization of complex microstructures, a composite micro-and nano-measuring instrument was developed based on multi-probe sensing and precise positioning platform multiplexing technology. To realize high-speed scans of a large area and fine measurement of a small area, a White Light Interference(WLI) microscope and an Atomic Force Microscope (AFM) were integrated into the bridge structure of the instrument.A macro/micro-driving platform was designed and tested to meet the positioning range and accuracy requirements. The instrument could work on in a contact or noncontact mode.Furthermore, different geometrical parameters could be characterized by the instrument. The instrument was calibrated using a nano dimensional standard to ensure accuracy and traceability. The WLI resolution of the instrument can reach 500 nm in the lateral direction and 1 nm in the vertical direction. The AFM resolution is better than 1 nm in both the lateral and vertical directions. Lastly, a micro ball was measured and its surface geometrical characteristics were obtained by large-range imaging and small-range fine scanning. Thus, the instrumental capability was verified.

The Muscle Computer Interface (MCI) system is one of the areas of active interest in virtual reality and human-computer interaction research. The main problem associated with the MCI was the EMG signal classification, to facilitate the effective combination of an MCI system with deep learning methods. Surface EMG signals include high-density transient signals and sparse multi-channel signals. The former was analogous to an image that can be recognized by a CNN network. The latter was studied in this investigationin which an MCI system with an MYO armband was realized. Sparse multi-channel EMG signals were long-term sequence signals with a high correlation between time and time that can be recognized by an RNN network. We proposd a combined RNN network architecture to recognize gestures with multi-stream feature sequence signals that were obtained by extending the original signals in the time-domain and time-frequency domain. The accuracy of the net is 90.78%. We perform cross-validation without a self-training set using 35 individuals, and the accuracy of the classification is 78.01%. The accuracy of real-time gesture recognition in the MCI system is 82.09%, and the action can be recognized within 61.7 milliseconds. We establish that the combined RNN nets can classify different gestures using EMG signals, and the MCI system performs well in generalization and real-time recognition.

Traditional graph embedding methods often use single graph structures for feature extraction (FE).However, these methods cannot effectively represent the complex intrinsic structuresof high-dimensional data. To address this problem, a Semi-Supervised Multi-Graph Embedding (SSMGE) algorithm was proposed for FE of Hyper Spectral Images (HSIs). First, the SSMGE method constructed one of each of intra-and inter-class hypergraphs and intra-and inter-class graphs through intra-and inter-class neighbors of labeled samples.In addition, it constructs an unsupervised intrinsic hypergraph and a penalty hypergraph using unlabeled samples. The fusion of graphs and hypergraphs can effectively characterize the complex relationships in high-dimensional data. The SSMGE method not only effectively reveals the intrinsic structure of an HSI by exploring the collaboration of graphs and hypergraphs but also enhances the discriminative ability of extracted features in low-dimensional embedding space.This enables improved classification performance of HSI data. Experimental results on the PaviaU and Urban hyperspectral datasets show that the overall accuracies of the proposed method reached 85.92% and 79.74%, respectively. The SSMGE method can significantly improve classification performance as compared with some state-of-the-art FE methods.

Improving the quality of Cone-Beam Computed Tomography (CBCT) reconstructed images under complicated noise conditions was critical for CBCT systems. In this study, an image reconstruction method of CBCT based on a hybrid Poisson-Gaussian maximum likelihood function was proposed. First, a CBCT image reconstruction model suitable for describing a mixed Poisson-Gaussian noise environment was studied. The model contained a fidelity term based on a mixed Poisson-Gaussian maximum likelihood function and a constraint term based on a three-dimensional total variation regularization method. The fidelity term was used to constrain the reconstruction result to match the observed value as closely as possible in the mixed noise model, where as the constraint term was used for noise removal and preserving the edge and detail information as effectively as possible. The proposed model was further solved using the separable approximation and extended Lagrangian methods. Finally, the effectiveness of the algorithm was verified using both simulation and real data. The results indicate that the proposed method exhibited a maximum improvement of 2.1 dB as compared to other methods evaluated usingsimulated data. In visual comparisons, the proposed method demonstrated the best denoising performance. We can thus conclude that the proposed method is effective forCBCT reconstruction under low dose conditions.

The combination of Faster-than-Nyquist (FTN) theory with modulation technology can effectively improve the spectral efficiency of atmospheric communication systems. In this paper, the FTN theory was introduced to atmospheric optical communications, and an FTN optical transmission system suitable for log-normal turbulence channel was proposed. Moreover, the equation for the Bit Error Rate (BER) of the proposed system was derived and further analyzed using the MonteCarlo method. The results indicate that the FTN scheme can considerably improve the spectral efficiency of atmospheric optical transmission systems. For example, when the signal-to-noise ratio is 18 dB, and the S.I. is 0.4, the spectral efficiency can reach 1.7 Baud/Hz, whereas it is only 1.56 Baud/Hz without the FTN scheme. Regarding the effects of atmospheric turbulence on system BER performance, when the S.I. is 0.4 and BER is 3.8×10-3, the error performance decreases by approximately 1 dB. Compared with the significant improvement in spectral efficiency, the decrease in BER performance is worthwhile.Therefore, FTN technology can be introduced into atmospheric optical transmission systems to improve spectral efficiency.

In medical multi-atlas registration, to improve the limitations of low efficiency and poor accuracy caused by large initial position differences, complex shapes, and local residual differences, a fine registration algorithm that used Iterative Closest Point (ICP) was proposed based on the bidirectional distance ratio. The proposed algorithm was based on the coarse registration method followed by fine registration, where the former was processed by principal registration analysis.In the fine registration algorithm, the K-Dimensional tree was initially used to perform a nearest-neighbor search to improve the searching speed of corresponding point pairs. A bidirectional matching method was then proposed for each point, and the bidirectional distance and ratio were calculated. To further improve the accuracy of the registration, an exponential function was introduced to determine the probability that the point pair belongs to the correct match. The final transformation matrix was then obtained using Singular Value Decomposition. To evaluate the feasibility and effectiveness of the algorithm, experiments were designed using Stanford point cloud data and two sets of CT cardiac point cloud data registration. The results show that the average error during registration is reduced by 21% using this method compared to the classical ICP algorithm, which is 13% lower than the error obtained using the trimmed ICP (TrICP)algorithm. In the cardiac point cloud data registration experiment, this method is accelerated to 1.77 s compared to the TrICP algorithm, which has a value of 15.5 s.Therefore, the proposed method has high efficiency, accuracy, and stability in solving the registration problem associated with the three-dimensional cardiac point cloud data.

The SIFT descriptor has poor real-time performance and conventional binary descriptors are poorly robust to scale, rotation, and viewpoint changes. They can be improved by optimizing the sampling pattern and adding gray-value differential invariant comparisons. This study proposed a binary descriptor with a higher robustness. First, a sampling pattern with scale association and number marking was presented. Then, each sampling point of the sampling pattern was rotated to a specific position to ensure that the descriptor was invariant to scale and rotation. Subsequently, the influence of sampling pairs on the descriptor was analyzed, and 128 sampling pairs after machine learning were chosen. Finally, intensity comparison and gradient absolute value comparison were selected to build the descriptor. The image keypoint detection was based on the difference of Gaussians method. Experiment results show that the proposed descriptor is 84% and 67% faster than the SIFT descriptor in descriptor construction and descriptor matching, respectively. Its accuracy is 3% to 5% higher than that of the conventional binary descriptor in image matching with view change, and the recall rate is more than 30%. The descriptor presented in this study is suitable for time-critical image matching.

To solve the problem of detecting small infrared targets ininfrared optical systems under complex backgrounds, an information processing model based on visual feature integration was established, and a dim small target detection method based on visual feature integration was proposed. First, the difference of Gaussians model of the receptive fields of retinal ganglion cells was used for processing the primary information of infrared images and initially detecting dim small targets. Then, the features containing the dim small targets were extracted by spatial and frequency channels. In the spatial channel, the second-order differential Hessian matrix was constructed with image information, and the local extremum was determined by calculating the basis truth and determinantto extract the structural component features. In the frequency domain channel, wavelets were used for decomposing the image frequency domain to extract the features of high-frequency components containing dim small targets. Finally, the spatial channel and frequency channel were integrated to extract the dim small targets from complex backgrounds. Experimental results indicate that when the false alarm rate is 10-3, the average detection probability is 95.17%. The proposed method can basically meet the requirements of stability, reliability, and high accuracy.

Finger vein recognition algorithms are usually based on grayscale images of vein distributions. However, because of the limitations of image acquisition devices, the uncertainties in the illumination intensity and the complexity of the tissue around the finger vessels (among other factors)even after image processing cause the resulting grayscale images to possess irregular shadows and non-venous characteristics, which may reduce the accuracy of the recognition results. Therefore, this paper proposed a new finger vein identification algorithm based on binary images to minimize the interference of non-venous factors in the identification process. First, the features from accelerated segment test algorithm was used to extract the pixel points at the edges of vein textures as feature points, and then the feature vectors were constructed. Further, to improve the matching precision, a new matching algorithm based on circular neighborhoods was proposed. The weighted matching distance was used to describe the degree of similarity between images. The average recognition rate of the proposed method when applied to the finger vein database published by Shandong University is 0.993, and the equal error rate is 0.019 6.These results demonstrate the effectiveness of the algorithm and provide a new basis for the design of vein recognition algorithms.