Infrared imaging is an important method to detect the thermal wake target on sea surface of underwater vehicles. By utilizing the modeling and simulation method, based on the sea surface temperature distribution characteristics of the wake region, combined with the space coordinate transformation and projection mapping method, the infrared imaging simulation process of the thermal wake target was realized. Under the same detection conditions, the distribution of the radiant energy of the thermal wake with the band, and the variation of the shape of the hot wake in the image plane and the brightness changing law of the infrared radiation under different azimuths and heights were obtained. The effects of noise on the thermal wake under typical working conditions and the contrast and sharpness of the sea surface, as well as the difficulty of detecting and identifying the thermal wake target were demonstrated.The simulation results show that under the same detection conditions, the radiant energy of the thermal wake in the 8 μm~12 μm band is much larger than the 3 μm~5 μm band;when the length of the detection path is constant, with the angle of the detection zenith increasing, the observed brightness of the infrared radiation of the thermal wake decreases, and the area occupied by the thermal wake in the image plane gradually decreases.

During the movement of the monocular robot, the camera pose cannot be recovered due to the extremely short base line between the key frames.Aiming at this problem, an iterative optimization algorithm of essential matrix was proposed to quickly decompose the relative pose relationship of robot by using the known specific reference objects. In this method,firstly, the estimated value of essential matrix is obtained using a specific reference with clear three-dimensional information. Then, the essential matrix is iteratively optimized with the matched feature point groups of adjacent key frames and the estimated value of essential matrix. Finally, the estimation of translation vector is obtained by the constraint relation of essential matrix, and the estimation of rotation matrix is obtained based on Lie group. The unique solution is obtained by further optimization. The problem that the relative pose of the camera cannot be restored due to the too short base line is avoided in this method. The experimental results show that the relative pose relationship of the robot can be quickly solved and optimized in this new method, and the absolute error of relative displacement estimation is less than 0.03 m, which is superior to the traditional method.

Heat generated by the absorbed pump light always results in a non-uniform distribution of temperature in the laser crystal, which leads to a non-uniform distribution of refractive index in the crystal. Therefore, a phase modulation can be produced when the oscillating beam passes through the crystal, thus the oscillating light field can be redistributed and deviated from the ideal mode distribution. Because of the filter function of the resonator, the resonator filters out the beam profile that does not match the ideal resonator mode and causes loss. By calculating the thermal loss of the fundamental mode and the conversion ratio of the low-order mode to the high-order mode, the temperature field distribution in the crystal was simulated with MATLAB, the energy interchange between the modes in the resonator was studied, which could provide a theoretical basis for suppressing the high-order mode and reducing mode loss in solid state laser.

An dispersion-managed thulium(Tm)-doped fiber laser was designed and demonstrated. By adjusting the pump power and the intracavity polarization state, the stable stretched-pulse can be firstly achieved. The center wavelength of stretched-pulse is 1 939.4 nm and the pulse width is 482 fs. The maximum output power of stretched-pulse is 15 mW, and the corresponding single pulse energy is 0.52 nJ. When increasing the pump power to 645 mW, the noise-like pulse can be realized with appropriately adjusting the polarization controller, and the center wavelength is 1 940.1 nm. The mode-locked pulse achieved has femtosecond(fs)-order spike and picosecond(ps)-order pedestal. The maximum output power of noise-like pulse is 20.4 mW and the corresponding single pulse energy is 0.7 nJ. Compared to traditional soliton, the mode-locked pulses achieved under dispersion management have higher pulse energy. In addition, the designed Tm-doped fiber laser can be used as the seed source of the main oscillation power amplification and chirped-pulse amplification structure for improving the pulse energy and further expanding the practical applications of the 2 μm high energy fiber laser.

Aiming at the application requirements of spectral observers for test site calibration of satellite remote sensors, we designed a spectral radiation acquisition circuit system with visible near-infrared (0.4 μm~1.0 μm) ,and conducted the analysis on the application and data combining with the spectral radiation acquisition module. We mainly introduced the selection and driving requirements of the linear array detectors , the design based on STM32 microcontroller unit (MCU) was completed. According to the radiation observation requirements，we processed the analog output signal of the linear sensor by instrumentation amplifier in the signal processing section，and used the A/D converter in STM32 for collection.By analyzing the results of digital signal output，the data collection of the radiation measurement system has achieved the accuracy of 1LSB (less than 0.8 mV). Furthermore, we applied the detector circuits system to the spectral radiation acquisition module, and carried out the outdoor application and data analysis . Results show that the deviation of the comparison result of spectral irradiance is less than 5%. The comparison results of the spectral irradiance prove that the whole system can be directly applied to the product and meet the reliability acquisition of field radiation measurement.

In order to test and analyze the fluorescence life of up-conversion nanoparticles (UCNPs), a system was designed for high-speed acquisition and measurement of fluorescence signal data of UCNPs. The system programmed the virtual original fluorescent signals through MATLAB, completed the filtering and data fitting, embeded the data program script into LabVIEW for calculating and processing, and displayed the fluorescent signal waveform, fitting data waveform, fluorescence number, fluorescence life and extreme of the life on the host computer. The research results show that the system software simulation and numerical calculation simulation have achieved basically consistent results, the error between the two is less than 3%, with good stability and accuracy. This research work has great guiding value for the development of UCNPs fluorescence lifetime measurement system.

The focusing of optical measurement system is one of the important steps in the process of obtaining high-quality images. When the EO measurement system during target tracking and imaging is unable to timely focus, the obtained fuzzy images can adversely affect the attitude, tracking, recognition and radiation characteristics measurement of the target. Aiming at this problem, a new phase-image hybrid passive focusing structure was designed on the basis of the summary of current focus detection methods in ground-based field.In addition, a focusing detection prototype was made . Through the compared experiments on the prototype between the proposed method and the traditional image focusing detection experiment, the phase focusing accuracy of the system is better than 0.08 mm, and the compound focusing accuracy is better than 0.02 mm, which can effectively improve the focusing efficiency of the system.

In order to obtain high imaging quality of numerical aperture(NA)1.35 lithographic projection optical system, except for controlling the wavefront aberration,the polarization aberration also should be optimized during design process. The polarization aberration of NA1.35 lithographic projection optical system was expressed by the use of Jones pupils and physical pupils, and the magnitude of polarization aberration was analyzed with bi-attenuation and retardation. According to the max incident angles of different optical surfaces, corresponding coatings of each optical surface were designed to optimize polarization aberration of optical system. After optimizing coatings,the diattenuation and retardation of NA1.35 lithographic projection optical system decrease to 0.021 8 and 0.057 2 rad separately compared with regular coatings，it means that the polarization aberration of optical system decreases. Using lithographic simulation software Prolith to simulate the lithography performance with regular coatings and optimized coatings separately, the result shows that the image contrast with optimized coatings increases 4.4% ,proving the effectiveness of the method for optimizing polarization aberration of NA1.35 lithographic projection optical system.

Aiming at the problem of large number of image feature mismatches, an oriented FAST and rotated BRIEF (ORB) image feature point matching algorithm based on sparse optical flow method was proposed. Firstly, the feature points were violently matched to obtain the initial matching point set. Then the sparse optical flow method was used to calculate the feature point motion vector, and the two-dimensional coordinate position of the feature point in the image to be matched was estimated, and the feature points far from the estimated position were eliminated. Finally, the random sampling consistency algorithm was used to perform geometric verification to further optimize the matching result, so as to eliminate the effect of mismatching. The experimental results show that compared with the ORB operator, SIFT operator and SURF operator, the accuracy of this algorithm is increased by 21.6% on average, and the accuracy of random sample consensus FAST and rotated BRIEF (RANSAC-ORB) algorithm is increased by 2% on average; moreover, the algorithm has good versatility for image illumination transformation, perspective transformation, fuzzy transformation, rotation and scaling transformations and illumination variations.

The reconstruction algorithm of single-frame lensless microscopic imaging was investigated. The experimental setup of the lensless microscopic imaging system and the two reconstruction algorithms of L-R(Lucy-Richardson) and ASM (angle spectrum method) were introduced. Comparing the resolution of the USAF resolution plate image reconstructed by the two algorithms, according to the Rayleigh criterion，the amplitude image resolution acquired by the SSM is up to 3.10 μm and the calculation time is the least (0.9 s), which proves ASM is the best reconstruction algorithm of single-frame lensless microscope imaging. In addition, the cell imaging experiments were performed using a lensless microscopic imaging system combined with ASM for image reconstruction. Results show that the lensless imaging field of view is 4.4 times than that of a 5× microscope, and its resolution is between 5× and 10× optical microscopes. The statistical advantage is obvious, and it has broad application prospects in the biomedical field.

Aiming at the problems of contrast reduction, saturation reduction and color migration of images collected in foggy environment, an image defogging algorithm based on full convolution neural network is put forward. First, the proposed three scales convolution neural network is used to study the fog of the mapping relationship between foggy image and medium transmission map, gradually produce the refine medium transmission map；secondly, the foggy image is used as a guide map to refine the forecasting medium transmission map, so as to make the edge information of the image more smooth; finally, the value of atmospheric light is estimated according to the dark channel prior theory, and the fog-free image is recovered by the atmospheric scattering model. The fog-free image obtained by this method not only causes no loss of useful information in the image, but also restores the color of the image naturally. Experimental results show that the algorithm proposed is superior to other comparison algorithms in both natural fog images and fog images produced by Middlebury Stereo Datasets, and the restored images have better contrast and clarity.

Aiming at the false alarm or missing problems due to the insignificant and obscured targets in sequence images , an indirect localization method for building targets based on automatic selection of auxiliary targets was proposed to improve the stability of target tracking. In this method ,firstly, the region with stable shape and significant gray level in the infrared image was automatically selected as the auxiliary target. After locating the building target, its relative positional relationship with the target was extracted, and then,the relative position was used to indirectly locate the target. Finally, the target was located by fusing the result of direct target recognition and the result of indirect target location,which could update the auxiliary target timely to ensure that it is always in the field of view. The experimental result shows that the false alarm or missing problems can be solved by indirect location,and it can improve the accuracy and robustness of the algorithm during the process of target location.

A steam humidity measurement model and a wet steam parameter inversion optimization model were established based on the laser back-axis angular scattering method in order to measure the steam trubine humidity at the final stage accurately. According to the optimization models, the particle swarm optimization (PSO)algorithm was used to perform multiple inversion optimizations on the simulation data which was added the Gaussian white noise and the experimental data of the simulated cylinder. The obtained inversion results were compared with the artificial fish-swarm algorithm and the traditional uniform search method as well. When the PSO algorithm was used, the inversion results of r0.5,K and N were 0.05, 0.66 and 0.51% respectively, the inversion time was 306.41 s. When the artificial fish-swarm algorithm was used, the inversion results of r0.5,K and N were 2.96, 19.98 and 4.68% respectively, the inversion time was 411.05 s. When the uniform search algorithm was used, the inversion results of r0.5,K and N were 5.00, 27.14 and 7.95% respectively, the inversion time was 246.42 s. The results show that the PSO algorithm can overcome the shortcomings of both artificial fish-swarm algorithm and uniform search method, which can obtain high precision, stable and reliable inversion results in a short time. It provides more accurate data for wet steam parameter measurement and theoretical basis for other particle parameter measurement inversions.

As the resolution of the computational spatial light modulator has become larger, the computational complexity of the three-dimensional dynamic display of the hologram is also higher and higher, which makes new requirements for the holographic calculation speed. The fast calculation of hologram by slice-based method is realized by means of the graphics processing unit (GPU) parallel computation processing method, which accelerates the calculation of the Fresnel diffraction transform algorithm by taking advantage of the GPU parallel multi-threading and the image two-dimensional Fourier transform in tomography. At the same time, by calling the GPU underlying resources and the stream processing of the program in compute unified device architecture (CUDA), the intermediate delay wait is effectively reduced. The comparison of the calculation speed shows that the calculation speed is greatly improved compared with the calculation on the CPU, and the method based on GPU parallel calculation is about 10 times faster than the method based on CPU calculation.

Aiming at the problem of small field of view of long-range aerial camera, the large-scene ground imaging was realized by utilizing the small-array camera combined with the optical components of the system through the sweeping in the roll direction. The mathematical analysis of the sweeping process was carried out, and the scanning instructions required for sweeping were obtained. Several image motion compensation techniques commonly used at present were introduced, and various compensation methods were compared. The correctness of the mathematical formula was verified by simulation experiments. The simulation experiments show that the scanning command can be used to acquire large scene images on the ground by small array aerial camera.

The uniform light sources are widely used in optic imaging fields ,as well as the development and calibration of remote sensing instruments. For irradiation uniformity, an integrating sphere structure with an annular baffle was proposed. Based on Monte Carlo method, the structure was analyzed and optimized by changing the parameters including light source type, baffle position, slotting size and number of light sources. Results show that the average irradiation uniformity of the Lambert light source is 2.32% higher than the collimated light source. When the baffle is located close to the edge of integrating sphere, the irradiation uniformity is slightly lower than other positions. Considering the radiation energy and illumination uniformity, the open ratio is suggested to be limited in the range of 30%~35%. The radiation power has a linear relationship with the number of light sources,and the nonuniformity fluctuates with the number of light sources from 2.46% to 3.38%. The optimized annular baffle structure is proposed as irradiated uniform light source in practical application scenario.

In the null testing of large-diameter and fast-ratio aspherical mirrors, the incident light rays are sharply turned at a large angle within a short distance, resulting in nonlinear distortion of the test result of the interferometer, and the positional accuracy and error removal efficiency of the computer controlled optical surfacing (CCOS) polishing are seriously affected. In order to correct the image distortion of off-axis aspherical mirror in compensation test, a method for correcting the nonlinear distorted image was proposed. The center position of the distortion was determined by the concentric annular method and the mapping relationship between the tested mirror and the interferogram was established by using ray tracing .In addition,the distortion of a 520 mm×250 mm off-axis paraboloid primary mirror of an optical system was corrected.Results show that the position deviation between the correction result and the surface map of the workpiece is reduced to less than 1 mm, which meets the working requirement of CCOS polishing.

Quality inspection of package is vulnerable of illumination. Aiming at this problem,a defects detection algorithm based on the improved gradient magnitude(GM) similarity was proposed. Based on the existing defects detection method by gradient magnitude similarity, the local binary pattern (LBP) operator was integrated into the proposed algorithm.The algorithm utilized the characteristics of rotation invariance and gray scale invariance of LBP operator, and combines them with the gradient magnitude features of the image to be used in the defects detection of packaging, which improves the algorithm robustness to illumination.Experimental results show that the proposed algorithm has better anti-illumination ability in comparison with traditional methods,and under different illumination conditions, the detection accuracy of the proposed can reach 96.57%. Therefore, the proposed algorithm can be widely used in quality inspection of package and improve the defects detection accuracy.

Non-diffracting moire fringes are widely used in the field of precision measurement, and the accuracy of finding the central positions of the fringes is the key to realize high-precision measurement. By analyzing the characteristics of the non-diffracting moire fringe images, a method to seek the central positions of the fringes was proposed. Firstly, two local concentric circle regions, including the central spots, were extracted through the feature of intensity distribution. Secondly, the local concentric ring detection was performed to determine the initial center set. Thirdly, the initial center set was clustered to determine the initial positions of the two spot centers. Finally, the abnormal points were deleted and the more precise positions of the center of the two central spots were obtained by iterative algorithm, realizing the positioning purpose of the two central spots of the non-diffracting moire fringes. Theoretical and experimental results show that the method can determine the center position of two spots at the same time quickly and accurately, and the error is less than 1 pixel.

The groove curved error in the grating can directly affect the diffraction wavefront quality of grating . Aiming at the problem, we established the mapping relationship between the grating curve error and the diffraction wavefront, built the optical path for measuring the curve error of the grating line, discussed the principle of the active control technology of the grating ruling machine, and used the active control technology to compensate the grating curve error on-line. Finally, we carried out the contrastive experiment to verify the feasibility of the method. The above method can compensate grating curved errors accurately and effectively, and the grating wavefront can be increased from 0.074 λ to 0.038 λ, increased by 48.6%, at the same time it can effectively solve the problem of the curve of the grating surface. The research results can be used to improve the quality of mechanical ruling grating, which has important theoretical and applied values.

In order to avoid a series of traffic accidents caused by improper use of high beam lamps，and to give drivers a better driving experience, an adaptive driving beam (ADB) illumination system was designed, which can actively adjust the distribution of high beam. Based on the theory of non-imaging optics, the illumination optimization design method was adopted and the area was divided reasonably by calculation to control the brightness and extinguishing of corresponding LED, so as to achieve the different types of high beam. Through tracing the light with Monte Carlo simulation method, the simulation result shows that the final result of light distribution fulfills the current only for adaptive driving beam (ADB) light distribution regulation. When all the LEDs work，the maximum illumination value is 111 lx, and the efficiency of the module optical system is 35.7%.

In the Fourier transforms profilometry (FTP), the band-pass filter that extracts the fundamental frequency component has a significant influence on the reconstructed accuracy of the 3D profile. For traditional 2D symmetrical filters such as Hanning window or Gaussian window, their cut-off frequencies are limited by the spatial frequency of projection fringe, namely the distance between the fundamental frequency and the zero frequency. This often leads to the incomplete extraction of fundamental frequency components and the loss of useful information. To solve this problem,we designed a 2D asymmetric filter with the elliptical bottom shape on the basis of 2D Hanning filter. It holds different frequency response and cut-off frequency towards different frequency directions and thus one can extract more complete fundamental frequency component. The simulation experiments of step object and array object show that the reconstructed accuracy of the proposed 2D asymmetric filter improves respectively by 39.5% and 55.6% for these two objects compared with traditional Hanning filter.

Calculation formulas of parameters which are needed for establishing a linear Fresnel concentrator model in SolTrace software were derived by geometrical optical principles, and a modeling method was given by the example. The results show that for the linear Fresnel reflector which is composed of reflectors with 21 rows, 0.38 m width, 4 m length, the compound parabolic concentrator (CPC) with 45° maximum acceptance angle and the vacuum tube with 5.3 m height, the energy flux density on the vacuum collector tube gradually increases and the uniformity becomes better with the increasing of the solar incident angle. When the solar incident angle is greater than 40°, the energy flux density and the uniformity tend to be stable. What’s more, the energy flux density on the vacuum collector tube is bigger and the uniformity is better for the CPC with involutes + cusp reflectors than the CPC with involutes + parabolas. The results have a directive function for the popularization and application of the linear Fresnel concentrator.

In order to investigate whether the radiometric responsivity of the detector in large-scale dynamic range is constant under the illumination of monochromatic light, a laser-based linear measuring device was established, and the measurement uncertainty of the device was evaluated. According to the dual optical path superposition method, the linearity of the photodetector was measured in both the power mode and the irradiance mode. For the power mode, the nonlinear coefficient was measured with the laser power varying over 6 orders of magnitude,while for the irradiance mode, the laser was guided into an integrating sphere to output the optical radiation, and the nonlinear coefficient of the detector under 3 orders of magnitude changes was measured. Results show that when the current of the photodetector is varied from 0.2 nA to 0.2 mA in power mode, the nonlinear fact of the detector is less than ±0.02%. When the current of the detector is changing from 200 nA to 0.2 nA in irradiance mode, the nonlinear factor is also less than ±0.02%. It can be seen that the laser linear device can realize two modes of power and illuminance measurement, and meets the measurement of at least 6 orders of magnitude or more.

In order to obtain the surface features information of the workpiece to be grinded, a method for surface detection of grinding workpiece based on laser scanning was proposed. The robotic arm was used to drive the laser sensor to scan the grinding workpiece placed in the laser measuring plane,thereby obtaining the three-dimensional coordinates information of the workpiece in the laser measuring plane, and obtaining the characteristic boundary point of the workpiece by the height change between two adjacent scanning points. Then,the extreme points and least squares of the x-axis and y-axis coordinates were combined to fit the analytical expression of the workpiece boundary in the laser measurement plane, and the offset and rotation angle of the workpiece relative to the laser measurement plane were further determined. Finally, the surface features information of the workpiece was obtained by using the position and attitude information of the workpiece in the laser measuring plane. The experimental results show that the detection error of the workpiece surface is 0.11 mm and the average detection time is within 1 s, which meets the requirements of workpiece surface features detection.

The hydrocarbons in the drilling fluid can indicate the hydrocarbon content of the formation.The detection of oil and gas concentration in the stratum is of great significance for the identification of true and false oil and gas,especially for the accurate interpretation and evaluation of oil and gas layers.Based on the unique advantages of continuous,rapid and direct detection of samples by laser Raman spectroscopy,we carried out researches on quantitative Raman detection of hydrocarbon concentration in drilling fluid.The laser Raman online detection system was built under laboratory conditions which was used to detect C7-C14 normal paraffin and benzene.The n-octane was selected as the marker of alkanes in water-based drilling fluid and the mathematical models of different characteristic peaks of n-octane in water-based drilling fluid were established based on the least square method.The vibration intensity at the frequency shift of 1 298 cm-1 has a good linear relationship with n-octane content.The benzene was selected as the marker hydrocarbon in the diesel oil drilling fluid and the mathematical model of different characteristic peaks of benzene in oil drilling fluid were established based on the least square method.The vibration intensity at the frequency shift of 986 m-1 has a good linear relationship between the vibration intensity and the content of added benzene.The experimental results show that the laser Raman spectroscopy can be used for in-situ detection of hydrocarbon concentration in drilling fluid and has good stability and repeatability in order to provide a new way for the inversion of oil and gas concentration and improving the accuracy of oil and gas reservoir identification.

A new synthetic strategy for fiber grating was proposed. For a given reflection spectrum of a fiber Bragg grating, the index profile of fiber Bragg grating could be calculated by a layer peeling algorithm. Some superimposed fiber grating with different apodisation functions were used to fit this index profile. The special reflection spectrum of a fiber Bragg grating could be realized by controlling the superimposed fiber gratings parameters adjustment.The relative error between the reflection spectrum of synthetic superimposed fiber grating and the goal spectrum is less than 4.0%. It is a sample and quick method for fiber grating synthesis problems which is easy to implement.

Fiber optic oxygen sensors based on the principle of fluorescence quenching have always been the focus by many research works. A simple and low-cost method for manufacturing fiber optic oxygen sensor was introduced.Based on the principle of fluorescence quenching,this method was realized by the fluorescence material platinum octaethyl porphyrin (PtOEP) coated on the fiber end.In the sensor,the fluorescent material was excited by the violet light of 395 nm and guided by the Y-shaped optical fibers,and furthermore, the luminescence timing diagram of the fluorescence was recorded using a spectrometer USB2000+ of Xipu Optoelectronics.The value of PtOEP (I0/I100)-1 is 0.78, that is to say,the sensitinity of the sensor is 0.78, and the Stern-Volmer diagram shows good linearity characteristics,and the response time from the air environment to the oxygen is 5 s while the response time from oxygen to air environment is 24 s.These results show that the fiber optic oxygen sensor based on fluorescence quenching principle has higher sensitivity and faster response time.

We proposed and demonstrated a multi-wavelength noise-like pulsed passive mode-locked erbium-doped fiber laser. The 980 nm semiconductor laser was used as the pump source and the erbium-doped fiber (EDF) with length of 2.5 m was used as the gain medium. The mode-locking operation was achieved by using a nonlinear amplifier loop mirror (NALM) structure. It is proved by self-correlation trace that the output pulse is a noise-like pulse. The 3-dB bandwidth of noise-like pulse is 17.2 nm, the side mode suppression ratio is 47.7 dB, the repetition rate is 5.434 MHz and the single pulse energy is 7.9 nJ. In order to achieve flat multi-wavelength output, a Sagnac ring interferometer was added to the NALM structure, and the flat multi-wavelength noise-like pulse was obtained with the number of maximum wavelength of 5 and the flatness of 1.995.