In recent years, lighting simulation equipment used in space environment simulation has attracted much attention, including solar lighting simulators, light array lighting simulators, infrared heating cages and so on. Solar illumination simulation equipment mainly includes lamp room, integrator, window, mirror, etc. Because of the particularity of vertical illumination simulation structure, the metal mirror needs to adopt hanging structure. Combined with the characteristics of high irradiation and high heat energy of illumination simulation system, it will produce higher heat radiation energy on the surface of metal mirror. The temperature uniformity and thermal deformation index under the load are very important to the irradiation uniformity of illumination simulation equipment, that is, the high irradiation hanging metal mirror has high requirements for thermal stability.The thermal stability of the hanging meter-level metal mirror used in high irradiation environment simulation equipment was systematically analyzed, and the thermal stability analysis of the single meter-level metal mirror used in vertical illumination simulation equipment was mainly conducted, including thermal simulation verification and analysis, mechanical simulation analysis and optical simulation analysis. Finally, the structure illumination simulation equipment. The proposed thermal stability analysis method provides a guiding method for the structural design of the hanging metal mirror, a key component in the subsequent high irradiation vertical illumination simulation equipment.

Aiming at the stereo matching problem in digital speckle projection 3D measurement technology, a semi-global stereo matching algorithm based on Inverse Compositional Gauss-Newton iterative algorithm is proposed.The algorithm uses the matching cost value based on the Census transform to measure the similarity between pixels, and the cost aggregation algorithm enhances the robustness of the matching cost to noise;On this basis, a reverse synthetic Gauss-Newton algorithm based on the first-order shape function is proposed to achieve sub-pixel parallax optimization;Finally, an erroneous parallax elimination algorithm coupled with uniqueness check, left-right consistency check and connected region constraints is proposed to eliminate the erroneous parallax caused by occlusion and noise. Experimental results show that the algorithm can accurately, densely and completely reconstruct the three-dimensional topography of complex objects, and the reconstruction accuracy is better than 0.06mm.

Speckle noise in digital holography has a great influence on image quality, and one of the reasons for the formation of speckle noise is that the coherence between reference light and object light is too high, which leads to speckle effect and introduces unnecessary noise to recorded hologram. Speckle denoising is realized by using electrically controlled liquid crystal scatterers to decolorize coherent light. By adjusting the voltage to change the scattering effect of the electrically controlled liquid crystal scatterer in the object light path, the holograms irradiated by the liquid crystal scatterer under various scattering effects are recorded, and the equivalent appearance index of the reconstructed object image under various scattering effects is different, and the optimal result is obtained through comparative analysis. Then the hologram is reconstructed by superimposing the hologram under the optimal scattering effect and the effect is better than that of the single image. Through theoretical analysis and experimental verification, it is proved that the proposed method has great application value to improve the imaging quality of digital holographic system.

Aiming at the problems of low-resolution lidar range image edge detection accuracy, poor anti-noise performance, and the need to set the threshold manually. An edge detection algorithm with improved Sobel operator is proposed. The improved algorithm increases the accuracy of range image edge location by adding a gradient template and uses a median filter to calculate the local edge threshold. Therefore, it can realize adaptive threshold calculation and enhance the anti-noise performance. Then FPGA hardware acceleration is used to achieve real-time edge detection of the algorithm. In addition, the lidar range image acquisition system is designed and the algorithm is verified through experiments to meet the requirements of real-time edge detection.It is better to extract the complete edge information of the range image, which has certain engineering application value.

Aiming at the problem that the scanning speed and scanning accuracy cannot be taken into account in the fixed focus process of traditional confocal microscopy, a fast surface profile measurement method based on virtual double differential confocal is proposed. This method shifts the traditional confocal response curve forward and backward by a specified amount to obtain virtual front focus and virtual back focus signals. Then the virtual front and back focus signals are subtracted from the original confocal signal respectively, and the subtraction results are added together to obtain a virtual double differential curve. The virtual double differential curve has the characteristics of large linear range and high fixed focus sensitivity. When the surface profile is acquired under the condition of large scanning distance, the rapid and high-precision fixed focus of the sample surface profile can be achieved. Simulation analysis and experimental results show that, compared with traditional differential confocal microscopy imaging methods, the virtual double-differential confocal surface profile measurement method can increase the imaging speed of geometric topography by more than 2 times while maintaining high precision and fixed focus.

A new method combining microscope and microfluidic chip technology was proposed to realize the dynamic microscopic imaging of red blood cells, which was used to measure and analyze the multi-angle morphology of red blood cells (RBCs). Firstly, the microfluidic chip is used to make the red blood cells rotate or turn in a specific position by the contraction/expansion structure; then, using Matlab frame processing to collect the video image, through gray adjustment, median filtering, morphological processing, centroid method, object recognition, dynamic tracking, etc., to achieve the effective identification of blood cells and accurate classification and counting; finally, multi-angle morphological analysis was performed on the classification results, and the maximum diameter of RBCs with different morphology was measured. The accuracy of the proposed method is verified by analyzing the single cell multi-frame associated images. The blood of diabetic patients and normal people was tested by using the method of multi-angle morphological analysis of RBCs. The results showed that the average maximum diameter and the distribution width of diabetic erythrocytes were larger than that of normal erythrocytes. The method of microfluidic chip combined with image processing can realize high-throughput detection and multi-angle analysis of RBCs in diabetic patients and normal people, which provide a new method for multi-angle morphological measurement of other types of sample cells.

Aiming at the over-correction problem of the fringe template measurement method in the lens distortion correction, the phase analysis of fringe-pattern combined with the distortion model is adopted to realize the measurement and correction of the lens barrel distortion. Firstly, the standard cosine fringe-pattern is employed as the measuring template and captured by the camera with wide-angle lens to obtain the distorted fringe-pattern. Four-step phase shifting analysis, which is excellent at spatial localization analysis, is employed for the phase analysis. And then the distortion center and the radial distortion phase are obtained. According to the even-order polynomial model of the radial distortion, the distortion parameters are calculated by numerical curve fitting. Finally, the image correction is achieved according to the distortion parameters and the distortion center. Numerical simulation and real distorted image experiments have been carried out to validate the method.

As an important part of evaluating the performance of glass-ceramic materials, bubble defect detection can guide the cutting and reprocessing of the original glass-ceramics to give full play to its stable optical characteristics. Due to the problem that the background brightness of the field of view is uneven and the outer periphery of the glass-ceramic is brightened when the circular light source illuminates the cylindrical glass-ceramic, a glass-ceramic bubble detection algorithm based on grayscale and geometric characteristics is introduced. Possible bubble outlines are obtained by Canny edge detection according to the local image gray feature, and then the false outlines are removed according to bubble geometric characteristics to obtain the real bubble defect information. Experimental results show that the algorithm can effectively extract bubble defects and improve the accuracy of bubble defect detection.

High precision measurement of temperature changes is of great significance in modern technology. It is usually hard to measure ultrasensitive temperature changes directly. Based on the fact that temperature changes can cause the phase changes between the o and e light, a polarization dependent weak-value-based system is constructed to amplify these phase changes that will alter the intensity of detected light, from which the temperature changes of multi order quartz wave plate can be inferred.. The precision can reach 0.0042℃, which is 9 times higher than the balanced homodyne detection.

Spectroscopic modeling of soil element content based on machine learning and deep learning is a research hotspot in soil chemical composition detection. In order to improve the accuracy of soil element content spectroscopy modeling based on convolutional neural network, a particle swarm optimization optimized convolutional neural network soil nitrogen element content spectral analysis model is proposed. Smoothing and standard normal transformation of soil samples was used to reduce the impact of noise on modeling. A convolutional neural network structure suitable for regression was designed. The hyperparameters of convolutional neural network such as kernel parameters, learning rate, number of iterations were optimized by using particle swarm optimization. The wavelength range of the visible/near infrared spectrum is 225~975nm, by analyzing and modeling the nitrogen content of 177 groups of soil samples in Qingdao area, the model’s results show that, compared with PLS, CNN and other modeling methods, the PSO-CNN model proposed in this paper has a higher prediction accuracy. The coefficient of determination of the test set is 0.9707, the root mean square error is 0.8818, and the ratio of performance to standard deviation is 5.88.

The incidence and mortality of liver cancer are increasing year by year. Finding effective methods to improve early screening and diagnosis of liver cancer is an important way to reduce mortality. Quantitative detection of serum alpha-fetoprotein (AFP) is an important auxiliary method for early screening and diagnosis of liver cancer. The commonly used AFP detection methods have the disadvantages of high cost, long detection time and unsuitable for large-scale population screening. The method of combining a biochip scanner based on the principle of laser confocal and a near-infrared fluorescent-labeled antibody chip is used to achieve the quantitative detection of AFP, and compare it with the commonly used Roche electrochemiluminescence immunoassay method to detect AFP content in patient samples. The results show that there is a high correlation between them.

The lens light-screen detection sensor with L-shaped LED light source constitutes a large area rectangular detection light screen, which is usually arranged in a closed indoor test target to measure exterior ballistic parameters. The sensitivity of the detection area directly affects the measurement accuracy. The spatial distribution of sensitivity of 10m×10m large area rectangular detection light screen with L-shaped LED light source is been studied and analyzed , and the len edge effect and light propagation path is considered .The detection area is divided into small blocks, the change of luminous flux caused by the same caliber projectile passing through different blocks is calculated by digital simulation, and the spatial distribution of sensitivity of rectangular detection light screen area is obtained. The results show that the sensitivity distribution of the rectangular detection area is symmetrical. When it is closed to the main optical axis of the lens and near the lens, a larger detection sensitivity in the detection area will be obtained. The research results provide a theoretical basis for the subsequent large target exterior ballistic testing technology.

According to the requirements of processing, testing and lifting of a 2.7m spherical mirror, an inspection device for support testing, posture adjustment and turnover of mirror body is designed, which can meet the requirements of the support accuracy of the knife edge instrument in the rough polishing stage, the posture adjustment requirements in the mirror body test, and the overturning requirements of the mirror body back structure processing and assembly. For different working conditions, the stiffness and strength are analyzed, the force model of the mirror body under various working conditions is established, and the supporting accuracy, stiffness and stress of the mirror body are analyzed. The results show that the mechanical structure of the device meets the requirements of strength and stiffness; the support stress of the mirror body is less than 1MPa under various working conditions, which is less than the required stress of microcrystalline; the accuracy RMS of the mirror body with the sling side support is 14.15nm, which meets the measurement accuracy requirements of rough polishing stage.

In recent years, laser cleaning technology has been gradually applied in the field of cultural relics protection. Nanosecond laser was used to clean the yellow layer (rust) and black layer (a mixture of CaSO4 or PbCO3 with dust) on the surface of ceramic relics. The ablation effect and cleaning effect of pollutants on ceramic surface were controlled by changing the laser power, scanning speed and frequency of cleaning. The optimum cleaning parameters of the two pollutants were explored by means of naked eye observation and laser scanning confocal microscope. The best cleaning parameters of two kinds of pollutants on the surface of ceramic cultural relics were obtained, providing data support for the restoration and protection of ceramic artifacts.

Through laser cladding test, metallographic test, residual stress detection and static tensile test of cladding specimens, and the variation law of mechanical properties of laser cladding specimens were studied. It was found that the thick weissenite structure was produced in the remelting zone of the cladding specimen, and the residual tensile stress was introduced into the cladding specimen during the cladding process. Compared with the matrix specimen, the tensile strength of the cladding specimen was reduced to a certain extent, and the elongation of the cladding specimen was significantly reduced. The experimental results show that the mechanical properties of the cladding specimens are reduced in different degrees compared with the matrix materials, that is, the matrix materials have a certain degree of damage after laser cladding. Here, this kind of damage is quantitatively analyzed and a quantitative evaluation model of cladding damage is established, which provides a theoretical basis for improving the quality of laser cladding.

The existing visible light multi-carrier communication systems suffers low transmission rate and critical inter-frame interference, the Salp Swarm Algorithm is taken advantage of to optimize the modulation module algorithm of visible light multi-carrier communication systems, in order to increase the data transmission rate of visible light communication. The characteristics of multi-carrier modulation of visible light communication system is analyzed in detail, and comprehensively considers the important parameters of relative modulation systems, it also presents a reasonable quality evaluation function. The exploration operation of Cuckoo Search is introduced the into Salp Swarm Algorithm to enhance the global search power of that algorithm, finally the improved Salp Swarm Algorithm is taken advantage to optimize the multi-carrier modulation module of visible light communication systems. Simulation experimental results show that the parameters of visible light communication multi-carrier modulation can be optimized with the Salp swarm algorithm, and the transmission speed of the visible light communication increases effectively.

In the complex infrared scene of the battlefield, the target is easily confused in the background due to the irregular gray level distribution, the blurred edge of the object and the lack of texture features. Due to the limitation of the computing performance of the embedded platform, most of the deep learning detection algorithms are difficult to apply to portable devices and cannot achieve fast and effective object recognition. An object recognition method based on moving object extraction and efficient machine learning model is proposed. This method firstly realizes the target pixel-level segmentation through motion detection, and locate the single target after morphological processing. Then select lightweight deep network features or contour features according to the computing performance of the embedded platform, train softmax model to achieve object classification. This algorithm is transplanted to the embedded platform, and the object recognition experiment is performed on the open source infrared image sequence, which can realize the simultaneous positioning and classification of multiple objects, and the processing speed is up to 56FPS. Experimental results show that this method can effectively identify infrared targets in complex backgrounds in real time.

The advantages and disadvantages of traditional edge detection operators are analyzed; LVQ neural network is introduced to detect image edge; the detection principle and training process of network weight of LVQ is detailed, in which the result of traditional operator detection is taken as teacher signal and the input signal is based on a set of features of 5 × 5 neighborhood, such as median feature, direction feature and Kirsch operator direction feature of gray image, the calculation formula of those features is given; based on different thresholds and teacher signal types, the edge of navel orange image is detected using LVQ neural network; the results show that edge shape of the image is independent on the type and threshold of teacher signal, and as significant advantages over traditional operator detection, edge continuity is improved and overexposure suppressed.

To resolve the difficulty and low efficiency problem of camera calibration for large-scale vision system, a fast calibration method for camera defocus based on phase shift coding circle is proposed. Firstly, the feature points are extracted in the defocused state by using the phase shift coding circle. Secondly, the target function of monocular camera is constructed through utilizing spatial location information given by the calibration target, and then make use of the nonlinear least square method to solve the internal parameters of the camera. Finally, by establishing the optimization function of the binocular system, the external parameters of the camera system are solved. This method does not need to move the camera or target, and can complete the rapid calibration of the large-scale vision system camera in the defocused state by shooting 3 phase-shift circle images. The experiment results shows that the deviation of the calibrated focal length is less than 0.5%, and the maximum inverse projection error is 0.33pixel, which solves the difficult and low efficiency problems of large-scale vision system camera calibration.

In view of low illumination images existing in many application fields, a low illumination image enhancement method based on quantum behaved particle swarm optimization is proposed. The method extracts non-textural regions of low illumination images, and limits the gray level range of non-textural regions to reduce the artifact situation while enhancing the images. A hierarchical probability density function for Histogram Equalization is designed, the improved quantum behaved particle swarm optimization algorithm is applied to search the optimal parameters of hierarchical probability density function. Finally, validation experiments are carried on typical low illumination images, the results indicate that the proposed method not only improves the contrast of images, but also enhances the detail information and visual quality of images.

The impulse noise is one of the main noise sources during the imaging processes, meanwhile, it is hard to remove high density impulse noise with traditional filters. To solve this problem, an image denoising algorithm for impulse noise based on asymmetric neural networks is proposed. In this algorithm, the steganalysis rich model is used to extract the convoluted noise feature maps of the noisy image. then the feature maps of original image and noisy convoluted version are delivered to two identical convolutional neural networks. The l1 loss and l2 loss are combined as the total loss function of the neural networks, it takes advantages of high visual quality of l1 loss and strong convergence of l2 loss. Experimental results show that the proposed denoising method outperforms the filter based denoising algorithms on different densities of noise, it also performs better on high density impulse noise.