The motion shadow is conglutinouswith the object, and has the consistency of motion. It is often misde-tected as a part of the moving target. The existence of motion shadowchanges the shape of the moving object and influences the further analysis of the foreground of the moving target. To solve this problem, a motion shadow re-moval algorithm based on improved firefly optimization algorithm is proposed. The optimal threshold is obtained by optimizing the 2-Otsu distance measure function based on the improved glowworm algorithm which is based on the influence of the best position in the population history, and then the image is segmented and the moving shadow is removed. Compared our method with the traditional 2-Otsu method, particle swarm optimization (PSO) optimize 2-Otsu method, firefly optimization algorithm (FA) optimize 2-Otsu method, the experimental results show that the algorithm are 2.69, 1.42 and 1.21 times faster than the other three methods in the presence of shadow. Besides, it is superior to the other three algorithms in terms of region consistency, shadow detection rate and recognition rate. The effectiveness of the method is verified.

Automatic target recognition (ATR) technology has always been the key and difficult point in the militaryfield. This paper designs and implements a new DRFCN in-depth network for military target identification. Firstly, the part of DRPN is densely connected by the convolution module to reuse the features of each layer in the deep net-work model to extract the high quality goals of sampling area; Secondly, in the DFCN part, we fuse the information of the semantic features of the high and low level feature maps to realize the prediction of target area and location in-formation in the sampling area; Finally, the deep network model structure and the parameter training method of DRFCN are given. Further, we conduct experimental analysis and discussion on the DRFCN algorithm: 1) Based on the PASCAL VOC dataset for comparison experiments, the results show that DRFCN algorithm is obviously superior to the existing algorithm in terms of average accuracy, real-time and model size because of the convolution module dense connection method. At the same time, it is verified that the DRFCN algorithm can effectively solve the problem of gradient dispersion and gradient expansion. 2) Using the self-built military target dataset for experiments, the re-sults show that the DRFCN algorithm implements the military target recognition task in terms of accuracy and real-time.

Aiming at the problems of the existing vehicle object detection algorithm based on convolutional neural network that cannot effectively adapt to the changes of object scale, self-deformation and complex background, a new vehicle detection algorithm based on multi-scale context convolution features is proposed. The algorithm firstly used feature pyramid network to obtain feature maps at multiple scales, and candidate target regions are located by region proposal network in feature maps at each scale, and then introduced the context information of the candidate object regions, fused the context information with the multi-scale object features. Finally the multi-task learning is used to predict the position and type of vehicle object. Experimental results show that compared with many detection algorithms, the proposed algorithm has stronger robustness and accuracy.

In recent years, the use of new technologies combining infrared thermal imaging and digital holographic imaging to observe the targets in the fire field has become a current research focus. In theory, flame and smoke have no effect on long-wavelength infrared digital holography, but in the real fire environment, large particles of dust from the combustion will interfere with the light path, seriously increasing the reconstruction noise of the hologram. This paper proposes a new image processing algorithm to suppress the noise of infrared digital holographic reconstruc-tion. The algorithm uses a bilateral filter combined with the Laplacian pyramid algorithm to separate the details and energy layers of the holographic reconstructed image, filters the detail layer, and then superimposes the separated layers back into the reconstructed image by the inverse Laplacian pyramid algorithm. Therefore, the resolution of the reconstructed image is improved, and the simulation of the fire field environment proves that the algorithm has a significant effect on improving the resolution of the reconstructed image of the infrared digital hologram.

In order to overcome the limitation of current image enhancement algorithms for non-uniform illuminationimages, a brightness equalization algorithm is proposed to preserve the detail information in low illumination region and normal illumination region at the same time. The algorithm uses the adjacent frequency and position of pixels to generate illumination filter, so it can effectively separate illumination information and reflection information with de-tails. The illumination threshold is used to divide different illumination areas to compensate for low illumination brightness, so as to balance the image brightness. The experimental results show that compared with the classical naturalness preserved enhancement algorithm (NPEA), the average peak signal to noise ratio of the image in-creases by 15.4%, the average enhancement degree increases by 245.0%, and the average brightness step dif-ference decreases by 25.4%. The results of the proposed algorithm can maintain the details of different illumination areas while balancing the brightness and obtain a better visual effect.

Since the traditional algorithm may cause problems such as slow running speed and more mismatching points when perform stereo matching on underwater environment, the ORB characteristics detection and curve re-striction has been applied in this paper. Firstly the image should be detected so as to find out the characteristics, generate the descriptor, and match the feature points. Then the underwater curve restriction can be deduced ac-cording to the law of refraction combining internal and external parameters of camera. Finally the mismatching points can be decreased by means of underwater curve restriction. The experimental results have shown that in the case of effectively controlling mismatches, the speed of this algorithm are faster than traditional SIFT algorithm combined with curve restriction. As a result, it is of practical significance to improve the speed of underwater binocular vision system.

Aiming at the problem of high tolerance sensitivity and difficult adjustment of 30 times continuous zoom TV,the effects of eccentricity on the MTF (modulation transfer function) of the optical system were analyzed by opticalsoftware. The results show that the central error of the front mirror group is sensitive to the asymmetric aberration. Inthis paper, the structural form of the spacer ring machine program is optimized, so that the lower surface of the lensis automatically centering. As the moving component of the system, both zoom lens group and compensation lensgroup are the key factors affecting the system image quality during the zooming process. In this paper, mechanicalcentering tooling is used to make the central axis of the moving assembly parallel to the axis of the guide rod. Theoptical axis of all components is corrected by the optical axis of the front mirror group. The optical system is preciselyadjusted, and the optical resolution of the small field of view reaches 2.43″, which is close to the limit of diffractionresolution.

Aiming at the problem of large workload and strong subjectivity for manual retinal vessels extraction, this paper proposes a retinal vessel segmentation method that combines regional growing strategy, pulse coupled neural network (PCNN), a Gaussian filter bank and a Gabor filter. First, 2D Gaussian filter bank and 2D Gabor filter are combined to enhance the shape retinal blood vessel region and strengthen the contrast between the blood vessel and the background. Then, PCNN with fast linking mechanism and region growing idea is implemented to achieve automatic retinal vessel segmentation in which the unprocessed pixel with highest intensity is set as the seed, and the adaptive linking weight and stop conditions are adopted. The experimental results on the DRIVE fundus database show that the average accuracy, sensitivity and specificity are 93.96%, 78.64%, 95.64%, respectively. The segmentation results have less vascular breakpoints and clear micro-vessels. This work has promising application value.

A method for fast diagnosing gastric cancer is proposed, by combining optical fiber Raman spectroscopy system matching the gastroscope with the ratios of the spectral integral energy. we complete the detecting of Raman spectra from 17 samples of normal gastric mucosa and 12 samples of gastric adenocarcinoma mucosa using the optical fiber Raman spectroscopy system (excitation wavelength of 785 nm light, power of 50 mW, the CCD tem-perature to 80 ℃, acquisition time 1 s). Then, the original Raman spectra were pretreated, through reducing the baseline and smoothing by fast Fourier transformation (FFT). Finally, according to the characters of Raman spectra, Raman characteristic peaks were analyzed. At the same time, we compared the ratio of integral energy of conti-nuous band (1500 cm-1~1700 cm-1) and non-continuous band (1100 cm-1~1200 cm-1). The results show that the in-tensity of Raman peak of gastric adenocarcinoma at 1002 cm-1、1073 cm-1、1450 cm-1、1655 cm-1 belonging to phenylalanine and proteins are higher than that of normal mucosal relatively. From continuous band (1500 cm-1~ 1700 cm-1) and non-continuous band (1100 cm-1~1200 cm-1), the ratios of the spectral integral energy of gastric adenocarcinoma were different with normal mucosa markedly(independent samples t test, P<0.05), and with the ra-tio of the integral energy for use as a diagnostic index, obtained the higher accuracy (97.5%~98.5%), sensitivity (91.7%) and specific degrees (100.0%).

Silicon dioxide (SiO2) is one of the most widely used in various optical system as film material. The mi-cro-structure and defects of SiO2 films are of great importance to the functions and performance of these optical systems. In this paper, the absorption edge characteristics of single layer SiO2 films prepared by electron beam evaporation, ion assisted deposition, and magnetron sputtering are investigated in detail via calculating their ab-sorption edge spectrum, which is divided into three regions: the strong absorption, exponential absorption, and weak absorption regions. The bandgap, Urbach tail, and concentration of oxygen deficiency centers (ODC) are obtained by analyzing the measured absorption spectrum. By analyzing the bandgap, Urbach tail, and ODC data of SiO2 films prepared with different deposition techniques and annealed at different temperatures, the atomic arrangement as well as micro-defect information of SiO2 films are obtained and compared. Such information of SiO2 films are im-portant to the preparation of high-performance optical coatings employing SiO2 as the low refractive index material.