Space debris is posing a serious threat to human space activities and is needed to be measured and cataloged. With the development of space exploration technology, the measurement accuracy of debris is constantly improved, the measurement station is constantly increasing, and the accuracy of orbit determination and prediction of space debris is needed. By using the precise ephemeris of TerraSAR-X, we simulated photoelectric measurements of space debris, and evaluated effect of the empirical forces compensation model on the orbit and the prediction accuracy of space debris based on the same multi-stations’ simulation data of right ascension and declination. According to the orbital determination results by simulation data of right ascension and declination, the accuracy of about 50% data arcs increased about 25%.

Aiming at solving the problem of the impact of earth and atmosphere radiation on star extraction accuracy of the star sensor，the model about the irradiance effect on the star sensor was proposed by adopting infinitesimal method and analysis of the variables affecting the magnitude of the irradiance were made. The mechanism of the influence of earth and atmosphere radiation was explained，and defocusing simulation of earth and atmosphere radiation with different magnitude was achieved by White Gaussian Noise (WGN) with different variance and simulation comparison of the accuracy of star extraction was made under the condition of the earth and atmosphere radiation with different position and time. The simulation results indicate that earth and atmosphere radiation can lead to the decrease of the accuracy of star extraction, which can be adopted to evaluate the anti-earth albedo performance of the lens hood of star sensor.

In the system of real-time orbit determination of synchronous satellite, an algorithm based on simple features fusion is proposed to reduce the consuming time on star extraction from star image of CCD drift-scan. Firstly, segment star targets and background to decrease the influence of noise and improve the contrast of star image. Then, compute part maximum pixel gray value and mean contrast of partial region as star target feature and fuse the features, by counting optimum weights, to make the star target be sharper. The precision of star extraction is higher than traditional star extraction method. The amount of calculation of the algorithm which can be easily fulfilled is small, star extraction time is saved and the accuracy of star extraction is improved. The experiment results of processing the star image whose resolution is 1 528×1 528 taken by drift-scan CCD proves the validity of the algorithm.

In order to study spiral spectrum spread after Laguerre-Gassian (LG) beam propagating through atmospheric turbulence, the method of multiple phase screen is proposed to simulate the transmission of LG beams carrying Orbital Angular Momentum (OAM) in atmospheric turbulence. Propagation of Laguerre-Gaussian beam through atmospheric turbulence is simulated using phase screen based on Fourier transform following the non-Kolmogorov spectrum and the effect induced by atmospheric turbulence is computed quantitatively for different topological charges, wavelengths, exponent parameters, structure constant of the refractive-index fluctuations, inner scales and outer scales. The simulation results show that atmospheric turbulence deteriorates spiral spectrum of LG beam. It is also shown that system performance is closely related to topological charge, wavelength, exponent parameters, structure constant of the refractive-index fluctuations, inner scales and outer scales.

In order to solve the problem that majority of ship detection algorithms show the low accuracy and slow rate, a fine rapid ship detection algorithm based on salient feature guidance is proposed. First, the candidate target area is located through visual saliency model based on the local and global integration, and candidate target slices are gained through the region extraction. Then, the improved means clustering method is used to divide the target slice into super pixels. Finally, ship target regions are gained through filtering super pixels belonging to the target to achieve a fine segmentation after the integration of a significant figure and super-pixel segmentation. Experimental results show that the algorithm proposed in this paper can quickly and accurately locate the target ship, accurately depict the object contour, and is more conducive to subsequent follow-up work.

A semi-supervised graph-based clustering method is presented with composite kernel for the hyperspectral images, mainly to solve the problems existed in an algorithm called Semi-Supervised Graph-Based Clustering (SSGC) and improve its performance. As for the realization, it firstly reforms the Radial Basis Function (RBF) by adopting semi-supervised approach, to exploit the wealth of unlabeled samples in the image. Then, it incorporates the spectral angle kernel with RBF kernel, and constructs a composite kernel. At last, the use of K-Nearest Neighbor (KNN) method while constructing the weight matrix has greatly simplified the calculation. Experimental result in Indian Pine and Botswana hyperspectral data demonstrates that this algorithm can not only get higher classification accuracy (1%～4% higher than SSGC, 10%～20% higher than K-means and Fuzzy C-Means (FCM), but effectively improve operation speed compared with SSGC.

In order to overcome the weak of the limit ability of preservation of edges and easy to produce visual artifacts in some super-resolution methods based on dictionary learning, we propose multi-dictionary learning imagesuper-resolution method with edge-enhanced, which can effectively restore the image edge details. Firstly, the training image patches will be classified by using K-means, and then quickly learn multi-dictionary pairs by employing the Boost K-SVD algorithm. During the super-resolution reconstruction, the method adaptively selects the optimal dictionary pairs for sparse decomposition and recovery. To improve the visual quality of edge after image reconstruction, we employed direction-preserving regularization according to the input test low-resolution (LR) image, meanwhile learning the natural image database edge sharpness statistics prior to constraint the image reconstruction of edges. The experimental results demonstrate the effectiveness of the proposed algorithm.

Due to the background noise and the detector noise, the centroid error of Shack-Hartmann wavefront sensor is relatively large. A method to filter the noise in sparse region is proposed. First, the sample image is generated by using the two-dimension Gauss model and the over-complete target dictionary is constructed. Then, the sub-image blocks of the test image are extracted successively and the corresponding coefficients are calculated with the constructed over-complete target dictionary. Since the coefficients between the noise and the target are large different, the difference by setting a threshold is used to subtract the target. Experimental results show that, the target can be well subtracted, while the centroid deviation and the RMS and PV of the centroid are smaller than using the method of subtracting threshold

Bombers in our army have not been equipped with the Electro-Optic Targeting Systems (EOTS), which cannot observe targets at night and as a result cannot conduct precise bombing with the all-optic targeting systems as their bombing targeting systems, so the EOTS is needed. To solve the problem of multiscale image fusion in the process of implement the EOTS, the purpose of image information fusion has been tried to achieve through the multiscale edge fusion. With the fusion data compression, the influence of noise on the fusion result has been tried to reduce at the same time. Based on the multiscale edge theory, an image fusion method for bombing EOTS is proposed, and the effectiveness of the proposed method is verified by experiment. Results show that the proposed image fusion method based on multiscale edge has robustness against mismatch and noise, and provides a theoretical basis for the realization of the bombing EOTS.

Trying to provide a more effective algorithm for real-time counterpoint and alignment in the image, we propose an estimate method using affine transformation modal to calculate image translation rotation parameters, which is different from the traditional iteration method in searching image. It’s the first time to provide the algorithm which can calculate target and alignment image registration parameter by single step, using the model of the least square of the affine transformation homography. And the effectiveness of this algorithm is validated through the corresponding experiment.

In order to overcome the difficulty that texture feature space dimension of multi-band infrared images is large and hard to analyze its imaging differences synthetically, a three-level feature selection model was proposed. Firstly tested consistency in the form of two contrast, and it was selected the features that difference distribution laws were the same. Secondly, the function of independent sample test based on SPSS software was used to test difference significant degree, and the features whose differences that were not significant were eliminated. Then, the correlation of features was analyzed, and one of the features stood for the features with high correlation. At the same time, the rest of features were reserved. Experiments prove that a few features selected by this model can reflect features differences of image, and it will provide references for target recognition and so on.

Aiming at insufficient details information extraction and higher complexity algorithm when the infrared and visible light image fusion is processed, an infrared and visible image fusion algorithm based on the Non-subsampled Shearlet Transform (NSST) and the improved Non-negative Matrix Factorization (NMF) is proposed. Making use of NSST to decompose source images on multi-direction and multi-scale sparse, low-frequency components and high frequency components are obtained. The fusion method of the improved NMF is adopted in the low frequency subband. The fusion rule for the combination of the Laplace energy and visual sensitivity coefficient is used to the high frequency components. Finally, the fusion image is obtained after executing the NSST inverse transformation. Experimental results show that the fusion method can not only guarantee the definition of the fused image, but also shorten the running time of the algorithm.

In order to manufacture MOPA structure pulse single-frequency laser used for laser coherence imaging, the changes of temporal coherence of pulse single-frequency laser with the increment of laser power is required to be studied. This paper adopted acousto-optic Q-switch pulse single-frequency Nd:YVO4 laser as seed，and adopted Nd:YAG poweramplifier modules to double the power of seed laser. The Michelson interferometry was used to complete the measurement of the contrast of fringes of seed laser and amplifier under the optical path difference from 0 m to 6 m. Then, the change of temporal coherence of pulse single-frequency laser with the increment of laser power was primarily obtained. The measurements of basic parameters of seed laser and amplifier such as the property of single-frequency, pulse energy, pulse width and peak power etc was completed.

Frequency conversion characteristics, produced by interact between laser and the object, are of great significance for target detection. The echo signal of frequency up-conversion has weak strength, pulse stretching, phase shift characteristics etc. According to this echo feature, we use the PIN photodiode as detector, and extract echo signal using lock-in amplification technology. With the help of the preamplifier circuit and AC signal amplifier circuit, the transformation and amplification of echo signal can be achieved, as well as phase sensitive detector, which is used to realize echo signal extraction. Echo detection system is set up and experimental tests show that detection system based on lock-in amplifier can suppress noise signal, as a consequence, the system can extract echo signal from background of big noise. When target at 1.5 meters in front of detection system, the output signal-noise ratio (SNR) of the system is 16.9 dB.

Energy is concentrated in the middle part of the beam for the traditional narrowband laser beam shaping. This energy distribution approximates a Gaussian distribution which can not meet the requirements of large field line Pushbroom 3D imaging laser radar technology. A novel beam shaping technology is proposed to solve the problem of Pushbroom lidar which needs a narrow strip of the laser beam. According to the optical properties of cone mirror, we use Biaxial Axicon to change the spatial energy distribution of the laser beam, and then energy distribution of the laser beam are shaped "U" through the cylindrical lens. Experimental results show that the "U" shaped narrowband of the laser beam can meet the requirements of large field of linear array push-broom imaging 3D laser radar.