摘要:A theory and methods for automatically recognizing a hundred classes of targets are presented in this paper. It consists mainly of both description models and classification models of targets. A related system based on it is set up yet, a classification model for arbitrary combination among a hundred classes of targets can be fast generated, and correct recognition rate more than 99% is assured.
关键词:Targets recognition;Description model;Classification model;Automatic target recognition system
摘要:The estimation of phases from measurements of phase differences plays an important role in radar interferometry, adaptive optics, compensated imaging and image processing. The derivation of Possion's equation from the least squares estimation with the concept of the 2 norm is presented in this paper. It has been shown that the equation can be solved by use of the cosine transforms.
摘要:In this paper we present algorithms by using statistic/geometric features and simulating human vision mechanism to extract lines and arcs for vectorization of engineering drawings. Experimental results for complex line drawings confirm performances of these algorithms.
摘要:In this paper, an algorithm for real part reconstruction of partial K space data in MRI is presented and realized. The test result is given and discussed.
关键词:Magnetic resonance imaging;Reconstruction in MRI
摘要:In this paper, a neural network approach to DPCM(differential pulse code modulation) design for encoding of images is presented. This approach can deal with high-order statistics and nonlinear correlation which the linear method can not. Based on this method, this paper provides a predictor—2DNNADPCM, and presents an approach that divides an image into different sub-block according to different messages and requirements in order to achieve better results.
摘要:A new mean residual fractal image coding algorithm with fast decoding is proposed in this paper. Theoretical analysis and experimental tests have shown that the iteration times in decoding process are reduced significantly with an improved PSNR value. In addition, the proposed new method is further combined with adaptive quadtree partition to obtain a good trade off between compression ratio and image quality. Superior coding perfermance is confirmed by simulation tests as compared to the conventional fractal coding.
摘要:This paper analyses the principle of fuzzy vector quantization (FVQ) for image coding, gives three strategies of FVQ algorithm during the codebook design process, and presents six new fuzzy vector quantization algorithms with exponential membership function for image coding, named FVQE algorithm. Simulation results show that the performance of FVQE is comparable to FVQ for image compression, and a slight improvement on the convergence speed is obtained.
关键词:Image coding;Fuzzy vector quantization;Fuzzy K means algorithm
摘要:A new edge detection method and mathematical model of line moment described by chain code
are discussed in this paper. The constraints of line moment invariant-based matchimg are presented, and exact sub-pixel matching is realized by feature-based and least square method.
摘要:In many current multimedia database systems, query or retrieval is only based on descriptive keywords of media, it ignores another important information resource(the content of media. Content based image retrieval techniques usually use the color histogram of the full image. But this method can't contain the spatial features. In this paper, we introduce a color-pair retrieval method that derived from color histogram. It can reflect the spatial information of the images. Therefore it can retrieval images with clear boundary. The ret rieval results are not affected by image's size, rotation and also the little changes of light. The experiment result shows this method improves the retrieval effect.
摘要:This paper gives an illumination model to simulate atmosphere, cloud and fog in natural environment. For some natural landscape with a long distance of vision and a wide field of vision, such as mountain modeling, it has more realistic effect and can get 3-D topographic landscape map with weather features of cloud and fog. The results of the experiments show that our model is satisfactory, practical and effective.
摘要:Components GIS (ComGIS), a new technology based on Components Software, is a mainstream in GIS technology. In this paper, the concept and technological foundation of ComGIS are discussed. And some differences between ComGIS and traditional GIS, as well as the relationship between ComGIS and WebGIS are studied. Also, system designs and development models of ComGIS have been presented based on recent practices.
摘要:Based on the topological theory, this paper studied a new approach named"Node Searching" for transferring raster to vector. The main advantages of the new approach were: (1) high efficiency and high speed;(2) capable of building topological structure of vector data automatically during converting; (3) unnecessary to process raster data before converting; (4) not very sensitive to the number of raster image's column and rows. It was proved to be a good approach in experiments.