Font Size: a A A

Research On The Camouflage Design Based On Background And Comprehensive Evaluation Method

Posted on:2015-08-04Degree:DoctorType:Dissertation
Country:ChinaCandidate:Z WangFull Text:PDF
GTID:1108330482455706Subject:Mechanical design and theory
Abstract/Summary:PDF Full Text Request
With the development of high technology, the detection and recognition technology used in satellites and aircraft has been fully implemented in electronic and digital operation, and these techniques have also gradually realized the high precision, and high real time capability for the requirement of actual combat including reconnoitering and attacking the camouflage targets. In the face of the challenge, the research on camouflage technology is of great significance for both the scientific research and military field. Through the analysis of the existing camouflage technology and equipment operational needs, the project which military equipment can realize to generate the image as similar with background in real time is one of the important research direction. If the adaptive camouflage can be realized, quick response display material is prerequisite and the image which can generated quickly consistent with background character image is also necessary; In addition, it is also an important link of camouflage technology to evaluate camouflage image whether or not similar to background. Therefore, the camouflage technology based on texture synthesis rapid optimization algorithm and camooflage comprehensive evaluation method are proposed, and through theoretical analysis and experimental calculation, the method can be demonstrated.The main research content and achievements of the thesis are as follows:(1) According to visual principle of animal camouflage, design thinking of camouflage can be obtained. Through summary and research, it is found that in the long-term evolution process the surfaces of many animals which can show the same characters and features have formed highly similar appearance with environment background, so it is difficult to distinguish from environment. In the field of vision, texture is a kind of feature which can express characters and features of the things. Therefore, when designing camouflage images, it is not only that the color information and spot shape information of background will be retained, but also the texture information the same as background will be expressed. This is the principle that the camouflage design should follow in the thesis.(2) In order to solve quality problems of the camouflage image, Texture synthesis optimization algorithm of adaptive neighborhood based on pixels is proposed. A small piece of texture image sample can generate any size image with similar texture features through the algorithm, while its texture features can keep randomness. In the synthesis process, the sample texton can be extracted by using block coding theory and according to it the size of optimum matching neighborhood can be confirmed; For the phenomenon of inaccurate matching with strong edge texture in the synthesis, neighborhood matching optimization criterion is put forward to capturing the characteristics of local texture matching, and neighborhood matching can be carried out by the similarity of pixel mean value, variance and texture to improve synthetic quality. Contrast experimental results show that texture synthesis method of adaptive neighborhood based on pixel can synthetize texture image more effectively. This method which can synthetize very satisfactory results for a variety of different sample has better applicability.(3) Aiming at real-time requirements in adaptive camouflage, fast texture synthesis method based on pixel neighborhood correlation can be proposed through improving search mechanism of neighborhood on the basis of texture synthesis method of adaptive neighborhood based on pixel. The search method will put pixel neighborhood as a whole, and four relevant candidate neighborhood can be searched in the sample texture through the local correlation, Every searching will have four times matches. If it is failure to find the eligible pixel, helix search mechanism can be used to search it. Experimental results show that the fast texture synthesis method greatly shorten the calculation time and can be used in adaptive camouflage image design on the basis of guarantee the quality of synthesis.(4) In order to solve the technical problems about universality and materials of the camouflage design, the improved clustering method is proposed to cluster quantitative synthetic images. this method makes high proportion of color in synthetic image as the initial clustering center in order to improve the efficiency of clustering; Then, the method establish clustering iteration error function until the iteration convergence, and several kinds of color can be selected as camouflage colors. Although the color of the image is reduced, it can’t affect camouflage effect because of texture feature information. The final image don’t only reduce chromatic aberration influence on account of printing or display device, but also the camouflage design of texture synthesis and clustering has a big advantage in time and quality through validating and contrast with the traditional camouflage.(5) According to the selecting principle of evaluation index and priority of visual identification, five characteristics have been selected as evaluation indexes, including brightness contrast, color feature, texture feature, edge shape, spot size. The method can compute separately ratio of brightness, color histogram, texture feature of Manhattan distance, moment invariant feature vector of Euclidean distance and db4 wavelet decomposition between background and camouflage images, and the similarity of characteristics can be as data sets of indexes. The entropy value method is put forward to calculating index effect weight. The evaluation index system can be established through integrating weighted index dataes into a whole, and the system provides data basis comprehensive evaluation of camouflage.(6)Grey clustering decision-making comprehensive evaluation method is proposed to evaluate camouflage image. Aiming at overlap problem of grey clustering ash class data in the existing whitenization weight function, grey clustering decision based on center whitenization weight function is proposed. This method will put the maximum in every grey class center as clustering nodes, then the optimum value can be identified as corresponding grey class by building evaluation measure matrix, While the objects in the same class need to compare the clustering results separately. According to the different gray level, the optimal, suboptimal and inferior results can be determined. Through the evaluation experiment of different camouflage images and the background, the evaluation method can make out best camouflage scheme for background and does not need artificial participation and quantitative data results also ensure the accuracy and objectivity to Guide making further improvement for images on the camouflage application. The evaluation method testify ulteriorly camouflage design advantages based on texture synthesis have a great advantage in quality.
Keywords/Search Tags:texture synthesis, camouflage painting, clustering algorithm, camouflage evaluation, neighborhood matching, entropy evaluation method, grey theor
PDF Full Text Request
Related items