Fusarium Micrograph Processing And Recognition System Development

Fusarium is a kind of widely distributed pathogens. It is great harmful to crop and cash crop that the toxins produced by fungi can cause many diseases of crops such as wilting, ear rot, decay, and even endanger human and animal health. During the last ten years, European and American scientists have made some breakthroughs in fusarium research, so that the understanding of the fusarium species which cause serious disease and produce toxin to crops developed with rapidity. In China, it is reported frequently that fusarium toxins is detected from cereal grain and feed. Meanwhile, the cognition of the fusarium feature should be improved. It shows that extraction and identification of the fusarium has important scientific and practical significance to our lives.Based on digital image processing technology, this dissertation develops a system of fusarium image processing and recognition which is integrated the knowledge of image processing, pattern recognition and extraction of the feature. This research has been put forward mainly on the subject of fusarium micrographs. Firstly, according to the gray feature of the fusarium microscopic images and based on the color information of the fusarium, it makes use of the vector median filter method for color images. The test result shows that the vector median filter method can operate at speed and produce a good effect on the filter. Secondarily, in the light of the characteristics of the fusarium microscopic images, it segments the fusarium microscopic images which adopts the method of gray-level threshold and achieves the desired results. Because of the segmentalized images have isolated little spots, burrs and little cavities, and such image noises will produce certain influence for the later recognition work, it adopts mathematical morphology method such as opening and closing operations to process the segmentalized image further and accomplish the purpose of eliminating the image noises. Finally, based on the biological features such as the shape and the veins of the fusarium, it recognizes the types of the fusarium with the SVM recognition method.This paper develops a set of fusarium image processing and recognition system depended on the Visual C + + 6.0 development platform and the Matlab7.0 test platform. In the system, the image processing part will use VC++ language compile directly, then parts of extracting the feature of the fusarium images and final recognition will be accomplished by using the method of VC++ which will be as a development platform to transfer the MATLAB language. MATLAB become the de facto standard of the mathematical calculation tools, with its powerful computing and graphics capabilities, a large number of stable and reliable algorithm library, a rich library of image processing function, simple and efficient programming language. With the visual output result and applicable overall performance, you can easily change the image. Especially, with minimal code, the internal image processing functions greatly simplify the procedure of the complex image processing under the VC environment...