Study On Infrard Target Segmentation And Recognition Algorithm

With the developing of infrared photographic technology, the use of these technologies become more and more broadly in modern vectoring, searching and tracking. It is very meaningful to apply infrared techniques exactly in many fields. In the same time, the requirements of infrared images processing become more and more rigor, especially in the analysis of region of interest, such as target detecting, tracking and recognition. In modern infrared system, multi-channel photographic techniques are often applied to improve the performance of system, and make the system more robust and accurate. To effectively using such multi-channel information, the thesis's researches emphasize on the precise segmentation and fused recognition in the complex background.The main work in this paper is as follows.1. Proposed a new image segmentation method based on tensor field. Tensor expression, for its abundant information contained, become more and more important in modern image processing technology. A few methods of obtaining image tensor field have presented in reported research. By defining a new structure assume in local image, a new image...