Application Of Object Recognize Based On Infrared Image Sequence

The problem of Multi-Object detection, based on low contrast Infrared image, is a pop task in recent years. Recently, there are many practical methods about the problem. Because of its complexity and the low contrast, infrared image is undetectable. Thereby, according to the discussion and analyses of others research, and the statistic characteristic of infrared image, this essay provides a new method of region image enhance based on bidirectional histogram equalization, and a method of moving object detection based on image sequence. The result of the experimentation and theory show its efficiency for the multi-object detection.The first part of the essay introduce the development of the IV, and the result and product of IV produced by many cooperates and academes. And it recommend the Assistant Navigation System,Self adaptive Navigation Control System and so on. At last, we introduce the development of the Machine view since 1980s.The technology of image enhance was researched in this paper. Analyzing many kinds of technology of Image preprocess, analogizing according to experimentation, we provide a new method of image enhance adapt to the infrared image. First, according as the interested object, we select a region which should be enhanced. Second, image in this region was be enhanced with bidirectional histogram equalization; Then we can get a binary image which contain some objects. Therefore, the technology of object detection based on image sequence will be more efficient.This essay provides a scheme which comprises the method of the object detection, recognition and tracing. The result of the experimentation show the technology of moving object detection base on image sequence, which is according to the continuity of the movement of object, is efficient for the multi-object detection.