Combining The Fuzzy Theory Of Visual Recognition And Tracking Technology

With the development of digital image and computer visual technology, target recognition and tracking is applied gradually in automatic monitor system, intellectualized traffic system, and military automatic collimation and exact navigation, and reveal great prospect in agriculture and industry field. In the paper, focusing on target recognition and tracking technology of the color image, a method of automatic target recognition based on fuzzy theory was proposed, and a method of fuzzy control of an automatic dynamic object tracking was proposed by combining target's area and ubiety in the image. Then, a foundational experimental system based on small locomotor robot was developed and some experiment to validate the proposed methods was actualized.The main component of this paper is as follows:(1) A algorithm to convert Hue and Saturation in HIS color space to a color variable C by utilizing the fuzzy theory was proposed and threshold of color image segmentation was achieved automaticlly, and then, based on this, filtering, erosion and dilation of color image was processed. Final, color target recognition and tracking can be reallized effectively.(2) Combining with the arithmetic of computing color variable C, a method of fuzzy control of an automatic dynamic object tracking according to target's area and ubiety in the image was proposed, and the function of searching and tracking dynamic target automatically was realized. This method does not need to compute the distance and speed of object in the tracking process, so it can be more brief and practically.(3) A target recognizing and tracking system was developd to validate the proposed methods in the paper. This system include: a small locomotor robot, a cloud terrace, the hardware of cotrol system, the control software and image processing software. By using this system, target recognizing and tracking of color target can be realized.