The Study And Realization Of Centralized Soccer-robot Vision System

People acquire about 80% information from outside world by their vision system. The intelligence machine is endowed with human being's vision function that is very important. Chromatic image is the best approach to the world, so there are most advantages in chromatic image processing than in gray-level image. At the same time, to build up a fast reacted vision system as human being's is the research emphasis of those above fields. Robot-soccer vision system is just based on chromatic image processing, and with a great demand for its real-time performance, and then it can be an ideal subject for this research field.In the paper, according to the condition and characters of robot-soccer game, it analyzes the skill fields of robot-soccer system. For the purpose of building up the vision system, it discusses the principles and methods of chromatic image processing, especially focuses on the theories and techniques of segmentation and recognition of chromatic image and how to increase the system real-time performance. The paper constructs the computing vision theory for the PC-based real-time chromatic robot-soccer vision system, and then analyzes the components of this algorithm structure and completes the design of soft ware and hard ware.How to complete chromatic image pre-processing and segmentation and object orientation in real-time is the key of the robot-soccer vision system. The paper puts forward a histogram-based bi-threshold automatic method to segment chromatic image, which can enormously expand the theory and application of gray-level image processing. It discusses how to increase the efficiency of the chromatic image segmentation. After the compare, we adopt a CLUT method based on division of color space to segment the chromatic image. To achieve the real-time demand, it studies the scan strategies of image. After ensuring the precision, the speed of the system is increased.The paper includes 7 chapters. In chapter 1, it introduces the history and status of robot-soccer, and the content and meaning of the paper. In chapter 2, the design ideas of MiroSot system and every model systems are discussed. In chapter 3 and 4, the principles and methods of chromatic image processing are discussed, and many of arithmetic are compared each other. In chapter 5, the skill keys of robot-soccer vision system are discussed and the scan strategies of image are most important. In chapter 6, the design of soft ware and hard ware is statement. In chapter 7, it sums up the work of the paper and presents the problem of the paper.