The Illumination Invariance Vision System For Androsot

Robot soccer grows rapidly in recent years, including FIRA and ROBOCUP two major events. The subject of this paper is based on Androsot, which is a humanoid robot match in FIRA.This paper shows a software platform solution about Androsot, and gives the detail of team CUG in Androsot match.The current Androsot uses semi-autonomous control mode,which means that the robot has no independent capability of vision and decision.Therefore, Androsot system consists of four subsystems include robots,vision, decision and communication.Humanoid robot in robot system has 16 degrees of freedom,and no vision or decision system.So the visual and decision capacity is depending on the external global vision and centralized control decision making.The digital camera, which is hanging on the central top of the field, and the image processing program constitute the vision system.The camera captures the global image continuously;the program processes the image and transfers the position data to decision system.The decision system records and analyzes the data, then makes the strategy according to the position data. After that, the decision system converts the strategy to robot command and push to communication system.At last the communication system sends the command to the robot and the robot acts automatically.This paper focuses on the vision system and finds that light influences the system obviously. When the light intensity changes,in order to recognize the object correctly, the vision system must capture the color patch repeatedly;this means a lot of unnecessary time waste.Therefore,the vision system can get improvement in adaptability of the light intensity In this paper, color database adaptive algorithm is designed according to computer vision theory, color theory and the relevant experiment data. This algorithm calculates the average light intensity in the field, and maps it to the color database.Then the color database can be generated automatically.Color database adaptive algorithm is able to handle the light problem generally. However, the light intensity comes to the extreme situation,the vision system stops working.Therefore, the improved color database adaptive algorithm has been designed.The improved algorithm designs a function to evaluate the image, which is captured by the camera.The improved algorithm also designs a method to adjust the inner parameter of the camera according to result of the function.If the color shifts,adjust the white balance parameter; if the light intensity is not proper, adjust the shutter or the gain parameter. When the result of the function is in acceptable range,the improved algorithm is finished and then the vision system is able to work normally. The improved algorithm achieves the adaptability to light intensity of extreme situation.