The Research Of Middle-Size League Robot Soccer’s Identifying Color-Undefined Ball Based On Limited Condition

In recent years, as the platform of small high-techcompetition and the carrier ofartificial intelligence and Robotics, robot soccer competitionhas been the general concern of national and international scholars. Soccer robots relate to the key technologies in manyscientific fields of machinery, electronics, control, pattern recognition, image processing and artificial intelligence and so on, and have high demands for real-time and accuracy.Middle-Size League isa hot topic of robot soccer’s research nowadays, andan important event in the Robot World Cup Soccer Games.Middle-Size League requires the use of autonomous robots. Asthe important modules to access to environmental information, the vision system performance directly affects the overall performance of robot.Nowadays, the most of the vision systemuse color information to identify the field environment. Technical Challenge in Middle-Size Leaguetest the ability of identifying color-undefined ball,further request the robustness and real-time toidentification and Location System.In the background of the middle-Size League in the Robot World Cup Soccer Games, Our key research is achieving the robot soccer sidentificationof color-undefined ball in givenconfiguration or a small amount of hardware upgrade. Tohigh quality of equidistant panoramic image, extractits edge and process it by mathematical morphology. And then identify and locate the color-undefined ball by Hough transform and HIS location system. Ifpanoramic image is low quality and serious distortion,using edge matching and color histogram segmentation can also maintain the recognition rate.