The Research On The Vision System Of The Autonomous Robot Soccer

Robot soccer match is a kind of high technology match in the world.It combines the technique from information,automation,mechatronics,communication and AI,and provides a platform to test all these skills.With the development of robotic technique, autonomous robot soccer system(ARSS) has become one of the hot focuses in this field.Vision system is an important part of ARSS.The assignment of vision system is to grab the image of ground,then analyze and dispose the image,accordingly get the motion state information of objects(ball and both robots).Finally it supplies strategy system with all the information,used to command the robots.The thesis discusses the research of the vision system of the ARSS,constructed the hardware platform of the vision system with DM6437,developed the software of the vision sytem with CCS and DSP/BIOS,achieved real-time digital image acquisition and pretreatment.Based on the experi- merits of Some algorithms of moving target decting which is backgroud subtraction,frame difference,edge decting algorithm and colour decting algorithm.The paper proposed a new target decting algorithm which is based on the combination of the frame fifferencd and edge decting.We used this algorithm to achieved the real-time target detecting.Besides,we constructed the camera model of the single Camera and gained the picking-up of the gravity of the target,achieved the target location.Practices demonstrate the algorithm can detect and locate the target commendably. Besides,It sets up a platform for upper research with the embedded hardware.