Research On Robot Palletizing System Based On Embedded Machine Vision

Since the birth of the industrial robot, it plays an important role in human industry. From early special robot to today’s open, modular, intelligent robot, it has made a tremendous advance. Palletizing robot solves problems like low stacking load, high labor costs, low efficiency and is widely used in modern industry. With the improvement of industrial automation level, teaching robots can not meet the needs of industrial production. However, in traditional machine vision system, there is certain gap in real-time demand, which set a higher requirement in it.Based on characteristics of palletizing robot and industrial demand, this paper used embedded system to achieve machine vision in order to meet the real-time requirement and constructed a embedded machine vision palletizing robot system based on six joint robot. First, the paper described the purpose and significance of the research, introduced current research on machine vision system, embedded system and palletizing robot. Second, completed the camera demarcation analysis by Zhang Zhengyou camera calibration method. Third, analysed and simulated different image acquisition and processing algorithm to complete the identificaition of the workpiece position and other operations, then simulated in MATLAB. realized in DSP embedded processsors. Fourth, transplanted the algorithms to DSP and realized recognition and localization of workpiece. Finally, transfered recognition results to the host machine, providing location information and operational basis for subsequent palletizing processes.