Path Planning For SCARA To Avoid Obstacle Based On Image

SCARA(Selective Compliance Assembly Robot Arm)is a type of industrial robots with a wide range of applications,mainly composed of three rotation axes and one translation axis.The application scenarios of SCARA include palletizing,welding,packaging and assembly.In order to ensure the efficiency of operation,it is usually required that the working space is free of obstacles,and then a fixed path is planned.With the increasing complexity of operation scenarios and the improvement of equipment compactness and intelligence,the demand for autonomous path planning capabilities and obstacle avoidance functions of SCARA has become more urgent.Combining machine vision technology,using a wide-angle camera to construct the image collecting system,through camera calibration and hand-eye calibration,the conversion from the pixel coordinate system to the world coordinate system is completed.The image obtained by the camera contains background information.In order to realize the extraction of the location information,the obstacle and the background need to be segmented.After comprehensively evaluating various image segmentation and edge extraction methods,the mean shift algorithm and Canny edge extraction algorithm are finally selected to complete the extraction of the location information.On this basis,to solve the problem of redundant nodes in the obstacle avoidance path planning process of the basic A* path-finding method,a vector-based intersection determination method is proposed,and the A* path-finding method is improved,and proposed according to the bow height error limit,smooth A*path-finding method is used.In addition,the application of Dynamic Window Approach(DWA)in obstacle avoidance path planning is also studied.Aiming at the shortcomings of DWA in the global path search ability,combined with the improved A* algorithm,a DWA integrating A* is proposed to realize more flexible local obstacle avoidance path planning.On the basis of the above research work,an extended edge location information is completed,and the safe distance in the workspace is converted to the pixel distance in the pixel coordinate system,which realizes the safe distance control of the obstacle avoidance path.A visual simulation environment for obstacle avoidance path planning algorithms is built based on Open CV,visual simulation of path planning is realized.The Ratio of Path Length(RPL)based on the obstacle avoidance path of the basic A* path-finding method is proposed as an evaluation index of different obstacle avoidance path planning algorithms.Finally,the quality of obstacle avoidance paths of different path planning algorithms is compared and analyzed.To verify the research content,a wide-angle camera was added to the experimental prototype platform to form a image collecting system and conduct obstacle avoidance path planning experiments.On the basis of camera calibration and hand-eye calibration,the extraction of extended edge location information on the working plane of the SCARA is completed through image processing.On this basis,the obstacle avoidance path planning capabilities of the improved A*path-finding method,the smooth A* path-finding method and the DWA integrating A* are verified.Experiments show that the built image collecting system can realize the extraction of location information,and the obstacle avoidance path planning algorithm adopted has high path planning quality.Image-based obstacle avoidance path planning technology for SCARA has high feasibility.