Research On Underwater Robotic Arm Grasping Technology Based On Visual Dynamic Tracking

As one of the components of underwater robots,underwater robotic arms play a crucial role in underwater tasks.However,there are still problems with low accuracy,long operation time,and low intelligence in underwater robotic arm operations.The use of underwater cameras to collect underwater visual information and the application of visual dynamic tracking in the work tasks of underwater robotic arms can greatly improve the accuracy and intelligence of underwater robotic arm operations.Therefore,the research on underwater robotic arm grasping technology based on visual dynamic tracking has important practical significance in the application of underwater robot engineering.The main research content of this article is as follows:Firstly,to address the issues of camera lens distortion,visual degradation of underwater images,and underwater noise interference,Zhang Zhengyou’s chessboard calibration method is used in water to eliminate lens distortion.Additionally,underwater visual images are filtered and denoised to reduce the interference of underwater noise on visual feature details in the image.Finally,image enhancement methods are used to increase the color contrast of underwater images and improve the quality of underwater visual images,Prepare for visual based underwater target recognition and dynamic tracking in the following text.Secondly,the SURF algorithm is used to extract feature points of target objects from underwater images,and then template matching method is used to identify underwater target objects.Subsequently,visual tracking algorithms are used to track the target object.In response to the high accuracy and timeliness requirements of underwater operations,the characteristics of several visual tracking algorithms are compared.The extended Kalman filter method is used to visually dynamically track the underwater target and obtain its pose information.Subsequently,the binocular camera is selected to obtain underwater visual image information,and a robotic arm visual control system is built based on ROS.The extended Kalman filtering algorithm is used for visual tracking experiments in a laboratory pool,and simulation grasping experiments are conducted in Gazebo.The experimental results show that the extended Kalman filtering visual tracking algorithm can accurately obtain dynamic target pose information and achieve dynamic tracking and grasping of the robotic arm.Finally,the robotic arm visual tracking and grasping experimental platform is established in the laboratory,and after camera calibration and visual image preprocessing,an extended Kalman filter visual tracking algorithm is used for robotic arm tracking and grasping experiments to verify the feasibility and accuracy of the constructed visual tracking system and the adopted dynamic tracking and grasping method.The experimental results indicate that the application of visual dynamic tracking method in underwater robotic arm grasping is feasible.