Anti-disturbance grabbing of underwater robot based on retinex image enhancement

摘要

Underwater Vehicle Manipulator System (UVMS) has the characteristics of weak illumination, strong nonlinearity and many disturbances when operating in underwater environment. It is difficult to control. In this paper, after constructing the mathematical model of manipulator based on task space, a fusion image enhancement algorithm based on Retinex theory is proposed for the problem of low illuminance and blurred target recognition. The bilateral filtering process, color pre-correction, estimated reflectivity and dynamics are proposed. The compensation, color compensation and other processes realize the fine perception of underwater targets; using the extended Kalman filter (EKF) and the spatial feedback linear transformation matrix to construct the robot anti-interference controller to compensate the internal and external uncertainty disturbances and achieve underwater stability operation; The target ranging method is used to solve the joint angles of the manipulators, and the tracking and grasping of the target objects are completed. Underwater experiments show that the algorithm can not only improve the dynamic stability of the UVMS system, but also ensure the accuracy of underwater robot grabbing and optimize system control performance.