Underwater Target Tracking Algorithm Based on Optical Flow

AbstractBecause of uneven underwater light and high contrast and background noise, the underwater target detection and location need higher requirements. In this paper, the optical flow target tracking algorithm based on Kanade-Lucas-Tomasi (KLT) algorithm is implemented on OpenCV platform, and the algorithm is verified in the water pipeline intelligent inspection competition environment. Aiming at underwater target detection and location, this algorithm uses optical flow method to achieve high detection certainty and fast operation speed of image frame difference method. Using the incremental PID control, the closed-loop control of underwater vehicle underwater motion is realized with a more stable control error.KeywordsOptical flow methodLocationUnderwater robotLucas-Kanade optical flow