The Obstacle Avoidance for UAV Based on Improved Frames Difference and Optical Flow

Abstract

Abstract-As to the outdoor real-time change has a bad influence on the obstacle detection and avoidance for UAV with monocular-version, this paper propose a moving obstacle detection algorithm based on three frames subtraction with adaptive iterative threshold and optical flow. Based on improved three frames subtraction, it can quickly separate the moving object from the UAV's forward version and simplify the complexity of the optical flow algorithm. By using optical flow method in the zone where moving obstacles detected, which effectively decreases the influence in brightness constant and calculates the closer obstacle to avoidance. Through the experiment, results show that this method can detect the moving object and calculate the obstacle need to avoidance in good real-time performance and robustness.