Abstract: The system structure and related key technologies of the robot used to inspect the aircraft skin defects are studied.The general structure and key components which enable the climbing robot to maneuver over the nonsymmetrical exterior surface of an aircraft are presented.By analyzing the functions of absorption,motion and inspection during robot operation,the structures of the perception and control systems are presented.A path tracking control algorithm based on trajectory linear control(TLC) is designed to compensate errors caused by the driving components.The vision-based classification and identification algorithms for rivet crack defects are provided.The validity of the algorithms is verified by experiments.