Research on Ripple Algorithm in Automatic Target Recognition

Automatic target recognition (ATR) is increasingly important in a wide range of video and image applications, such as machine vision and security monitoring. In order to reduce the complexity and time lag in unmanned aircraft vehicle system (UAVS) for signals surveillance, force protection and strike, this paper proposes a novel ripple algorithm for ATR. This algorithm uses concentric circles to extract features of object and the features are quasi invariant to target translations, rotations and scaling. Through outdoor experimentation, it is proven that the ripple algorithm is correct and effective with high-speed in program running time. The recognition rate is more than 90% and the code runs on the order much faster than the implementation based on the traditional algorithms.