基于峰谷点特征的红外目标自动识别方法

Algorithm for Automatic Infrared Target Recognition Based on Peak Valley

研究了红外目标图像峰谷点特征提取的方法，提出了一种目标识别相似性度量准则，证明了这种相似性度量在尺度和小角度旋转等变化情况下的不变性，并在此基础上提出了一种基于目标峰谷点特征的自动目标识别算法．真实场景下长波红外目标图像实验证明了该算法的正确性和有效性．

A method for extracting peak valley features of an infrared target image is investigated. A similarity measurement criterion for the target recognition is defined, and demonstrated its invariants under scaling and rotation with the small angle. An infrared target recognition algorithm based on peak valley features is presented and tested on real long wave infrared images. The experimental results have proved its validity.