提出了一种新的红外目标边缘检测算法,利用局部熵值分析法和过渡区提取相结合的方法,通过Top-hat与Bottom-hat变换,建立有效平均熵和对应灰度值的曲线关系,选择合适的阈值,在此基础上执行形态学闭合操作。仿真结果表明,该方法在去除不必要的细节的同时,能很好地提取目标的边缘信息。

A new algorithm for infrared target edge detection is presented, based on local entropy value method and the transition region extraction method. Through the Top-hat and the Bottom-hat transform, building a corresponding curve between effective average entropy and its gray value, selecting appropriate threshold, and finally morphological operators are executed to detect image edge. Simulation results show that the method can remove the unnecessary details, while efficiently extract target edges.