基于重构算法的一种微弱目标检测后处理方法

A Post Processing Method in Dim Target Detection Based on Reconstruction

重点研究了形态学重构算法在微弱目标检测和跟踪中的应用,主要体现在对原始图像的背景估计过程和检测后处理过程。对于经过背景估计和二值化后的图像,若直接进行跟踪算法,由于虚警点数目太多,会严重影响跟踪算法的效率,同时也给跟踪结果造成了负面效应。因此,考虑在二值化之后,跟踪之前,引进后处理过程,以消除虚警点。在已知目标大致尺寸的情况下,选用不同的两个标记重构,同时消除小于目标的虚警点和大于目标的虚假目标。

This paper researches the application of morphology reconstruction algorithm in dim point target detection and tracing, mainly in background estimation and post processing stage after detection. If the tracing algorithm is applied directly to a binary image after background estimation, as the great number of false targets, the efficiency will be influenced. In addition, this will bring some negative effects to tracing result. As a result, this paper considers introducing a post processing stage, after binary stage, before tracing. When some information of target size is known, double markers could be applied for reconstruction. This will eliminate the false targets, the size of which is larger or smaller than real ones.