光学相关复杂背景低对比度目标的识别研究

Low Contrast Target Recognition in Cluttered Background with Optical Correlation

摘要:

复杂背景的低对比度目标识别一直是目标探测跟踪的关键技术之一。针对光电混合联合变换相关器，以多小波变换为基础，应用GHM多小波对目标联合图像处理，并采用多尺度非线性方法，通过非线性函数有效地改变各尺度上的细节分量，从而改变图像对比度，达到增强目标图像的目的。对复杂背景的低对比度目标图像进行了大量计算机模拟实验和光学相关实验，结果表明，这种基于多小波图像增强算法，对复杂背景的低对比度目标能够获得明亮的相关点，提高低对比度目标的识别率。作为实例，文中给出航行中低对比度船的图像识别。

Abstract:

The optical correlation recognition of low contrast targets in cluttered background is one of key techniques in target detection and tracking. In this paper, based on the multi-wavelet transform, GHM was applied to process the targets joint image in the hybrid optoelectronic joint transform correlator (HOJTC). The multi-scale nonlinear enhancing method is adopted, which achieves the aim of image enhancement by changing detail components among different scales. Computer simulations and optical experiments of low contrast targets in cluttered background show that the image enhancement method based on multi-wavelet can acquire bright correlation points, and increase the recognition ratio of low contrast target. As an example, image recognition of low contrast ship on voyage was given.