低对比度目标因其灰度对比度低、边缘模糊等缺点,使得联合变换相关器无法将其从混杂的背景图像中辨别出来,达到成功识别的目的.针对这一问题,采用了基于Curvelet变换的图像增强算法对目标联合图像进行处理.作为超小波分析范畴的Curvelet变换,因具有极强的方向性,成为比小波变换更适合分析和理解图像特征的多分辨率分析工具.文中采用不同的方法分别调整了Curvelet变换后的高、低频系数,增强了目标的灰度对比度和边缘信息.以低对比度坦克图像为例,增强后的目标对比度由原来的4.16％提高至29.37％.计算机模拟和光学相关实验结果均表明,增强后的联合图像获得了明亮的相关点对,成功实现了低对比度坦克的自动识别.%The target in low contrast image cannot be discriminated from cluttered scene by using joint transform correlator as its indistinct edge. Aimed at this problem, an image enhancement based on Curve-let transform was adopted to process the target' s joint image. In the wavelet analysis, the Curvelet transform is a better multi-resolution tool to analyze and understand the image feature. Both gray contrast and edge can be enhanced by adjusting the high and low frequency coefficients of Curvelet transform with different methods. As an example, a tank' s image with low contrast was processed by using Curvelet transform. The gray contrast of original image was increased from 4. 16% to 29. 37% . The computer simulation and optical experiment results show that a bright pair of correlation peaks are obtained after enhancing the image, and the recognition for the target in low contrast image is realized successfully.
展开▼
