小波域声呐图像自适应增强

Adaptive sonar image enhancement in the wavelet domain

声呐图像受噪声污染严重、对比度低,给后期的定位识别带来不便,而传统的处理方法容易造成边缘模糊.针对这一问题,提出了一种图像自适应增强算法.该算法利用形态小波对声呐图像进行自适应的多分辨率分析,分别增强不同尺度上的信号或细节,通过多通道重构图像的加权实现去噪和对比度提高.仿真结果表明该算法快速有效,对高斯噪声和冲击性噪声都具有较好的鲁棒性,处理后的声呐图像边缘细节信息保留完好,得到了理想的增强效果.

Noise pollution in sonar image is significant, contrast is low, and this creates problems for object location and recognition. Moreover, traditional methods can easily fuzz edges. To deal with this problem, an adaptive image enhancement algorithm was proposed. This algorithm gives adaptive multiresolution decomposition of a sonar image with morphological wavelets, and enhances the signals or details in different scales separately, so denoising and contrast improvement can be performed by weighting reconstructed images from different channels. Simulation results showed that this algorithm is fast, effective and robust to Ganssian noise and impulsion noise. After processing, edge details of the sonar images are preserved better and significant improvements in the enhancement effect can be obtained.