DCT域遥感影像的自适应遗传优化增强

Remote sensing image enhancement by adaptive genetic algorithms optimization in the DCT domain

本文在DCT域中提出了一种基于遗传算法的遥感影像增强技术,针对DCT域遥感影像增强参数需人工设定的问题,利用自适应遗传算法对其进行优化选取:设计了一种考虑偏差的适应度函数以驱动遗传算法进化,通过无分块思想的使用实现了无人工干预、无块效应的遥感影像增强。该方法不但保持了遥感影像的原始色调而且显著提升了遥感影像方差、平均梯度、信息熵等质量测度,有效改善了遥感影像的视觉效果,便于后续影像处理及应用。

The paper studied on the realization of remote sensing image enhancement by Genetic Algorithms in the DCT domain. The parameter of image enhancement in the DCT domain requires artificial setting, so it optimized the parameter with adaptive Genetic Algorithms automatically. In details, it designed a novel fitness function considering deviation index to drive evolution of Genetic Algorithms, and managed to realize remote sensing image enhancement without artificial affection and blocking artifacts by virtue of no-blocks theory in the DCT domain. The method discussed here could improve the contrast of intensity, gradient and entropy including keeping hue of remote sensing image. Besides, the method could also improve the visible effects of images as well as convenience for further image processing.