摘要
大多数的遥感图像存在视觉对比度低、分辨率低的缺点,因而在对遥感图像分析之前,通常都需要通过遥感图像增强技术对图像进行增强处理。当前的图像增强方法有很多种,文中引入了一种新颖的图像增强方法,即多尺度Retinex(MSR)算法。这种增强技术尤其对能见度差、分辨率低的图像有很好的效果,因此适合于对遥感图像的增强处理。此外,还引入了几种常用的经典图像增强方法,如直方图均衡法等。为了证实所引入的MSR算法对遥感图像的增强效果优于其他的增强方法,在实验中,将经典的图像增强方法和MSR算法分别应用于增强一幅典型的遥感图像并对比增强后的增强效果。实验结果表明MSR算法在对遥感图像的增强中可以取得满意的效果并且优于其他的图像增强方法。
As most of the remote sensing images have little visual contrast and poor resolution, thus, the remote sensing image enhancement technique is almost always applied before the remote sensing images are analyzed. Numerous image enhancement techniques are available, In this study, a novel image enhancement technique,multi - scale Retinex （MSR） algorithm, is introduced. This enhancement technique is effective especially for images with poor visibility or low resolution. Therefore this technique is very suitable for enhancing remote sensing images. Besides, several classic image enhancement techniques, say, histogram equalization, are also introduced in this study. To illustrate the effectiveness and superiority of the MSR technique on renaote sensing image enhancement, the technique is compared with these classic enhancement techniques by enhancing a remote sensing image with poor visibility and little visual contrast. The results show that MSR technique is satisfying and superior over the other enhancement techniques.
出处
《计算机技术与发展》
2008年第2期70-73,共4页
Computer Technology and Development
基金
安徽省自然科学基金资助项目(070412059)
安徽大学研究生创新计划项目(20073050)
