基于Shearlet变换的遥感图像增强

遥感图像具有整体亮度偏暗、对比度较低和目标与 背景区分不明显的特点,遥感图像增强技术对于改善图像的对比度、突出某些局部细节等起 着积极的作用,图像的多尺度系统已经成功应用在图像处理中。经典的多尺度系统曲波(Cur verlet)变换、 轮廓波(Contourlet)变换无法将连续性与数字世界进行统一处理,而Shearlet变换是目前多 尺度领域 内唯一满足这一性质同时还提供对图像的最优稀疏表示的多尺度系统。本文提出一种基于Sh earlet变换的遥感图像增强 算法,首先将遥感图像进行Shearlet分解,然后对Shearlet变换产生的低频系数进行模糊对 比增强,对各尺度各 方向的高频系数进行模糊增强。实验结果表明,本文算法增强的结果,在主观上能获得很好 的视觉效果,客观指标中图像熵和均值有了大幅度的提升。

Remote sensing image enhancement based on Shearlet transform

Remote sensing image with the characte ristics of overall dim brightness,low-contrast and no obvious distinction betwe en th e target and background,makes the remote sensing image enhancement technology pl ay an active role for improving the contrast of the image and highlighting some loc al details.At present, the image multi-scale system has gained successful applications in image proces sing,wavelet transform, Curverlet transform,Contourlet transform and some improved algorithms based on t hem.One of the most common shortcomings of the frameworks of the system Curverlet and Co ntourlet is lacking of providing a unified treatment between the continuity and the digital world.S hearlet system is the only one which satisfies this property,yet still optimally delivers sparse approxi mations of images.Based on Shearlet transform,we propose a new algorithm for remote sensing image enha ncement.First,the remote sensing images are transformed into the low frequency coefficients and high freq uent coefficients by Shearlet transform;second,we use fuzzy contrast enhancement on the low frequency coefficients of Shearlet transform;third, we use fuzzy enhancement on the high frequency coefficients of each scale and di rection.Experimental results show that in the subjective side,we obtain a good visual effect,and objectivel y,the image entropy and mean have a significant promotion.