多尺度Retinex的低照度航空图像非线性增强

针对低照度航空遥感图像亮度低、对比度低,阴影区域目标难以准确判读的问题,提出了一种基于多尺度Retinex的图像非线性增强方法。该方法采用递归迭代提高高斯滤波处理速度;采用灰度极值去除线性拉伸,尽可能保留原始细节信息;采用动态范围调节因子控制输出图像对比度。实验结果表明,灰度极值去除比例、动态范围调节因子和最大环绕尺度等重要参数的取值变化对图象处理效果具有影响。给出了上述3种参数合适的取值范围,有效提升了低照度航空遥感图像的判读效果。

The non linear enhance image for low illumination aerial based on multi-scale Retinex

In order to solve the problem of low luminance and low contrast in low illumination aerial remote sensing image,which is difficult to interpret accurately in shadow area,a new method of nonlinear enhanced image based on multi-scale Retinex is proposed.This method adopt recursive iteration to improve the processing speed of Gaussian filter,applying the gray-scale extreme value to remove linear stretching to keep the original details as much as possible,using dynamic range adjusting factor to control the output image contrast.The experimental results show that the value change of the important parameters,such as gray-scale extremum ratio,dynamic range regulation factor and maximum surrounding scale,influence the processing of image.In this paper,the suitable range of the three parameters is given,and the interpretation effect of low illumination aerial remote sensing image is improved effectively.