动态方向加权二维多级中值滤波的图像处理

在图像解译和识别过程中,经常需要对图像进行增强处理,以提取重要的信息。结合机器视觉、图像处理等图像增强方法,根据二维多级中值滤波原理,改进该算法,由于二维多级中值滤波在图象处理中是对图像进行固定的中值滤波处理,不能随着图像的复杂性进行变化处理,方向性不强,窗口大小不变,利用动态方向加权二维多级中值滤波方法进行图像增强处理,根据图像复杂的变化动态调整窗口大小和权值,该动态加权处理增强方法突出了视觉处理目的,达到了满意的视觉增强效果。图像处理前后的相似系数达到0.868 3,在原始图像比较复杂的情况下,可以改善后期图像处理过程,提高图像分析(解译)的正确性和有效性。

Directional Weighted Two-Dimensional Multi-step Median Filtering in Image Processing

Enhancement of remote sensing image is necessary in the process of interpretation and recognition to extract important information.In this paper,combined with some existing method of image enhancement in machine vision and image processing,we advanced an improved arithmetic according to the principal of two-dimensional multi-step median filtering.Because two-dimensional multi-step median filtering is a fixed median filter and it does not change along the complexity of image,we used dynamic directional weighted two-dimensional multi-step median filtering to enhance the image,which dynamiced change window size and adjust weight value based on complicated variation of the image.This method emphasizes the purpose of vision processing,and obtains satisfied effect,similar coefficient arrived at 0.868 3 between before and after the image precession.Moreover,even to the complex original image,this method can improve the validation of post image processing,such as image analysis of interpretation and provide better enhanced image for remote sensing interpretation.