A Fast Approximation Method of Median Filtering for SAR Image

Abstract

The invention discloses a median-filtering rapid approximation algorithm of a SAR image. According to a characteristic that a multi-view SAR intensity image approximately obeys Gamma distribution, the intensity image is converted into an amplitude image form, through calculating a global maximum value and a minimum value of the amplitude image, an interval range where an amplitude image median is located is estimated, and based on the range, the amplitude image is quantized linearly to several gray levels, and then the gray level whose cumulative distribution function value is 0.5 is searched. When a number of pixels in the gray level is small, a mean value of the pixel intensity of the level is taken as a final result of the median filtering; otherwise, carrying out fine ordering on the intensity images of the pixel level. In the invention, a small range of an amplitude domain median is quantized, which is good for accurately positioning the median, and a calculated amount is basically irrelevant to a number of elements which need to be ordered so that a search calculation amount can be greatly reduced and a problem that large window and multi-temporal SAR image median filtering efficiency is low is solved.

Description

【Technical field】 [0001] The invention relates to the technical field of remote sensing image processing, and relates to a synthetic aperture radar (Synthetic Aperture Radar, referred to as SAR) image coherent speckle noise suppression method, specifically a median filter fast approximation method for SAR images. 【Background technique】 [0002] Synthetic aperture radar is not limited by factors such as weather, geography and time. It can perform high-resolution imaging on the ground and has certain penetrating power. Therefore, it is widely used in military reconnaissance, resource detection, environmental monitoring, surveying and mapping, and geographical conditions. Monitoring and other ground remote sensing applications. With the successful launch of many radar satellites, time-series radar images have become an important aspect of current remote sensing applications. [0003] Since SAR works based on the coherent imaging mechanism, serious speckle noise exists in the S...

AI Technical Summary

Problems solved by technology

[0007] The technical problem to be solved by the present invention is to provide a fast approximation method for median filtering of SAR images, that is, to solve the problem of low calculation efficiency of multi-temporal or large sliding window median filtering in SAR intensity images. The present invention is based on multi-view SAR The intensity image approximately obeys the characteristics of the Gamma distribution. First, set the sliding window parameters, convert the intensity image into an amplitude image, and count the maximum and minimum values ​​of the amplitude image, and then estimate the interval of the median value of the amplitude image, and based on this range The magnitude image is linearly quantized to fewer gray levels, in order to accurately locate the gray level of the median value of the magnitude image by calculating the sliding window distribution accumulation function, and effectively improve the calculation efficiency of the median filter

Images