Non-local mean value speckle suppression method for integral image number approximate semi-declining polarized SAR image

Abstract

The invention provides a non-local mean value speckle suppression method for an integral image number approximate semi-declining polarized SAR image. The method comprises the following steps: obtaining a full-polarized SAR image; setting a search window and a local block window for each central point in the full-polarized SAR image; constructing a ratio logarithm product distance; calculating an offset integral graph; calculating a distance between the center point of the center block and a local point through the integral graph, and recording the total distance of the offset of the center point; performing the operation until the local block distance in the search window of each pixel point in the full-polarization SAR image is calculated; and finally, calculating the weight of each pixelpoint according to the search window, then carrying out weighted average on the full-polarized SAR covariance matrix, estimating the pixel points of the whole image, and obtaining a corresponding image after speckle suppression. On the premise that the performance of the algorithm is not reduced, the operation time of the algorithm is obviously superior to that of an existing non-local mean valuealgorithm.

Description

Technical field [0001] BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to the technical field of coherent spots inhibiting methods, in particular,,,,,,,,,,,,,,,,,, Background technique [0002] Polarimetricsynthetic Aperture Radar, Polsar has been widely used in environmental monitoring, land resource monitoring and disaster assessment due to total work characteristics. However, due to the inherent coherent imaging mechanism of the POLSAR image, the POLSAR image is present in the coherent spots, and the coherent spots are difficult to interpret the POLSAR image. In the coherent spatient suppression method, the non-local mean is a very active method. In recent years, domestic and foreign scholars have been widely concerned. [0003] The non-local average value algorithm is due to the similarity of the block, and the similarity between the block is made, it has a large advantage in image denoising, but there is a huge amount of operation. [1][2] ...

AI Technical Summary

Problems solved by technology

Although the algorithm can achieve the purpose of reducing calculations to a certain extent, but the accuracy decreases to a certain extent

Cozzolino et al. [9] A fast non-local mean algorithm based on a lookup table is proposed. Because the lookup table needs to be quantized, there is a certain loss of precision.

Images