Image uniformity and color uniformity method based on distributed framework

Abstract

The invention discloses an image uniformity and color uniformity method based on a distributed frame. The overall flow is as follows: 1, the global standard deviation and average value of the input image to be uniformity and the standard image are calculated; 2, the uniform color image is divided to be uniform according to the designate block size and the block parameters are calculated; 3, the calculated block parameters and the parameters required by Wallis filter formula are sent to the calculation node for calculation, and each block is processed to be homogenized and output in parallel. The invention is based on LLTS distributed framework and image read-write library under the framework, realizing the distributed processing and parallel output of the uniform color algorithm, it can effectively reduce the strong dependence of the algorithm on resources and improve the efficiency of the uniform color algorithm. It can solve the defects that the computational efficiency of the singlemachine version uniform color method is heavily dependent on the memory of a single machine and the performance of the processor, and can meet the needs of fast processing of high-resolution and large amount of data of remote sensing images.

Description

technical field [0001] The invention relates to an image processing method, in particular to an image uniform light and color uniform method based on a distributed framework. Background technique [0002] As the spatial resolution of remote sensing images increases to the meter or even sub-meter level, the data volume of multispectral images after single-scene fusion reaches tens of gigabytes. With such a large amount of data, it is impossible to meet the requirements of fast processing completion by relying on single-machine processing to uniformly light and color each fused image. Due to the obvious color difference between the images taken at different times, if they are stitched without processing, the stitched images will inevitably have uneven colors, which cannot meet the visual appearance and color requirements of the picture. [0003] The Wallis filter method is a common light and color uniform method. Its idea is to correct the uneven distribution of brightness, c...

AI Technical Summary

Problems solved by technology

However, the existing Wallis filtering method is mainly applied to stand-alone software through plug-ins, which is seriously affected by the operating environment resources and hinders the improvement of remote sensing images, especially the calculation speed of large-scale remote sensing images.

In addition, due to the limitation that one process can only operate one file, the calculation results can only be output to one file sequentially, which indirectly leads to an increase in processing time

Images