Research On Remote Sensing Image Color Equalization Method Based On Block Adjustment

Satellite remote sensing image is an important data source for obtaining global geographic information.However,due to the influence of various factors such as the imaging time of the optical remote sensing image,the lighting conditions during imaging,and the different features and terrain of each scene image,color inconsistencies between images can be caused.This phenomenon is a direct factor that affects the production quality of subsequent digital products.Even if the image adopts high-precision geometric positioning and geometric stitching methods,due to the color inconsistency between images,a clear image dividing line will still appear.If it is not processed,It is difficult to achieve the "one picture" effect of the measurement area visually.Therefore,this paper focuses on the method of color equalization between images,and the main work is as follows:1)This article summarizes the current research status of image color equalization algorithms,studies and discusses the existing dodging algorithm for single image and color equalization algorithm for multi-view images.In this paper,the advantages and disadvantages of various existing algorithms are compared and analyzed through experiments.According to the experimental results,a better and more reliable dodging algorithm is selected as the dodging processing algorithm before the color equalization algorithm in this paper.2)This paper proposes a color equalization algorithm for satellite remote sensing images based on block adjustment.Firstly,the overlapping areas of the image are divided into uniform grids.The Histogram of Oriented Gradient(HOG)algorithm is used to extract the features of each grid and filter the corresponding grids based on the gradient distance and the 3 times median error of the difference between their gray values.To solve the problem of large time overhead in the process of extracting the grid,Open Mp-based multi-thread parallelism is used to speed up the calculation.Then,the average gray value of the filtered corresponding grid is used as the adjustment variable,and the quadratic polynomial is used as the adjustment model,virtual control is used to solve the problem of uneven distribution of pixel values,The overall average gray value of the overlapping area is used as the constraint condition of the weakly connected area,and the color transformation parameters of each image in the measurement area are solved as a whole.Finally,the obtained parameters are used for color correction for each scene image to obtain a color balanced image sequence.3)This paper compares the color equalization algorithm based on the block adjustment with the traditional color equalization algorithm.The results show that the algorithm in this paper is superior to the Wallis filter algorithm and histogram matching algorithm in terms of visual effects and image quality maintenance.The algorithm in this paper fully considers the color difference of the overlapping areas of the images,and solves the overall color transformation parameters of the images.After processing,the color transition of the overlapping areas naturally requires no further processing of the overlapping areas.The results of the algorithm of this paper in the application of large-scale image color equalization show that the processed images have similar color information visually,there is no obvious stitching line after stitching,the color transition at the stitching is natural,and the overall brightness of each channel of the image is more balanced than before processing.There is no obvious loss of image quality,which proves the feasibility and effectiveness of this algorithm.Finally,the algorithm in this paper is applied to the color equalization processing of Landsat5 true-color images across the country,and good results are obtained.