Research On Contrast Enhancement For High Resolution Remote Sensing Image

Remote-sensing satellites can quickly acquire large-scale surface images with rich content.The images are an objective reflection of the real state of the earth's surface.They have been widely used in military detectives,emergency rescue,agricultural production estimation,and ecological monitoring.However,the images have contrast decreased,uneven illumination and other issues due to the characteristics of imaging equipment,cloud,fog,water vapor and other factors,affecting the visual effects of images and the capabilities of information interpretation,and even limiting the application of remote sensing images.It is an important research topic in the field of remote sensing image processing to improve image quality,enhance image contrast,and increase interpretability of information.In this paper,the contrast enhancement of remote sensing image is studied in depth.Firstly,the basic theory of image enhancement is analyzed and summarized.Here,we have studied several mainstream image enhancement methods,such as Histogram Equalization(HE)in spatial domain,Brightness-preserving Bi-Histogram Equalization(BBHE),Adaptive Gamma Correction with Weighting Distribution(AGCWD),Gradient and Intensity Histogram Equalization(GIHE),and Histogram Compacting Transformation with Linear Stretch(HCTLS).And the advantages and disadvantages of these methods are summarized.In order to effectively evaluate the image,the commonly used index system for image evaluation is also analyzed.According to the significance of each index and the general evaluation method of the remote sensing industry,this paper uses a combination of subjective evaluation and objective evaluation to evaluate the image quality.On the basis of summarizing and analyzing the existing research results,this paper proposes a detail enhancement algorithm for contrast enhancement of remote sensing images.This algorithm improves the problem of local detail loss and contrast reduction in some regions of Histogram Compaction Transform(HCT).First,by comparing the ratio count matrix and the gradient information of the original image and the enhanced image by Histogram Compact Transform Linear Stretching(HCTLS),the detail loss region is preliminarily determined.Secondly,using morphological hole filling,area filtering and other methods,the detail loss region is further determined precisely.Then,the region is linearly reconstructed utilizing the original image.Finally,the boundary of the image re-assignment region is smoothed,so that the transition of the boundary is natural and has no artificial trace.The results of experiments and evaluations show that the proposed algorithm in this paper maintains the local contrast and has more details when the overall contrast is improved compared with HCTLS.Uneven illumination is a common problem in remote sensing images,especially in aerial remote sensing images.The overall contrast enhancement method is often unsatisfactory for this type of image processing.The HCTLS contrast enhancement method has a good overall enhancement effect on a uniform-illumination image,but still cannot handle images with uneven illumination well.To solve this problem,this paper adds image dodging process to overcome the influence of uneven illumination factors.The experimental and comparison results show that the improved HCTLS algorithm can effectively improve the performance on this type of image by adding the MASK dodging algorithm.