| The images display is very important to RS image processing and is easily affected by many factors,such as noises,atmosphere,surface objects interaction,etc.These factors reduced the quality of the image display and had influence on image understanding and analysis of RS images.Image enhancement can improve image contrast and suppress noises,emphasized the global or local information in images,but were unable to highlight information of specified object.This paper presented a new method for RS images aimed at enhancing specified objects based on its spectral information,improving image visual effects and the display of specified objects at the same time.The paper combined methods of spectral mapping,color enhancement,image filtering and dynamic range adjustment to construct the process of color enhancement for RS images based on the spectral information of specified objects,and achieved automatic image enhancement for specified objects.Taking Taihu Lake blue-green algal blooms as object,select 10 typical RS images to test and verify the method.Research results and conclusions are as follows:(1)Object Spectrum and Spectral MappingStudied the relationship between the object spectrum and bands of RS images.Analyzed the advantages and disadvantages of current spectral mapping methods and tested these methods.(2)Specified feature information enhancementEnhance RS images based on color transform.According to the spectrum of object,weighted the color of specified features by normal distribution,and reinforce image detail and brightness by spectral mapping and bilateral filtering.The final step is adjusting the dynamic range of images and obtaining the enhanced color images.(3)Method ValidationSelect Landsat 8 OLI image on July 23,2013 to enhance the information of algal blooms based on the spectral of blue-green blooms measured in the field.The result indicate that Matched Filtering can highlight the brightness of bloom and the process and method presented can strengthen information of blooms and suppress the effects of clouds and other factors.Compared with the typical image enhancement,images enhanced by proposed method has relative color stability without the influence of clouds and make image processing and interpretation easier.Discussing whether the proposed method is affected by atmospheric correction or not,the result shows that this method is not affected by atmospheric correction and easier for automation application.10 Landsat images with different weather conditions of TaiHu Lake were used to verify the method comparatively.The result shows that this method has fine applicability.In conclusion,the paper presents a method and workflow that has good color enhancement effect for specified feature and achieve automatic color enhancement.This method improves the efficiency of image processing and analysis and contributes to applications of remote sensing. |
PDF Full Text Request