Research On Single-Polarization SAR Image Pseudo-Color Enhancement Method Based On Regularization

Synthetic Aperture Radar (SAR) as an all-day, all-weather microwave remote sensing imaging radar has unique advantages of military reconnaissance and earth observation. SAR is a gray image. The human eye is not sensitive to the gray information, only distinguish dozens of gray levels. But it can distinguish thousands of color, color image more favorable than the grays image in visual interpretation and the target identified. Therefore, study the SAR image pseudo-color enhancement method.The main contents are as follows:(1) According to a problem of noise in nonlinear coding method, introduced the region growing is improved; the histogram mapping color migration color is not rich, the introduction of clustering is improved; for Reinhard color transfer party to ignore details information and the fixed proportion channel cannot adjust the color, carries on the analysis from the theory, introduces the adjustment coefficient was improved. Through the experiment. The experiments proved that the method is effective.(2)The pepper applied the regularization method to the single polarimetric SAR image pseudo color enhancement for the problem of noise and not rich colors in SAR pseudo-color enhanced image. In the HSI color space, the paper constructs H, S, I three channel information by image processing method. SAR image denoise to post as I component. The Image coding with rainbow G band function as a H component. The regularization image as a S component.Among them, according to the SAR regularization image enhancement problem,the study the norm regularization model, Lorentz regularization model. Using Radarsat-2 data in pseudo-color enhancement experiments to prove the effectiveness of the enhanced approach.(3)The pseudo-color enhancement images used in image classification and interpretation. The pseudo-color enhancement results are carried out supervised and unsupervised classification, using Kappa coefficients, the overall classification accuracy as the evaluation indexes. The paper proves the pseudo-color enhancement based on regularization method is effective. And visual interpretates the image to prove the value of the method in the interpretate.