Severals Correct Medium Resampling Method Research

Remote sensing image geometric deformation in the imaging process is difficult toavoid, there are three schemes of remote sensing image geometric distortion correctionprocessing: system correction, using control points correction and mixed correction.Mixture correction is a staged correction scheme, it is often used in space remotesensing image, at first use the geometric distortion theory formula system of geometriccorrection, i.e. geometric correction, then use the ground control points on the geometricdistortion to do further correction, i.e. geometric precision correction. Mixturecorrection is the most commonly used scheme, this paper is mainly about the influencesof different resampling methods for space-borne SAR image geometric correctionaccuracy in remote sensing image geometric correction.The resampking is an important processing process in the geometric correction, itis directly affects the accuracy of image geometric correction, geometric correction ofand computational efficiency, so the selection of proper resampling method ofgeometric correction is a significant work.This paper first briefly introduces the geometry distortion of the image, geometriccorrection principle, method, then describes the process of resampling in details. In thefull understanding of the nearest neighbor method, bilinear interpolation, threeconvolution method, and Sinc interpolation, use the C language to the four interpolationmethods of computer programming and a make detailed description of the four methodsof the realization process.Finally, by using the simulation of SAR images to achieve four above interpolationmethod was verified by simulation, is presented by using these methods to achievethrough the geometric correction of SAR images. Finally we arrive at the followingconclusions: the resampling process is geometric correction process is an extremelyimportant step; it can realize data resampling procedure can be divided into two steps:one is heavy sampling point conjugate point location, two is a heavy sampling pointpixel brightness value is calculated; the resampling method is through the correction ofspace the point for original space conjugate points to achieve correction, it can ensure the correction of space grid pixel points evenly distributed output; the resampling eachmethod has its advantages and disadvantages, the nearest neighbor method is simple,fast processing speed, but the precision is not high, is easy to appear the discontinuousphenomenon; bilinear interpolation image continuous and high precision, but thismethod has a low pass filter to suppress high frequency components of nature, make theimage appear fuzzy phenomenon; the three convolution interpolation accuracy value ishigh, with image equalization and clear effect, can get the high quality image, but largeamounts of calculation, and destroyed the original data;" neighbor weight method" with" the nearest neighbor method" and " bilinear interpolation " part of the advantages, there-sampling the image has a good sampling effect and quality; sinc interpolation Valuemethod is the optimum interpolation function, but the complex calculation process,large amounts of calculation, the practical application of specific ways depends on thespecific circumstances...