Font Size: a A A

The HJ-1 CCD Image Geometric Correction AutoGCs Software Performance Evaluation Based On The TM Images

Posted on:2016-09-10Degree:MasterType:Thesis
Country:ChinaCandidate:W Y ZhaoFull Text:PDF
GTID:2180330461972360Subject:Surveying and Mapping project
Abstract/Summary:PDF Full Text Request
In order to fulfill the global HJ-1 CCD image automatic batch processing,Shan XiaoJun,a remote sensing researcher at the Chinese academy of sciences designed and developed the accurate geometric correction software HJ - 1 CCD AutoGCs based on the characteristics of the HJ - 1 CCD images.Since the software has not been tested by a large amount of data after completion, it’s unable to confirm the resolution of the correction.To guarantee the accuracy and availability of the follow-up products,we use the software AutoGCs to automatically geometric correct the 2009 and 2010 HJ-1 CCD image with TM images as reference.Then choose check points on HJ-1 CCD image and the corresponding TM image,quantitatively evaluate the stability,efficiency,accuracy and applicability of the software on the basis of the choosen check points,realised the quantitatively evaluation of HJ - 1 CCD AutoGCs.And put forward the software improvement suggestions through analysising the problems in data processing.This paper briefly introduces the geometry correction affiliate programs checklogandmove and MonitorCorrectHJImage,solved the problems about the file path and software stability while the processing of AutoGCs software automatically accurate geometric correction,and write a program to count the quantity and type of the error files.While guaranting the integrity of original image data,the AutoGCs automatic accurate geometric correction software can properly deal with more than 98% of the data;the average time for processing one scene 4-band CCD image is 2.5 min.While ignoring the error with manual check points,the average point accuracy of river and ridge intersection point located in the mountains can be up to 5 pixel,the average point accuracy of road intersection and building inflection point in the mountains can be up to 2.6 pixel,the average point accuracy of river intersection point lacated in plains can be up to 3.9 pixel,the average point accuracy of road intersection and buildings in plains can be up to 2.5 pixel,meets the accuracy requirement of practical application.Since the variable target is easily affected by factors such as water and forest cover,when handling the huge time phase difference reference image and target image,the correction precision of software will be affected in some degree.Since the software has been fully packed,we can’t intervene the process of software compiling.when we encounter problems while data processing,the affiliated programs described herein can be used for auxiliary.In dealing with the variable target,we can gradually establish a complete reference library of TM image,containing full time phase TM images and update it.Try to choose the nearest time phase between reference image and target image in order to improve the precision of image geometric correction.At present the software can be promoted to some extent.
Keywords/Search Tags:HJ-1 CCD image, TM image, AutoGCs, software performance evaluation
PDF Full Text Request
Related items