Full-Text Search:
Home|Journal Papers|About CNKI|User Service|FAQ|Contact Us|中文
《Journal of Tianjin Normal University(Natural Science Edition)》 2018-02
Add to Favorite Get Latest Update

Improved algorithm for estimating noise in hyperspectral image

WANG Qian;LIU Pengfei;College of Geography and Environmental Science,Tianjin Normal University;Tianjin Geospatial Information Engineering Technology Research Center,Tianjin Normal University;  
To solve the problems of the block with spatial decorrelation often contains some object boundaries and much operand caused by repeated inverse derivation of matrix in spectral decorrelation,an improved spectral decorrelation algorithm(ISDOS) based on object segmentation was proposed,which could ensure the object in a block have similar spectrum,and the matrix inversion was demanded only one time to reduce the calculated amount significantly. This method was applied to es-timate noise in simulated image and Hyperion image respectively. The results show that on the one hand,the improved algo-rithm has faster computing speed,on the other hand,pixel spectral curve and its structural similarity of improved algorithm is closer to the simulated noise when the simulated image is estimated,and the noise spectral curve of improved algorithm fluctuates around 0 within a narrow range when Hyperion image is estimated,it illustrates that the improved algorithm is better corresponds to the general character of the true noise. Therefore,the improved ISDOS is proved to be effective.
【Fund】: 天津市科技计划资助项目(15ZCZDSF00390);; 天津师范大学博士基金资助项目(043135202-XB1701)
【CateGory Index】: TP751
Download(CAJ format) Download(PDF format)
CAJViewer7.0 supports all the CNKI file formats; AdobeReader only supports the PDF format.
©2006 Tsinghua Tongfang Knowledge Network Technology Co., Ltd.(Beijing)(TTKN) All rights reserved