Full-Text Search:
Home|Journal Papers|About CNKI|User Service|FAQ|Contact Us|中文
《Science of Surveying and Mapping》 2016-08
Add to Favorite Get Latest Update

An improved parallel remote sensing ISODATA algorithm

XIA Huiyu;Changjiang Nanjing Waterway Bureau;College of Remote Sensing and Information Engineering,Wuhan University;  
The conventional ISODATA classification algorithm becomes very time-consuming as the volume of remote sensing data increases.Parallel computing technique is an effective approach for addressing this bottleneck issue.To address the shortcoming of current parallel ISODATA algorithm based on MapReduce model,a scalable parallel ISODATA algorithm was proposed in this paper.The improved algorithm overcame the bottleneck issue by using parallel global subsampling method for data partitioning,centroids filtering algorithm for refining intermediate clustering representatives and single-pass final clusters mapping algorithm for getting the final clustering results.The experimental results showed that the improved algorithm was more efficient than current parallel ISODATA algorithm in processing the same scale of remote sensing computing,and it also obtained more accurate clustering results in processing larger size of remote sensing image.
【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