p.1205
p.1209
p.1213
p.1219
p.1224
p.1233
p.1239
p.1247
p.1251
Research on Method for Massive Pixel-Level Remote Sensing Image Processing Based on Hadoop
Abstract:
Recently, global change research has reflected the great challenge of massive distributed remote sensing image processing. Faced with such challenge, massive pixel-level remote sensing image processing reconstruction based on Hadoop is proposed, which focuses on the support of data format and the design of paralle computing. In order to support a variety of formats of remote sensing images and simplify the process of data parse, the processing flow transforms the remote sensing image into image information in binary format, as well as metadata information in xml format. Compared with converting to text format, there are two advantages for this conversion, reducing the amount of data after converted and remaining metadata information. To avoid MapReduce parallel computing performance interference caused by the algorithmic complexity, remote sensing image point operation is selected to do research about the design of parallel computing. The experimental results show that the proposed method has good scalability in the distributed Hadoop environment, along with the changing of the data quantity.
Info:
Periodical:
Pages:
1224-1230
Citation:
Online since:
July 2013
Authors:
Keywords:
Price:
Сopyright:
© 2013 Trans Tech Publications Ltd. All Rights Reserved
Share:
Citation: