p.2300
p.2304
p.2308
p.2312
p.2318
p.2323
p.2329
p.2333
p.2337
High Resolution RS Image Industrial Solid Wastes Extraction Based on SVM
Abstract:
In order to accurately extract various types of industrial solid wastes from high resolution RS images, a industrial solid wastes feature fast extraction algorithm was proposed based on SVM. The reasonable image pretreatment was conducted by anisotropic diffusion filtering firstly. It is because that high resolution RS image contains abundant information and industrial solid wastes heap was very complex, we proposed the classification algorithm based on 1-v-1 which could extract multi-class industrial solid wastes fast and accurately at once. The new algorithm improved both efficiency and accuracy of industrial solid wastes recognition. The experimental results show that the industrial solid wastes feature recognition of SVM has better advantages than conventional methods. The new algorithm can recognize not only shape features of industrial solid wastes heap but also its material and type and it is constructed to recognize multi-class industrial solid wastes with higher operation efficiency.
Info:
Periodical:
Pages:
2318-2322
Citation:
Online since:
March 2014
Authors:
Price:
Permissions: