ICESat-GLAS-based forest type classification using SVM

  • Authors:

  • Licun Li;Yanqiu Xing

  • Affiliations:

  • Center for Forest Operations and Environment of Northeast Forestry University, Harbin, China;Center for Forest Operations and Environment of Northeast Forestry University, Harbin, China

  • Venue:
  • AICI'11 Proceedings of the Third international conference on Artificial intelligence and computational intelligence - Volume Part III
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

In order to study the forest classification effect of large footprint lidar date we used SVM(support vector machine) method to analyze the ICESATGLAS (Ice, Cloud and Land Elevation Satellite - Geoscience Laser Altimeter system) date in WangQing Bureau, Jilin province. In analysis we first used IDL to convert the ICESAT-GLAS original binary data into ASCII format. Then we got a waveform by using matlab software. After we were corresponded the waveform data to the field investigation data in 2006 and 2007, we could get the forest types of the waveform figure. Then waveform parameters were extracted. We applied of the SVM classification method to analyze 62 groups of training sample and established a classification model. After that we used another 62 groups of test sample to test the classification model, the result shows that the SVM classification method can better distinguish the broadleaved forest between the coniferous forest. And the classification accuracy is 82.26%.