Landslide Fusion Detection Based on Features on Spatial Shape and Spectrum in Massive Data of Aerospace Remote Sensing

EasyChair Preprint 885, version 1

Abstract

Aiming at the problems concerning poor and ineffective detection on landslide in massive data by aerospace remote sensing, landslide fusion detection method by spatial and spectral features based on neural network is proposed in this paper. A fusion detection model based on neural network and the typical fundamental spatial shape model of for landslide are established. The detection accuracy of landslide in remote sensing image is improved by SIFT algorithm feature matching and transformation, spatial shape feature similarity comparison. The rapid landslide detection and accurate extraction of disaster information is achieved by the fusion detection model, when disasters break out in large area. The proposed method is verified by application experiments in several aerospace remote sensing data. The experimental results show that our proposed method is superior to many other contrast algorithms, which improves the accuracy of landslide detection.

Keyphrases: Big data of aerospace remote sensing, Remote sensing for disasters, Spatial and spectral features modeling, neural network

@booklet{EasyChair:885,
  author    = {Chen Shanjing and Xiang Chaocan and Kang Qing and Wang Zhenggang and Shen Zhiqiang and Zhou Ruochong},
  title     = {Landslide Fusion Detection Based on Features on Spatial Shape and Spectrum  in Massive Data of Aerospace Remote Sensing},
  howpublished = {EasyChair Preprint 885},
  year      = {EasyChair, 2019}}