Method for accurately identifying new landslide area based on multi-source time sequence remote sensing image

Abstract

The invention discloses a method for accurately identifying a new landslide area based on a multi-source time sequence remote sensing image. The method comprises the following steps: acquiring a slope map of a digital terrain model, a vegetation coverage index NDVI time sequence image and a surface radiation temperature time sequence image; calculating a simplified earth surface drought index TVDI based on an empirical parameter relationship between the earth surface temperature and the vegetation coverage index NDVI; tVDI values of remote sensing image files before and after a landslide are extracted to serve as time phase basic data before and after landslide area extraction, TVDI images before and after the landslide are resampled and subtracted to obtain a TVDI-D value of a change area, a threshold value for segmenting the landslide area and a non-landslide area is set, binaryzation is performed on the TVDI-D images, the binaryzed TVDI-D images are converted into vectors, and the vectors are converted into vectors; and screening the landslide candidate objects by taking the gradient information as a filtering condition so as to determine a landslide area vector. The precision of the landslide area detection method is improved in combination with multi-source remote sensing information, the false alarm rate in landslide detection is greatly reduced, and the credibility of a landslide detection result is also improved.

Description

technical field [0001] The invention belongs to the technical field of remote sensing image processing and the field of landslide monitoring, in particular to a method for accurately identifying new landslide areas based on multi-source time-series remote sensing images. Background technique [0002] As one of the large-scale geological disasters, landslides are extremely destructive and often cause a large number of economic losses and casualties. Therefore, the monitoring and detection of landslide disasters is very important. Traditional remote sensing visual interpretation and on-site survey methods consume a lot of manpower and material resources and are inefficient. Therefore, scholars at home and abroad have carried out a lot of research on landslide remote sensing detection. The early stage mainly focused on single-temporal remote sensing landslide detection and multi-temporal remote sensing landslide detection. detection. [0003] Single-temporal remote sensing lan...

AI Technical Summary

Problems solved by technology

[0004] The present invention aims at the general problem that the single-temporal remote sensing landslide detection method is easy to misjudge rocks, bare soil and roads as landslides and the multi-temporal remote sensing landslide detection method is easy to misjudge natural changes such as farmland and vegetation as landslides, etc., and provides A method for identifying landslide areas based on multi-source remote sensing time series images

Images