Forest fire hazard burning site extraction method based on multi-source remote sensing data

Abstract

The invention relates to a forest fire hazard burning site extraction method based on multi-source remote sensing data. The method comprises the following steps: S1, calculating a forest fire disturbance index based on high-time resolution remote sensing data, judging a pixel of which the forest fire disturbance index is greater than a first threshold as a burning pixel, determining the burning pixel position as a fire point position, and recording time information of the fire point position; S2, based on high spatial resolution remote sensing data, calculating normalized vegetation index changes of the pixels within a preset distance range around the fire point position before and after a fire disaster by taking the fire point position as a circle center, judging the pixels as burning pixels when the absolute value of the normalized vegetation index changes is greater than a second threshold value, and extracting all the burning pixels to generate burning site data. The method provided by the invention can accurately determine the whole fire occurrence range, improves the information extraction precision based on the remote sensing fire mark, and distinguishes the information fromother interference factors.

Description

technical field [0001] This application relates to the technical field of biomass burning satellite remote sensing detection. Specifically, this application relates to a method for extracting forest fire burnt traces based on multi-source remote sensing data. Background technique [0002] Wild fire is a natural phenomenon in the process of ecosystem succession, but it can also seriously affect regional ecosystems and global climate. This impact is mainly manifested as changing the chemical composition of the atmosphere, which may directly affect the balance of cosmic radiation and the global climate system; destroy the production function of regional ecosystems, reduce animal habitats and biodiversity; interfere with vegetation succession and biological nutrient cycles. Therefore, it is necessary to detect the extent of forest fires and the area of ​​burned areas on a global or regional scale, so as to control the fire disaster in time and accurately estimate the image of t...

AI Technical Summary

Problems solved by technology

But at the same time, it should also be noted that because the extraction of burnt traces has a high requirement for the time resolution of remote sensing data, the existing studies all use data with high time resolution, and these data often have low spatial resolution. For some small fires, the extraction accuracy is low, and the area of ​​fire trace extraction is smaller than statistical data or visual interpretation data, etc. [11]

Images