Rapid extraction method of mangroves by using Sentinel-2 images
In response to the problem that the spectrum of mangroves contains both vegetation and wetland information,it is difficult to achieve rapid and accurate extraction of mangroves using satellite remote sensing images.This paper compares and analyzes the spectral curve characteristics of mangroves in the red edge and shortwave infrared bands on Sentinel-2 images with other typical land features,and proposes a new type of mangrove extraction index NDMI(Normalized Difference Mangrove Index,NDMI)for rapid and accurate extraction of mangroves.Taking Pearl Harbour in Fangchenggang City,Guangxi as the research area,this index was compared with four commonly used mangrove extraction indices,and its applicability was evaluated in Qinzhou Harbour and Dongzhaigang East Harbour National Nature Reserve.The results show that NDMI can enhance the difference in reflectance between short-wave length infrared bandsand the red-edge band,and can more effectively distinguish mangroves from other typical land features compared to other commonly used indices;By using Sentinel-2 remote sensing data and combining it with NDMI index,mangrove information can be quickly and effectively extracted without dividing the intertidal zone.Moreover,the b12 band in the short-wave length infrared bands has better extraction performance,with an overall accuracy and kappa coefficient of 98.6％and 0.924 9,respectively.This study provides a fast and convenient new method for remote sensing extraction of mangroves.
Sentinel-2mangroveintertidal zonespectral curve characteristics of mangrovesNDMI
万方数据
维普
