Mangrove High-resolution Remote Sensing Monitoring Technology Research

Mangroves forests are distributed along coastlines in tropical and sub-tropical latitudes, mangrove wetlands are important ecological species and wetland habitats. In China's southern coast and the mouth of the Gulf, etc., exist a lot of mangrove. With bank protection, elimination of pollution and conservation of biological diversity and other ecological functions.Supported by the subject "Satellite Remote Sensing Investigation and Research of Islands and Coastal zones in Guangdong Section", taking Zhanjiang Mangrove National Nature Reserve and Shankou National Mangrove Nature Reserve near YingLuo port as example, development a remote sensing detection and identification technology of mangrove based on SVM (Support Vector Machine) . The main results are as follow:(1) With two mangrove Investigations, acquired 30 Ground control points of remote sensing images; using Landsat TM, ETM+ and SPOT images, acquire interpreting key of remote sensing; using SPOT images, getting four types of nterpreting key of remote sensing, including Avicennia marina, Acigeras corniculatum, Rhizophora candel, Bruguiera gymnorrhiza/Kandelia candle.(2) With Landsat TM, ETM+ , SPOT image as data source, using SVM methods, mangrove distribution of space, studying detection technology research of mangroves, The results shows that mangrove in TM, ETM+ and SPOT image detection accuracy can reached 84.7%, 87.5 %, 90.3%. Using three remote sensing images, studying mangrove dynamic monitoring, found that from 1991 to 2000 the mangrove area increased 440.5 hm~2, 2000 to 2004 the mangrove area increased 58.3 hm~2.(3) Based on SVM methods and four types of mangrove interpreting key based on SPOT Image, studying the mangrove species recognition technology research. The identify experimental results show that the Kappa coefficient of Avicennia marina, Acigeras corniculatum, Rhizophora candel, Bruguiera gymnorrhiza/Kandelia candle classification achieved 0.86, 0.81, 0.84 and 0.79. The overall classification of Kappa coefficient achieved 0.82.