文献类型：会议论文

英文题名：TROPICAL NATURAL FOREST CLASSIFICATION USING TIME-SERIES SENTINEL-1 AND LANDSAT-8 IMAGES IN HAINAN ISLAND

作者：Zhang, Lu[1,2] Wan, Xiangxing[3] Sun, Bing[4]

第一作者：Zhang, Lu

通信作者：Zhang, L[1];Zhang, L[2]

机构：[1]Key Lab Earth Observat Hainan Prov, Sanya 572029, Hainan, Peoples R China;[2]Chinese Acad Sci, Inst Remote Sensing & Digital Earth, Key Lab Digital Earth Sci, Beijing 100094, Peoples R China;[3]Chinese Acad Forestry, Res Inst Forest Resources Informat Tech, Beijing 100091, Peoples R China;[4]Beihang Univ, Sch Elect & Informat Engn, Beijing, Peoples R China

会议论文集：IEEE International Geoscience and Remote Sensing Symposium (IGARSS)

会议日期：JUL 28-AUG 02, 2019

会议地点：Yokohama, JAPAN

语种：英文

外文关键词：Classification; remote sensing; Hainan island; tropical natural forest

年份：2019

摘要：Tropical natural forest plays an important role in environmental change and biodiversity researches. However, the complexity of structures and its cloudy and rainy environment make tropical natural forest classification difficult. Taking Hainan, China as study area, we conduct a tropical natural forest classification study by combining multi-temporal synthetic aperture radar (SAR) images collected from Sentinel-1 satellite and optical images collected from Landsat-8 satellite in this paper. The backscatter coefficient, spectrum information, digital elevation mod (DEM), temporal information offered by multiple remote sensing data have been analyzed to identify the evergreen and deciduous broad-leaved forest, evergreen coniferous forest, tropical monsoon forest, typical tropical rain forest and other forest types. In addition, a two-stage tropical forest classification strategy is proposed based on support vector machine (SVM) classifiers, namely the primary land cover type classification and tropical natural forest type classification in which the time-series backscattering information is used. Finally, the Hainan tropical forest mapping image is obtained based on the proposed classification strategy and the overall accuracy reaches to 90% based on field survey data. The results show the effectiveness of the classification strategy on tropical natural forest classification.