文献类型：会议论文

英文题名：Forest mapping using ENVISAT and ERS SAR data in Northeast of China

作者：Li, ZY[] Pang, Y[] Chen, EX[] Schmullius, C[] Santoro, M[]

第一作者：李增元

通信作者：Li, ZY[1]

机构：[1]Chinese Acad Forestry, Res Inst Forest Resource & Informat Technol, Beijing, Peoples R China

会议论文集：25th IEEE International Geoscience and Remote Sensing Symposium (IGARSS 2005)

会议日期：JUL 25-29, 2005

会议地点：Seoul, SOUTH KOREA

语种：英文

外文关键词：ERS-1/2 tandem SAR; ENVISAT ASAR; forest; China

年份：2005

摘要：The need to monitor forest is dramatically increasing in China, particularly after the introduction of new laws against deforestation. The well known ERS-1/2 tandem mission provides a large mount of SAR for interferometric processing. As forests have special volume decorrelation mechanism, the Interferometric Land-Use (ILU) images were used to generate forest map in Northeast of China. The striped ILU images processed by ESRIN Interferometric Quick-Look (IQL) processor have been used to generate forest/nonforest map. The algorithms of long strip ELU data geo-reference and mosaic have been developed. Decision tree classifier was used to map forest/nonforest on a pixel basis. The algorithm depends primarily on interferometric coherence. The threshold levels are adapted to the radiometry of individual strips. The gaps among ERS are filled by Landsat TM data of the same phase. The limited quality of the ELU from IQL system and the partial coverage of China limited the generation of forest base maps for larger areas of China. Under the Dragon Forest project between China MOST and ESA, the ERS SLC SAR data will be processed and the SEBERIA-I methods will be adapted to generate a forest volume density base map of 1995-1996 in this region. The Envisat ASAR Alternating Polarisation (AP) data will be used to update the base map to 2004-2005. Several hundred scenes ASAR AP data have been received during the summer and winter of 2004. Three test sites have been used to develop methodologies, which located in Daxinganling, Xiaoxinganling and Changbaishan separately. The digitalized forest map and field plots are used as reference data. The results from test sites show: i) It is very potential and feasible to generate forest volume density map using winter ERS tandem data. ii) Forest can be well discriminated from non-forest with ASAR AP (HH/HV) data. iii) ASAR HV backscattering signals increase with forest canopy density.