摘要
利用资源卫星信息提取技术，探讨中巴资源卫星 (CBERS-1 CCD) 数据在红树林资源调查中的应用能力。由于红树林群落间光谱特征的近似性和数据本身光谱信息的限制，给分类信息提取带来了困难。将红树林植被分布区与陆地植被分开，结合实地调查结果准确选取各群落训练区，采用神经网络分类法分类，获得了精度较高的分类结果。在此基础上，对主要群落的空间分布特征进行了简要分析。
The capability of extracting the mangrove information is discussed using CBERS-1 CCD data, and the extracted remote sensed thematic information is useful for the protection, management, and restoration of mangrove. Combining with the in situ survey data, we may learn the basic features and get the area data of mangrove from the remote sensed imagery . Due to the similarity among mangrove communities and the limit of spectral information of the CCD image, it is difficult to extract the mangrove from CBERS-1 data. So the distributed region of mangrove is firstly isolated from the land plant in the image referencing the map, and the field surveying, sampling and validating are also given special attention, and at last the good classification result is got using BP neural net method. The remote sensing is faster and more efficient than the routine procedure, and may provide useful information in time.
出处
《海洋通报》
CAS
CSCD
北大核心
2003年第6期30-35,共6页
Marine Science Bulletin
基金
国防科工委中巴卫星应用项目
华东师范大学河口海岸国家重点实验室开放基金
天津市科委"天津海域遥感图像解译应用软件开发"项目(
关键词
中巴资源卫星
红树林
遥感调查
神经网络
分类
CBERS-1 CCD data
Mangrove
remote sensing survey
neural net
classification
