项目名称: 基于成像光谱数据图谱合一特点的赤潮优势种识别方法研究
项目编号: No.60802089
项目类型: 青年科学基金项目
立项/批准年度: 2009
项目学科: 武器工业
项目作者: 马毅
作者单位: 国家海洋局第一海洋研究所
项目金额: 23万元
中文摘要: 本项目利用实验室培养、围隔培养、现场观测、航空高光谱观测等多种技术手段建立了不同优势种赤潮水体光谱数据库,开展了不同优势种赤潮水体高光谱遥感反射率、后向散射、吸收特征分析。发展了基于光谱角度(SAM)和支持向量机(SVM)等方法的赤潮优势种识别模型,提出了基于色度学特征的赤潮优势种识别方法;开展了基于遥感反演的固有光学量进行赤潮优势种识别的技术可行性分析。集成了所发展的模型算法,开发了基于航空高光谱遥感数据的赤潮优势种识别软件,并投入业务化应用。
中文关键词: 赤潮;优势种;识别;高光谱
英文摘要: Based on laboratory culture, mesocosm culture, in situ observation, airborne hyperspectral remote sensing, the spectral library for red tide with different kinds of dominant species was established. The spectral characteristic of remote sensing reflectance, backscattering, absorption of red tide was analyzed in terms of the different species. Algorithms to discriminating different dominant species of red tide were proposed including Spectral Angle Mapping and Support Vector Machine. Besides, a chromaticity diagram based method to identify dominant species was also put forward. The feasibility of discriminating dominant species by remote sensing retrievals of inherent optical properties was analyzed in this project. The software of discriminating different species of red tide with hyperspectral image as input is built up and has been operationally in use.
英文关键词: red tide; dominant species; identification; hyperspectral