摘要
在对矿物光谱特征理解与归纳的基础之上,对矿物光谱特征进行知识化表达,利用数理逻辑和一定的判别规则实现对高光谱遥感影像矿物的自动识别与批量化信息提取。在ENVI平台上,利用IDL语言开发了高光谱遥感影像矿物分层自动识别模块(M ineral Auto-identification Modu le Based on Spectral Identification Tree:MAIM-SIT)。该模块已经在新疆东天山哈密地区利用HyM ap数据、西藏驱龙地区利用Hyperion数据以及美国Cuprite地区利用AVIR IS数据成功地进行了矿物识别,可识别的矿物或矿物组合可达10种以上,基本实现了高光谱矿物信息提取的智能化与批处理能力。
Spectral knowledge acquired through the understanding of mineral spectral features was used to perform automatic extraction of mineral type information based on mathematical, logic and some other decision rules in the hyperspectral imaging field. In this paper, a mineral auto - identification module for hyperspectral imaging data ( MAIM - HID) has been designed by IDL language on ENVI software. It has intelligence and batch processing capacity so that it can identify and extract as many as over 10 types of minerals or mineral groups directly. This module is applicable to aero Hymap and AVIRIS data as well as satellite Hyperion data. It already identified and discriminated some minerals in East Tianshan Mountain of Xinjiang and Qulong area of Tibet in China and Cuprite in U.S.A.
出处
《国土资源遥感》
CSCD
2005年第4期28-31,i0004,共5页
Remote Sensing for Land & Resources
基金
国土资源部"百名优秀青年科技人才计划"
国家自然科学基金(40201034)
国土资源部科研项目(2002206)资助
关键词
高光谱遥感
矿物自动识别
矿物光谱数据
IDL
Hyperspectral imaging data
Mineral auto- identification module
Mineral spectra
IDL