威海市近海水产养殖信息提取方法研究  

Information Extraction Method of Aquaculture in the Area of Inshore Coastal Waters,Weihai City

作　　者：李缨[1] 张翠芬[1] 帅爽 刘晰[3] 郝利娜[4] 

机构地区：[1]山东女子学院信息技术学院,山东济南250002 [2]湖北省国土测绘院,湖北武汉430010 [3]国家测绘地理信息局四川基础地理信息中心,四川成都610093 [4]成都理工大学国土资源部地学空间信息技术重点实验室,四川成都610059

出　　处：《湿地科学》2017年第5期651-656,共6页Wetland Science

基　　金：山东省高等学校科技计划项目(J12LN42和J15LN11);全国统计科学研究计划项目(2012LY022);山东省自然科学基金项目(ZR2012DL01)资助

摘　　要：单独利用遥感影像光谱信息进行近海水产养殖信息提取时,养殖水体与自然水体易混淆,而单独利用遥感影像纹理信息提取近海水域水产养殖信息时,单一大块养殖水体与自然水体又难以区分。针对上述问题,利用OLI影像数据,提出了一种综合遥感影像光谱与纹理信息进行水产养殖信息提取的方法。首先,在对研究区养殖水体类型进行光谱特征分析的基础上,采用主成分变换方法,对OLI影像光谱信息进行压缩、挖掘和选取;其次,分析灰度共生矩阵窗口尺寸和纹理特征统计量对研究区水产养殖区域的区分能力,完成纹理特征的选取,并将选取的纹理信息和光谱信息进行特征协同;最后,对特征协同数据进行多尺度分割,根据各水体类型间的光谱特征和纹理特征的差异,构建研究区3种养殖水体的模糊逻辑隶属度函数,实现对研究区水产养殖信息的自动提取。研究结果表明,该方法能较好地提取研究区水产养殖信息,总体分类精度达到97.93%。The aquaculture waters and natural waters in the area of inshore coastal waters were easy to confound, when aquaculture information was extracted by spectral information or texture information alone. A new extraction method was proposed by comprehensive spectral information and texture information for seaside aquaculture. Firstly, based on the spectrum analysis of aquaculture waters in the study area, spectral information of Operational Land Imager（OLI） images were compressed, analyzed and extracted by principal component analysis transformation. Then, texture information of OLI image was extracted by discussing the effect of the window size and texture statistics after gray level co-occurrence matrix computing upon the separability of aquaculture waters. The extracted spectral information and texture information were synergy used as a new data. Finally, the fuzzy classification of aquaculture water was carried out after the new data was multi-scale segmented and the fuzzy logic membership functions of the three types of aquaculture waters were built based on the water types＇ texture and spectral differences. The result showed that the proposed method was successfully extracted the aquaculture waters in the study area, and received a high total accuracy of 97.93%.

关 键 词：光谱 纹理 水产养殖 信息提取 近海水域 OLI 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置] S967.2[自动化与计算机技术—控制科学与工程]