威海市近海水产养殖信息提取方法研究

发布时间：2018-06-21 21:48

本文选题：光谱 + 纹理　；参考：《湿地科学》2017年05期

【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%。
[Abstract]:When the spectral information of remote sensing image is used to extract the information of offshore aquaculture, it is easy to confuse the culture water body with the natural water body, while the texture information of remote sensing image is used to extract the aquaculture information in the offshore waters alone. It is difficult to distinguish a single large aquaculture water body from the natural water body. In order to solve the above problems, a method of extracting aquaculture information based on the spectral and texture information of remote sensing image was proposed by using Oli image data. Firstly, on the basis of analyzing the spectral characteristics of aquaculture water bodies in the study area, the principal component transform method is used to compress, mine and select the spectral information of Oli images. The window size and texture feature statistics of gray level co-occurrence matrix are analyzed to distinguish the aquaculture area in the study area, the texture feature is selected, and the selected texture information and spectral information are combined. Finally, According to the differences of spectral and texture features among different water types, the fuzzy logical membership function of three kinds of aquaculture water bodies in the study area was constructed, and the information of aquaculture in the study area was automatically extracted. The results show that the method can extract aquaculture information from the study area and the classification accuracy is 97.93%.
【作者单位】：山东女子学院信息技术学院;湖北省国土测绘院;国家测绘地理信息局四川基础地理信息中心;成都理工大学国土资源部地学空间信息技术重点实验室;
【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
【分类号】：S951.4;TP751

【相似文献】