节点文献
基于语义相关的面向对象林地信息提取
Object-Oriented Forestry Land Extraction Based on Semantic Correlation
【摘要】 基于地物间的语义相关关系,结合面向对象的影像分析技术,提出了适用于高分辨率影像的林地信息提取方法,包括影像分割和林地信息提取2大步骤;并以福州市2003年6月成像的IKONOS卫星影像数据为例进行了林地信息提取实验。结果表明:①通过模糊和最近邻2种分类技术分别完成植被、林地信息的提取,提高了提取过程的合理性;②以面向对象的方式进行信息提取,不仅能避免细节信息的干扰,还能利用影像的纹理信息,这相当于对影像数据进行了优化,提取结果的精度和完整性均得到了提高。
【Abstract】 Based on semantic correlations among the cultures,a new method of forestry land extraction for high resolving power imagery is proposed by using object-oriented imagery classification in this paper,the proposed method included imagery segmentation and forestry land extraction.the forestland information extraction experiment was conducted by taking IKONOS satellite image data formed in Fuzhou in June,2003.The results showed: the vegetation and forestland information extraction through ambiguity and nearest neighbour classification technologies;Secondly,the information was extracted with object-oriented style,which not only could avoid detail information disturbance,but also could utilize the image lamination information,it meant the image data was optimized,both extraction result accuracy and integrality improved.
【Key words】 high resolving power image; image segmentation; semantic correlation; object-oriented; forestland information extraction;
- 【文献出处】 福建林业科技 ,Journal of Fujian Forestry Science and Technology , 编辑部邮箱 ,2009年03期
- 【分类号】S712
- 【被引频次】6
- 【下载频次】106