一种面向对象的高分辨率遥感影像信息提取方法  

Object-oriented High-resolution Remote Sensing Image Information Extraction Method

作　　者：胡金梅 董张玉[1,2] 杨学志 HU Jinmei;DONG Zhangyu;YANG Xuezhi

机构地区：[1]合肥工业大学计算机与信息学院,安徽合肥230009 [2]工业安全与应急技术安徽省重点实验室,安徽合肥230009

出　　处：《地理空间信息》2021年第9期10-13,18,I0001,共6页Geospatial Information

摘　　要：针对传统面向对象分类方法的不足,根据研究对象特征构建了一种改进的面向对象的高分辨率遥感影像信息提取分类方法。首先利用SLIC超像素算法对影像进行分割,并提取分割后影像的纹理、光谱和形状特征;再利用SVM分类器提取影像信息,区分相似性较高的耕地和道路;然后利用随机森林算法提取水体和人工表面;最后对不同地物信息的提取结果进行拼接,实现土地利用分类。结果表明,与传统的面向对象分类方法相比,该方法的分类精度更高。Aiming at the shortcomings of traditional object-oriented classification method,we constructed an improved high-resolution remote sensing image information extraction classification method based on the characteristics of research objects.Firstly,we used SLIC super-pixel algorithm to segment the images,and extracted the texture,spectrum and shape characteristics of the segmented images.Then,we used SVM classifier to extract image information,and distinguished cultivated land and road with high similarity.We used random forest algorithm to extract water and artificial surface.Finally,we spliced the extraction results of different characteristic information to realize land use classification.Experimental result shows that the classification accuracy of this method is higher than that of traditional object-oriented classification method.

关 键 词：面向对象 影像分类 高分辨率影像 SVM 随机森林算法 

分 类 号：P237[天文地球—摄影测量与遥感]