高分辨率遥感影像结合LiDAR数据的面向对象分类方法  

Object-oriented classification with high resolution image and LiDAR data

作　　者：许传阳[1] 李建红[1] 

机构地区：[1]河南理工大学测绘与国土信息工程学院,河南焦作454000

出　　处：《河南理工大学学报（自然科学版）》2015年第2期222-225,共4页Journal of Henan Polytechnic University（Natural Science）

基　　金：国家自然科学基金项目资助项目(41371105)

摘　　要：为了更准确地对遥感数据进行分类,结合Geo Eye高分辨率遥感影像和机载Li DAR数据,通过对分割参数、特征选择、分类规则等特征进行研究,提出采用面向对象的模糊分类方法——成员函数法选择实验区进行了分类研究。实验结果表明:该分类方法能够更有效地提取出建筑物、煤堆、灌木等矿区典型地物,总体分类精度达到93.92%,KIA为92.52%,分类精度相比单一遥感数据明显提高。With the development of high resolution remote sensing technology and the multi-source of data acquisition, the multi-source sensor data fusion method has become a hot spot in the field of remote sensing information extraction. This paper combined GeoEye high resolution aviation image with the LiDAR data, by the research on segmentation parameters, feature selection, and classification rules characteristics, puts forward the utilization of the fuzzy method for object-oriented classification--member function method to select a classification study of the experimental area. The result showed that this object-oriented classification method can be more effectively extracted the buildings, coal pile, shrubs, and other mining typical objects, and that the overall accuracy reached 93.92% , KIA is 92.52%. Compared to the single remote sensing data, the classification accuracy has been significantly improved.

关 键 词：高分辨率遥感影像 LIDAR 面向对象分类 成员函数 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]