结合机载LiDAR数据与航空可见光影像多层次规则分类建筑物变化检测

Building Change Detection Based on Multi-Level Rules Classification with Airborne LiDAR Data and Aerial Images

建议了一种结合Lidar点云与航空可见光影像的建筑物变化检测新方法,利用多层次规则分类算法解决这两种异元异构数据间建筑物变化检测难题。并建议了一种结合面积阈值的形态学后处理方法,从而形成一套完整的处理流程,可应用于实际生产。最终,利用中国吉林省长春市2010年机载LiDAR点云数据和2009年高分辨率航空影像对该方法的有效性进行了评价,通过与基于支持向量机(SVM)面向对象分类的建筑物变化检测算法比较,进一步对本研究建议的方法进行了验证与分析。结果显示,此方法效果理想,其精度优于基于SVM面向对象分类的建筑物变化检测方法。Kappa系数达到0.90,correctness达到0.87。

The present paper proposes a new building change detection method combining Lidar point cloud with aerial image,using multi-level rules classification algorithm,to solve building change detection problem between these two kinds of heterogeneous data.Then,a morphological post-processing method combined with area threshold is proposed.Thus,a complete building change detection processing flow that can be applied to actual production is proposed.Finally,the effectiveness of the building change detection method is evaluated,processing the 2010 airborne LiDAR point cloud data and 2009 high resolution aerial image of Changchun City,Jilin province,China;in addition,compared with the object-oriented building change detection method based on support vector machine（SVM）classification,more analysis and evaluation of the suggested method is given.Experiment results show that the performance of the proposed building change detection method is ideal.Its Kappa index is 0.90,and correctness is 0.87,which is higher than the object-oriented building change detection method based on SVM classification.