地理空间信息2024，Vol.22Issue(3)：11-15,5.
基于影像密集匹配点云的建筑物变化检测方法
Building Change Detection Method for Dense Matched Point Clouds of Images
摘要
Abstract
The changes of building height can be effectively detected by using the dense matched point clouds of images.We proposed a building change detection method based on dense matched point clouds of images that integrating deep neural network and spatial voxels.Firstly,we constructed an attention-driven deep neural network.Then,we used the cloth simulation filtering algorithm and vegetation indexes to remove the ground points and vegetation points respectively.Finally,we detected the changed buildings by comparing spatial voxels.The experimental results show that the miss detection rate of proposed method achieves 0%,and the false alarm rate is reduced by 48.64%compared with the traditional method,which illustrating that this method has the potential to greatly improve the efficiency of illegal building detection,meets the requirements of practical applications.关键词
密集匹配点云/建筑物变化检测/注意力机制/空间体素Key words
dense matched point cloud/building change detection/attention mechanism/spatial voxel分类
天文与地球科学引用本文复制引用
李正洪,全昌文,陈华江,陈敏,吕琦..基于影像密集匹配点云的建筑物变化检测方法[J].地理空间信息,2024,22(3):11-15,5.基金项目
广西重点研发计划资助项目(桂科AB22080077) （桂科AB22080077）
广西科技基地和人才专项资助项目(桂科AD20238044) （桂科AD20238044）
广西空间信息与测绘重点实验室基金资助项目(191851011). （191851011）
