以自动生成测地线活动轮廓(geodesic active contour,GAC)模型的初始曲线及改进其外力为出发点,提出一种基于LiDAR点云和随机影像数据,利用改进的GAC模型进行建筑物边界提取的方案。首先利用形态学交替序贯滤波自动获得模型演化的初始曲线;进而利用LiDAR深度梯度影像改进模型演化的外力,得到了改进的测地线活动轮廓(improved geodesic active contour,IGAC)模型。仿真试验表明,采用IGAC模型,可抑制弱边界泄漏,并提高建筑物边界提取的完整性和形状精确度。
ImprovedGAC Model based on the initial curve and the external force is proposed, It isa new method for automatic building extraction from LiDAR point clouds combined with the aerial image, The morphological alternating sequential filtering is used to access the initial curve automatically, and then the external force is improved by the LiDAR depth gradient image, so the improved geodesic active contour (IGAC) model is proposed, Experiment indicates that this method can inhibit the weak boundary leakage and produce high accuracy and completeness of building boundary.