结合形变模型和ACD方法提出了基于形变模型的3D表面自适应重建方法.同时引入了与图象统计特性有关的外力,使得表面重建结果与模型的初始位置无关,利用ACD方法使模型自适应地改变其拓扑结构;为了提高表面重建的速度和鲁棒性,提出了多尺度重建算法.该方法适用于形状、结构复杂的物体重建,实验结果证明了该方法的有效性.

This paper presents a method for 3D surface adaptive reconstruction based on a deformable model, which integrates a deformable model and ACD(Affine Cell Decomposition)method. An external force is developed from the statistical characteristics of image that drives the model towards surface of object with complex geometries and topologies. ACD method is used to adapt topologies of model by reparameterizing the deformable model. In order to improve the speed and robust of 3D surface reconstruction, the mul-scale surface reconstruction algorithm is proposed. Results of several experiments show the potential of the method for reconstruction some objects with complicated shape and topologies.