基于单幅图像数据的三维重构方法研究

针对传统的基于阴影恢复形状方法（SFS）的三维重构方法的局限性,综合运用计算机视觉、图像处理和计算机辅助几何设计理论与技术,提出了一种实用的三维重构方法。采用SFS方法获取三维形貌信息,利用图像分割技术提取物体的二维轮廓信息,通过三维形貌信息与二维轮廓信息的融合建立物体的数字化模型,通过B样条方法实现数字化模型的重构。实例表明,该方法能够有效地去除图像的背景信息,改善重构模型的分辨力,为基于图像的三维重构技术的广泛应用提供了条件。

Research on 3D Reconstruction Method Based on Single Image Data

Aiming at the limitation of traditional method for reconstructing surface based on SFS(shape for shading) method,a practical method of 3D reconstruction was presented.It was an integrated techniques of computer vision,image processing and computer aided geometric design techniques.Shape from shading method was applied to obtain the 3D shape of image,and the 2D contour of object was extracted through image segmentation technique.The numerical model of the object was established through message merging of 2D contour and 3D shape,and the model was reconstructed through the B-spline method.The experimental results demonstrate that the single-image based 3D reconstruction method given herein can get rid of redundant background of the image and improve the resolution of model reconstruction efficiently,and it will supply a helpful condition for the extensive application of image-based 3D reconstruction techniques.