An vanishing point is defined as the convergence point of lines in an image plane that is produced by the projection of parallel lines in real space. Vanishing points can provide strong cues for inferring information about the 3D structure of a real scene. In this paper, a novel optimal vanishing point estimation method is proposed. This method detects line segment clustering based on random sample consensus (RANSAC) framework, obtaining the maximum likelihood estimation (MLE) results by minimizing the derived Sampson error. This method is performed without any prior knowledge of the camera parameters or information of underlying 3D lines. The error analysis based on backward propagation method is proceeded to give quantitative evaluation of our estimation algorithm. The physical experiments are carried out to validate the proposed method.%空间一组平行直线在图像平面上所成的像的交点称为消失点.消失点可以提供大量的场景三维结构信息.本文提出一种新的优化的消失点估计方法.该方法基于随机采样一致算法(Random sample consensus,RANSAC)对图像空间中的线段进行聚类,通过最小化Sampson误差获得消失点的极大似然估计(Maximum likelihood estimation,MLE).该方法不需要预知摄像机参数及直线的三维位置信息.为了对该算法进行定量评估,构造了基于反向传播的消失点误差传递模型.实验结果验证了本文提出算法的有效性.
展开▼
