【摘要】 提出了一种自动获取棋盘角点的新算法。该方法采用由粗到精的多层次检测策略,通过综合使用Radon变换、Harris算子以及Forstner算子,能够快速准确地提取出图像中的棋盘点。与传统方法相比,该算法具有两个明显的优点:无须人工干预,能够自动确定角点位置;定位精度高,角点的平均位置偏差在0.1个像素以内。更多还原

【Abstract】 A new algorithm for automatically extracting chessboard corners is presented. In this paper, the coarse-to-fine scheme is adopted and all the necessary corners can be obtained quickly and accurately through integration of Radon transform, Harris corner detector and Forstner detector. The algorithm has two traits relative to the traditional approach: the position of the corners can be determined automatically while not requiring manual interposition, and high precision could be acquired and the mean position error is below 0.1 pixels.更多还原