Liu Hui, Shan Ligen, Feng Jiahao, Wang Shuanghao
School of Automation, Xi'an University of Posts & Telecommunications, Xi'an 710121, China.
Sensors (Basel). 2025 May 18;25(10):3180. doi: 10.3390/s25103180.
This article proposes a novel approach for corner detection in noisy checkerboard images, comprising several methodical steps: (1) an initial extraction of corners utilizing the cross features present in the edge image of the checkerboard; (2) the elimination of erroneous corners through an analysis of the periodic consistency among the detected corners; (3) the identification of the outermost corners and the subsequent generation of a rectangular bounding box based on the total number of input checkerboard corners; (4) the reconstruction of missing corners, which may have been obscured by noise, by leveraging the periodic characteristics of the corners. Experimental findings indicate that this methodology is capable of effectively detecting all corners of the checkerboard across varying levels of noise, thereby significantly enhancing the success rate of corner detection in noisy images. This makes the proposed method particularly advantageous for camera calibration in special scenarios where noise or contamination in checkerboard images is unavoidable.
本文提出了一种用于在有噪声的棋盘图像中进行角点检测的新方法，该方法包括几个有条不紊的步骤：（1）利用棋盘边缘图像中存在的交叉特征对角点进行初始提取；（2）通过分析检测到的角点之间的周期性一致性来消除错误的角点；（3）识别最外层的角点，并根据输入棋盘角点的总数生成一个矩形边界框；（4）利用角点的周期性特征重建可能被噪声遮挡的缺失角点。实验结果表明，该方法能够在不同噪声水平下有效地检测棋盘的所有角点，从而显著提高有噪声图像中角点检测的成功率。这使得所提出的方法在棋盘图像中不可避免地存在噪声或污染的特殊场景下进行相机校准特别有利。
