Corner detection method for calibration targets in digital image correlation applications based on attention mechanism super-resolution
Haoyu Wang, Guihua Li, Weiqing Sun, Ziwei Wang, Xu Wang, Mei Zhang · Optical Engineering · 2025
Camera calibration plays a crucial role in stereo digital image correlation (DIC) measurements. However, traditional calibration techniques often encounter challenges such as inadequate detail and blurriness during corner detection, particularly at low resolutions, which negatively impact the accuracy of both calibration and measurement. To overcome these limitations, we introduce a super-resolution corner detection approach based on an attention mechanism. By emphasizing the high-frequency information of calibration corners, the attention mechanism facilitates the reconstruction of low-resolution images, leading to more precise corner detection during calibration and improved camera calibration accuracy. Experimental results in DIC camera calibration demonstrate that the proposed method achieves lower reprojection errors. In addition, this method enhances the geometric reconstruction accuracy and improves the precision of dynamic measurements in DIC.