A novel high-precision dynamic calibration and distortion-free image generation method for line-scan cameras with an auxiliary frame camera

Existing line-scan camera (LSC) calibration methods typically rely on complex calibration targets and are constrained by the scanning direction, constant scanning speed, and an inability to calibrate the scanning direction. These limitations often lead to scanning distortion and reduced calibration accuracy of the measurement system. Moreover, the resulting scanned image cannot be reliably corrected or transformed. This study addresses these issues by proposing a high-precision dynamic calibration method for LSCs using an auxiliary frame camera. A dynamic scanning imaging model is established, and the geometric parameters of the LSCs are estimated utilizing an auxiliary frame camera. Moreover, lens distortion parameters are introduced to construct an optimization objective function that includes parameters, such as the scanning vector, thereby achieving high-precision dynamic calibration of LSCs. Additionally, a real-time pose–scan image correction and transformation method based on the line-scan velocity matrix is proposed for scanned image distortion. Practical calibration and correction experiments are conducted to analyze the influence of various factors on calibration accuracy, along with comparative evaluations against existing methods. The proposed method reduces the root mean squared (RMS) reprojection error of the line-scan system by 70.3% compared with existing methods. With the proposed image correction and transformation method, the relative deviation of aspect ratio is reduced by up to 90.69%. The findings show that the proposed method can effectively overcome key challenges in high-speed line-scan inspection.