【摘要】 本文探讨了一种用于液位图像标定的相机自检方法。实验表明：该方法基于YOLOv4目标检测的SIFT特征点提取和FLANN匹配的位移测量算法，可实现更精确的目标位移测量，不仅能应对日常使用中相机可能出现的轻微扰动问题，而且能有效地发现相机扰动，确保液位检测的准确性和稳定性。更多还原

【Abstract】 This paper discusses a camera self-test method for liquid level image calibration.Experiments show that the proposed method, based on SIFT feature point extraction of YOLOv4 target detection and FLANN matching displacement measurement algorithm, can achieve more accurate target displacement measurement, which can not only deal with the slight disturbance that may occur in daily use of the camera, but also effectively detect the camera disturbance to ensure the accuracy and stability of liquid level detection.更多还原