摄像机标定参数的再优化方法.docVIP

  摄像机标定参数的再优化方法# 周富强，崔毅* （北京航空航天大学仪器科学与光电工程学院，北京 100191） 5 10 15 20 25 摘要：为了提高摄像机标定结果的准确性和实用性, 采用基于三维测量坐标系的摄像机标定 参数优化方法，对初次优化的摄像机参数进行了二次优化，参数的二次优化准则更接近实 际的测量误差准则，避免了参数优化坐标系与视觉测量坐标系不一致带来的测量精度下降 等问题。实验结果表明，在测量坐标系中进行的标定参数再优化结果比传统的以图像坐标 系为基准获得的参数具有更高的标定精度和测量精度，对于视觉三维测量工程的实际应用 具有重要意义。 关键词：测试计量技术及仪器；摄像机标定；参数优化；目标函数；测量精度 中图分类号：TP391.41 A Re-Optimization Method in Camera Calibration ZHOU Fuqiang, CUI Yi (School of Instrumentation Science and Optoelectronics Engineering,Beihang University,Beijing) Abstract: In this paper, a re-optimization method based on three-dimensional measurement coordinate was proposed. Generally, the traditional camera calibration techniques iterate with the objective function by minimizing the discrepancy between the detected position and the calculated projection of the 3D object point in pixels of a 2D image point. However, the final goal of vision measurement system is to acquire the accurate space coordinate of the measured point. To solve the mentioned problem, we implemented re-optimization process in camera calibration, whose principle is closer to the measurement error criterion. The experiment results show that the proposed re-optimization approach is quite efficient to improve calibration accuracy and measurement accuracy compared with traditional method, which is also important for practical applications in three-dimensional vision metrology engineering. Key words: Measurement Technology and Instrumentation; Camera Calibration; Parameter Optimization; Objective Function; Measurement Accuracy 30 0 引言 摄像机标定是视觉测量中一个重要而关键的环节[1]，所谓摄像机标定就是指根据摄像机 模型，利用特征点的提取图像坐标和已知世界坐标求解模型参数，从而建立物体空间和图像 平面的映射关系[2]。摄像机模型参数通常分为线性参数和非线性参数。线性参数遵循理想的 35 40 45 针孔成像模型，包含内部参数和外部参数，其中内部参数为摄像机固有参数，不会因为摄像 机位置变化而改变，外部参数则反映的是摄像机坐标系与世界坐标系的位置关系[3-5]。非线 性参数是镜头的畸变系数，镜头的光学畸变会引起实际成像模型与理想针孔模型不一致，导 致检测到的图像特征点偏离理想位置[6-8]。 为了获得较高的标定精度，一般需要采用优化方法估计出摄像机的所有参数。传统的标 定参数优化方法是根据摄像机模型，将已知物空间特征点投影到图像平面上，得到特征点的 模型图像坐标，模型图像坐标与摄像机实际探测到的图像坐标存在偏差，通过非线性优化的 过程使得这种偏差达到最小，从而求出偏差最小时的参数，即摄像机参数的最优解。