煤炭产业是我国重要的基础产业，煤矿事故的频繁发生使得国家及全社会非常重视煤矿安全生产问题。随着监控技术的飞速发展，视频监测系统在煤矿生产方面得到了广泛应用。巷道形变计算是煤矿安全监测的重要内容，本文主要对煤矿巷道形变计算中摄像机标定过程及其参数优化方法进行深入研究，并开发了巷道形变计算系统。 首先分析了摄像机成像模型和常用的摄像机标定方法。在此基础上，为了提高摄像机标定精度，给参数优化方法提供良好的初始值，本文使用多幅模板图像标定左右两个摄像机的内部参数和畸变系数，再利用监测现场采集到的一幅图像标定外部参数。然后介绍了最优化问题的数学模型以及几种经典的优化算法，并根据非线性摄像机成像模型建立参数优化模型，提出基于LM算法的摄像机参数优化方法，同时对其求解过程进行了详细推导。最后利用优化后的摄像机参数和巷道面上特征点的图像坐标求解其三维坐标，根据不同时刻下的坐标值计算形变位移。 在VS2010的平台上开发了巷道形变计算系统，利用该系统对模拟巷道面上的特征点进行监测实验。实验通过精度评定标准计算优化前后的坐标误差，结果表明优化后的摄像机参数精度更高，验证了本文优化方法的正确可行性，并证明该系统可以精确求解特征点在空间中的三维坐标和形变位移。

Coal industry is an important basic industry in our country, the frequent occurrence of coal mine accidents makes the country and whole society pay great attention to the safety production of coal mine. With the rapid development of monitoring technology, video surveillance system has been widely used in coal mine production, The calculation of roadway deformation monitoring is an important part of coal mine safety monitoring. In this paper, the camera calibration process and its parameter optimization method in the calculation of coal mine roadway deformation are studied deeply, and the deformation calculation system is developed. Firstly, the paper analyzes the camera imaging model and the conventional camera calibration method. On this basis, in order to improve the precision of camera calibration, and provide good initial values for the parameters optimization method. This paper calibrates internal parameters and distortion coefficients of left and right cameras by using multiple template image, then, according to a picture of monitoring site to calibrate external parameters. Secondly, this paper introduces the mathematical model of optimization problems and several kinds of classical optimization algorithms. Then the parameters optimization model is established by nonlinear camera imaging model, the camera parameters optimization method based on LM algorithm is proposed, at the same time, the solving process of this model is also deduced in detail. Finally, the 3D coordinates are solved by the optimized parameters and the image coordinates of the feature points on the surface of roadway, and the deformation displacement is calculated according to the coordinates at different times. The roadway deformation computing system is developed on the platform of VS2010, this system is used to monitor the feature points on the surface of roadway. The experiment calculates the coordinate error before and after optimization, the results show that the optimized parameters of camera have higher precision. This optimization method is proper and feasible, and the system can accurately calculate the three-dimensional coordinates of the feature points and deformation displacement in space.