近年来,随着数字化技术的不断进步,普通数码相机性能和计算机性能的不断提高,数码相机得到了普及,许多摄影测量系统都以普通数码相机作为获取数据的主要传感器。因此,数码相机的检校（或标定）逐渐成为摄影测量领域以及计算机视觉领域的研究热点之一。 影像像点坐标的提取精度与效率是影响相机检校自动化程度和精度的重要因素，针对这一问题，本文设计了一种编码标志，通过对编码标志自动识别与提取，借助直接线性变换方法建立物-像投影变换关系，并利用该关系完成影像像点自动初定位。通过边缘检测、最小二乘椭圆拟合，最小二乘直线拟合等步骤，完成高精度定位影像像点中心坐标。实现了影像像点坐标的全自动、高精度量测。 本文以相机检校为研究对象，首先对摄影测量中的自检校光束法平差相机检校方法进行了研究与实现，并分析了该方法的特点。深入学习计算机视觉中的LM算法以及稀疏矩阵技术，并将其应用于自检校光束法平差，找出了够解决自检校光束法平差计算速度慢的问题，实现了基于稀疏矩阵的自检校光束法平差相机检校方法。 本文做了一系列的实验，先后验证了本文方法的检校效率、检校精度以及在应用中的可行性。实验结果表明，本文的方法自动化程度高、计算效率高、检校精度高，可满足无人机航空摄影测量以及普通近景摄影测量。

In recent years, with advances in digital technology, ordinary digital camera performance and continuously improve the performance of your computer, digital cameras gained popularity, many photogrammetric systems are ordinary digital camera as the primary sensor data acquisition. Therefore, the digital camera calibration has gradually become one of the focus areas of photogrammetry and computer vision. The extraction accuracy and efficiency of image point coordinate is affect camera calibration automation degree and precision of the important factors, we designed a coded signs, through the automatic identification and extraction of coded signs, established by direct linear transformation method things - like projection transform relationships and use that image as the relationship between the completion point automatic early positioning. By edge detection, least squares ellipse fitting, linear least squares fitting procedure to complete high-precision positioning image as the center point coordinates. Image pixel coordinates to achieve the automatic, high-precision measurements. The camera calibration as the research object in this paper, Method of photogrammetry self-calibration bundle adjustment camera calibration were studied and implementation, and analysis of the characteristics of the method. In-depth study of computer vision algorithms and LM sparse matrix techniques, and apply self-calibration bundle adjustment, to find a solution to be calculated self-calibration bundle adjustment problems of slow, sparse matrix based on the realization self-calibration bundle adjustment camera calibration methods. This paper made a series of experiments have verified the efficiency of this method of calibration, calibration accuracy and feasibility in the application. Experimental results show that the proposed method a high degree of automation, high computing efficiency, high precision calibration, to meet the UAV aerial photogrammetry and general close-range photogrammetry.