The angular error of multi-line array laser imaging detection and ranging(LiDAR) has a distance-proportional impact on measurement results, necessitating periodic calibration. To address this issue and enhance the calibration performance of such angular errors, this paper proposes an indoor calibration method based on the principle of image detection. Boasting advantages of high efficiency, low cost, high precision and strong universality, the method employs a charge-coupled device (CCD) camera and an image-parallel optical tube to acquire pixel coordinate observations of laser images, and then derives a relational model using spherical geometry to establish a quantitative relationship between angular errors and the observations. The experimental validation was carried out using a Pandar 40P LiDAR manufactured by HexaTech. The results show that the LiDAR horizontal angular error obtained by the proposed method is 176.02″, and the measured axial systematic error is 0.0757o, which is very close to the results obtained by the target-based method.. The point cloud data were corrected using the measured error of the method, and the accuracy of the corrected data was greatly improved (within ±4″). The repeatability of the data is also verified, providing a new solution for LiDAR angle calibration.