An Automatic Intrinsic and Extrinsic Calibration Method for Multi-Camera Systems Without a Common Field of View Under Weak Observation Conditions Based on Total-Station-Assisted Point Localization

To address the calibration challenges of multi-camera measurement systems without a common field of view under weak observation conditions, this paper proposes an automatic intrinsic and extrinsic calibration method based on total-station-assisted point localization. First, illuminated marker points are deployed on-site, while cameras and a total station are set up in appropriate positions. The cameras capture images of the illuminated markers and quickly compute the spot center coordinates using the grayscale centroid method, while the total station acquires the 3D world coordinates of the markers. Finally, the intrinsic and extrinsic parameter matrices of the cameras is solved within the RANSAC algorithm framework, enabling automatic calibration of multi-camera systems without a common field of view. A field experiment was conducted in a subway tunnel under weak observation conditions. The average reprojection error of the calibrated camera parameters was 0.65 pixels. The experimental results demonstrate that the proposed total station and camera joint point localization-based calibration method achieves high-precision camera calibration and holds promising potential for engineering applications.