Tracking method for optical center of projector based on binocular vision

Structured light 3D perception technology is widely used in conventional surface quality measurement. However, there exists a core issue that the insufficient accurate global coordinate tracking of the projector's optical center restricts the overall measurement accuracy. To solve this issue, we proposes a real-time tracking algorithm for the projector's optical center pose based on tracking markers. In traditional structured light systems, the projector's optical center must be obtained via inverse camera calibration. As a virtual optical reference point, it lacks directly detectable physical features, making accurate tracking difficult for conventional tracking devices. The proposed algorithm first attaches tracking markers to the surface of the measurement device. Then, it uses multiple sets of pose data for iterative calculations to establish a rigid transformation matrix between the tracking markers and the projector's optical center. During the actual measurement phase, the tracking device continuously captures the global coordinates of the tracking markers. By substituting these coordinates into the pre-established rigid transformation matrix, the real-time pose of the optical center in the global coordinate system can be quickly inverted. This method provides an accurate and stable foundation for the global stitching and management of structured light measurement data, effectively resolving the problem of difficult accurate real-time tracking of the optical center in traditional methods.