Exploiting Differential Optical Absorption for Automated High-Precision Spatial Calibration in Robotic Photoacoustic Imaging

Photoacoustic (PA) imaging, combining optical contrast and ultrasound penetration, is a promising biomedical modality for applications in tissue imaging and diagnostics. While robotic systems have been demonstrated to enhance the consistency and precision of PA imaging when compared with traditional freehand operation, the acquisition of accurate three-dimensional (3D) reconstruction from 2D probes necessitates a robust spatial calibration process. This calibration is imperative to establish the precise transformation matrices between the coordinate frames of the robotic arm and the PA image plane. However, a systematic and automated calibration framework specifically designed for robotic PA systems has not been reported to date, representing a critical bottleneck for their clinical translation. This paper introduces a novel, automated calibration framework for robotic PA imaging, demonstrated on a Universal Robots (UR3) arm with a custom-designed 2D PA probe. Integrating camera-to-robot and PA image calibration, the framework achieved a root-mean-square (RMS) reprojection error of 0.94 mm. The validation of the system was performed through the 3D reconstruction of a multi-target phantom, thereby confirming the framework’s capability for automated scanning and accurate spatial mapping.