3D Laparoscopic Surface Reconstruction Using Active Stereo

The patent segments the matching process into two distinct stages: coarse matching using low-frequency phase maps to establish initial correspondences, and fine matching using high-frequency phase maps to refine precision. This segmentation allows the system to achieve high measurement precision without requiring multiple sequential shots, thereby reducing image acquisition time for dynamic surfaces.

The patent applies partial action by using only the necessary frequency components for each matching stage. Instead of applying full-spectrum structured light patterns repeatedly, the system selectively uses low-frequency patterns for coarse matching and high-frequency patterns only where needed for fine matching, reducing the total number of patterns required and thus the acquisition time.

The patent transitions from a single-image-dimensional approach to a stereo-dimensioned approach by implementing a dual-camera system. This dimensional change provides redundant viewing angles that enable robust correspondence matching even under varying illumination conditions, as the stereo geometry constrains the solution space and eliminates ambiguous matches that plague single-camera systems.

The system employs feedback mechanisms where the coarse matching results from low-frequency patterns inform the fine matching process with high-frequency patterns. This hierarchical feedback approach allows the system to adaptively refine correspondences, improving matching reliability by using previous stage results to guide subsequent processing stages.

The stereo-based system performs self-calibration through geometric constraints inherent in the dual-camera configuration. The epipolar geometry and fundamental matrix relationships provide built-in reference frames that eliminate the need for external calibration targets or complex luminance compensation procedures, allowing the system to self-adjust to illumination variations without manual intervention.

The patent replaces the mechanical calibration procedures (physical calibration targets, manual projector alignment) with computational geometry-based calibration inherent to stereo vision. The system uses mathematical relationships between the two camera views to automatically establish correspondence, substituting complex mechanical adjustment operations with automated computational processes.