Improved Internal Propulsion Algorithm for Laser Centerline Extraction

Precisely extracting the laser centerline is one of the key technologies in line-structured light reconstruction systems, In response to the limitations of traditional internal propulsion algorithms, which are not suitable for extracting multiple laser lines and often lack precision in centerline extraction, An improved laser centerline extraction algorithm based on internal propulsion has been proposed. The problem of the internal propulsion algorithm being unable to handle the extraction of multiple laser lines has been addressed by incorporating a labeling mechanism and enhancing the internal propulsion search mechanism. After extracting the initial laser center points, the tangent equation is determined through the least squares method, allowing for the calculation of the normal vector of the laser fringes. Subsequently, sub-pixel coordinates are recalculated along the normal direction, and a linear smoothing process is applied to make the laser fringes smoother. Experimental results indicate that the root mean square error of this algorithm, compared to the gray-gravity method, has increased by 0.337 pixels, and compared to the traditional internal propulsion algorithm, it has increased by 0.128 pixels. It can stably extract features even in low-luminance regions.