Laser Stripe Center Extraction Based on Ridgeline Tracking Algorithm

Abstract:
In the online laser 3D vision measurement, the most important step is the extraction of the laser stripe center. Aiming at the problem that the existing laser stripe center extraction methods cannot balance the accuracy and speed, a new laser stripe center extraction method is proposed in combination with the principle of ridgeline tracking. The algorithm first uses Otsu algorithm segmentation and morphological top hat transformation to extract the binary mask, and uses the mask operation to preprocess the median filtered image to effectively remove large-area noise interference; then select the starting tracking point for the ridgeline tracking, use the mean square gray gradient method to calculate the normal and tangential direction of the tracking point, and uses the weighted gray centroid method to extract the sub-pixel center point in the normal direction, and continuously tracks the new ridge point until the entire laser strip center is extracted. The experimental results show that the proposed method can effectively reduce the noise interference, the extracted laser stripe center is smooth and flat with small standard deviation. Compared with the Steger method, the extraction speed is increased by about 22 times, which meets the requirements of industrial detection systems.