Timed evaluation of the center extraction of a moving laser stripe on a vehicle body using the Sigmoid-Gaussian function and a tracking method

Abstract

We timed the extraction of a moving laser stripe from a video using the Sigmoid-Gaussian function and the Hessian matrix method. A tracking model using the Kalman filter method was used to analyze the moving laser stripe projected on the vehicle body. The small region of the laser stripe predicted the subsequent frame based on the prior frame. To verify the extraction of the centers and the tracking effects of the moving laser stripe using the Kalman filter method, two videos of a moving laser plane scanning the vehicle body were captured in the experiments. The region including the laser stripe was predicted from the video frames. Moreover, we conducted two comparative experiments that provided 20 frames from the 10th frame to the 200th frame to evaluate the time and time reduction rate of the extraction of the laser stripe centers using this method and without tracking. The mean time reductions of the extraction in the two tests were 24.14% and 32.77%, respectively. Therefore, the time to extract the centers of the laser stripe was quantified using a Sigmoid-Gaussian function and a tracking method.