An efficient and accurate method for real-time processing of light stripe images

This paper describes a direction-guided method for real-time processing of light stripe images in practical active vision applications. The presented approach consists of three main steps. Firstly, we select two thresholds to divide the image into three classes, that is, the foreground, the background, and the undecided area. Then, we locate a feature point of the light stripe in the image foreground and calculate the local line direction at that point. Finally, we detect succeeding points along that direction in the foreground and the undecided area and directly exclude pixels which are in the normal direction of the line. We use synthetic and real images to verify the accuracy and efficiency of the proposed method. The experimental results show that the proposed algorithm gains a 1-fold speedup compared to the prevalent method without any loss of accuracy.