Extracting stone slab contours using sample windows and color information

Contour extraction of stone slabs is the first-step for optimization of cutting the irregular stone slab in real time. A color image segmentation method to extract the stone slab contour in the complex background is proposed. Firstly, we define a sample window and detecting window in HSI color image space, and extract the image feature of sample window as the prior knowledge. Traverse the whole image with the detecting windows and then obtain the candidate regions based on the similarity of probability distribution between the sample window and detecting windows. Then, determine object regions in which the H value of the pixel belongs to the double threshold range from the candidate regions. Finally, extract stone slab contour from the object regions. We extract stone slab contour in simple and beltline complex background using our method. Experiment results show that our method can reliably and accurately extract stone slab contour with simple and beltline complex background.