Area measurement method of paint bubble based on computer vision

An area measurement method of paint bubble based on computer vision is proposed. Firstly, through a camera calibration, the lens distortion is removed and the mapping between image coordinate and real world coordinate is constructed. Secondly, a novel color image segmentation method based on graph-cut, whose seeds can be auto-marked, is proposed, and with the method, paint bubbles are extracted from image. Finally, with camera calibrated parameter, the size of segmented paint bubbles is figured out. Experiment result shows that the proposed method gains higher precision than classical graph-cut segmentation method and can figure out the area of the bubble automatically.