摘要 |
PURPOSE:To efficiently and surely recognize the state of the component in a short time without adding any extra mechanism. CONSTITUTION:A mask image is put over an input image to extract a necessary part of the component shape, and the extracted image is tracked to extract feature quantities, which are combined to decide the state of the component. The input image is grasped at a time, the only necessary part is extracted from the input image, and the feature quantities are extracted from the extracted image to decide the state of the component according to whether the feature quantities match a prescribed state or not, so the state of the component can efficiently and surely be recognized in a short time without adding any extra mechanism. |