摘要 |
PROBLEM TO BE SOLVED: To automatically judge the discrimination whether a bottle mouth is normal, cracked or defective by binarizing an input variable density image after differential processing, determining the position of the bottle mouth on the basis of the determined central position of the bottle mouth, and determining whether it forms a continued pattern or not. SOLUTION: The image signal from a video camera 5 is inputted into an image processing device 6 in a flow 21, and the central position of a bottle mouth is determined in a flow 22. In a flow 23, the pattern of the bottle mouth is cut and inspected on the basis of the central position determined in the flow 22. The bottle mouth is judged normal when a continuous pattern is judged by the inspection (flow 24), and defective when the pattern is discontinued. The rectangular area including the present area on the image of the bottle mouth is cut on the basis of the central position, and subjected to differential processing, whereby the boundary laid along the bottle aperture can be detected. Further, even when the reflectance of the bottle mouth is fluctuated, the binarized pattern can be obtained by binarizing processing to surely detect the profile length to the image. |