Based On The Image Of Beer Bottle Mouth And The Automatic Detection Of The Bottom Defaced

It is important for small and medium-sized beverage producers that developing one kind of automatic system to realize on-line beer bottle detection. It is benefit for corporations to improve product quality and adapt the new techniques. How to implement such system based on digital image processing is the major of this thesis.This thesis carefully analyzed the model of brokenness and dirtiness, the exterior characteristics and the light transmission characteristics of beer bottles; studied and implemented one interpolation method to extend a circular region to rectangle; investigated and adopted one image enhancement method to recorrect the uneven illumination; proposed one method based on connectivity analysis to detect the broken occurred on bottle' s outer wall; modified the Hough Transform for circle to make a fast bottle locating; implemented one method based on projection to detect brokenness and dirtiness after the original image segmented and morphological filtered.Many experiments had made to testify the validity of methods proposed in this thesis.