Medical Bottle Package Defects Detecting System Based On Machine Vision

In the pharmaceutical packaging process, it appears easily sealed and lax circumstance for the bottle and cap. Therefore, the quality detection for the drugs packing becomes an important part in the process of the pharmaceutical industry production. This paper presents a medical bottle detecting system based on machine vision, and completing the package defects detection task for the installation of bottle and cap. The main researches and contributions of this thesis are as follows:1. Referring to the advanced achievements in medical bottle detection, a successful hardware experimental platform is built according to the practical requirements. Meanwhile, this paper clearly explained the selection principles and final parameters for the platform.2. Starting from the concept of digital image processing, we inspected the correlated algorithms for medical bottle images. Firstly, using MATLAB software for operations such as gray processing, image de-nosing, image enhancement, image segmentation and morphology processing; secondly, extracting the corresponding region of the medical bottle and making boundary signature by using 8-direction chain code method; then, calculating the number of pixels of the connected region and working out the characteristic values; finally, we can determine whether the medical bottle is qualified by the minimum distance method of template maching. In addition, we improve the defects detection scheme through increasing the image registration and subtraction operations, which makes the detection accuracy reached to 95%.3. The software part of this article uses MATLAB programming techniques, achieving good man-machine interface. The main interface is completed by MATLAB GUI. Users can be very intuitive and easy to interact with the computer.This thesis provides satifying solutions to most problems of medical bottle inspector. Thoretical study and practical experiments have proved the efficiency and feasibility of the algorithm for detection. Further research is supposed to benefit from the achievements of this thesis.