基于机器视觉的铝塑泡罩包装药品缺陷检测-测试计量技术及仪器专业论文.docxVIP

华 华 中 科 技 大 学 硕 士 学 位 论 文 II II Abstract Medicine in the aluminum- plastic blister packaging has received more and more favor among consumers for its sealed package and convenience. However, the defects are inevitable, such as missing tablets or capsules, broken tablets and so on, which will affect the quality of the products. Nowadays, manual inspection is generally adopted in many pharmaceutical companies, which is labor-intensive and low efficient. Machine vision method could provide a solution to solve these shortcomings, and has better reliability and robustness. This thesis presents an automated visual inspection experimental system, which detects defects of medicines including tablets and capsules in aluminum-plastic blister packaging. The system operates either in training or inspection mode. In training mode, the images of each type of defect-free blister pack are used to extract the model and establish the template library. In inspection mode, the images of blister pack to be inspected are captured, the type of an inspected blister pack is recognized and the defects of the medicines in the blister pack are detected by matching with the template library. The main content of the thesis comprise: An algorithm based on the least squares fitting is proposed to extract the image of medicine board area. The algorithm uses the Otsu method for image binarization processing, the edge of the board is obtained by using scanning, and the accurate board area is obtained by the least squares fitting. The algorithm can satisfy the requirements of real-time detection. The color histogram is adopted to recognize the medicine type correctly. In training mode, create the color histogram of each type of medicines in the RGB color space. In inspection mode, the type of inspected medicine is recognized by calculating the distance of the defined color histograms between the inspected image and each model. Different detection algorithms are designed for the ta