Research And Implementation On Capsule Defects Inspection Algorithms Based On Colored Image

With the increasing development of people’s living standards in the new century, problems for the security of medicine arise more and more by citizens and politicians. Manufacturer place a great emphasis on the quality of drug in order to the disqualified capsule damage people’s health in the market. To establish a comprehensive and stable capsule detection system has also been supported of each person in medicine. Various defects usually exist in the production of capsule due to influence imposed by processing and surroundings. Therefore, it is a must to inspect the flaws. Inspecting systems on market almost mainly involve the basis of gray images to achieve the purpose of finding out defects, by segmenting and quantifying. The study analyzes the multiple components of color image in different color spaces and provides a solution to low inspection efficiency in most cases when defects inspected using gray image. Experiments and practical applications have demonstrated the availability and efficiency of these algorithms.A encoder triggered sampling method has been proposed on the basis of the existing capsule inspecting machine based on gray image way. In this way, it can not only control the integrity of capsule but also reduces the CPU consumption. The system increasing the color camera in the original four gray cameras, uses the same PC as the hardware platform. It is based function library for HALCON algorithm and establishes the human-computer interaction system by MFC. The capsule system supplements each other between gray image and color image in order to realize the complete detection.In this article, an optical imaging system has been designed, which studies the characteristics of various defects capsule by through different color space of each component. The design can inspect various different capsule surface defects in different color image space of each component, using dynamic threshold segmentation, gradient threshold segmentation, and digital methods of morphology, so as to achieve the detection requirements for production process.