Study On System Of Capsule Characteristics Identification And Direction Shift Based On CCD

With the development of science and technology, it's important to arrange capsules in ordered in the field of drug manufacturing and packaging. It's obviously that the capsules'ordered arrangement can not only improve the quality and level of products, but also be beneficial to production. At present, the major method achieving capsule ordered is mechanical detection technique. There are efficiency and accuracy defects for this technique.Firstly, the capsule images collected by CCD are processed by non-uniformity correction and grey level transformation. And the result of image binarization is improved through partial enhance method. On the basis of histogram of oriented gradients and Hough transform, the direction detection algorithm for the capsule is advanced and simulated. Cansidering the speed requirements of real-time systems, this article derives the one-dimensional edge model of capsule, using least-squares linear regression theory. According to this model, infrared ordered arrangement for capsules is self-designed. The capsule image characteristics are collected by four pairs of infrared photoelectric sensor, which is possessed of the characters of higher reliable, higher processing speed, simple architecture, and lower cost. Based on the theory of the sticky wall effect on fluidics, the accelerating and pneumatic overturn mechanism with streamlined slipway and high-speed valve for capsules is designed. The high-speed valve driver based on PWM is also designed, which achieves the purpose of fast startup and shutdown and energy conservation. According to the experiences in field, the simplified detection system can work cooperatively with overturn mechanism, possessing high utility value.