Research on key technologies of medicine grain defect detection system based on machine vision

Yueqiu Jiang, Yujun Wang, Hongwei Gao, Shuang Ma

School of Information Science and Engineering,Shenyang Ligong University,Nanping Str. 6, 110159Shenyang, China

In the process of medicine grain production may generate many kinds of defects. If these unqualified medicine granules are not timely detected, it will not only affect the company's reputation but also the health of the patient. This paper mainly studied how to detect the unqualified medicine grain base on machine vision. It mainly consists of three kinds of common defects segmentation and defect area calculation. Firstly, preprocess the medicine grain image for the following procedures. Secondly, obtain the defect region by improved segmentation algorithm, in order to deal with three different drugs grain defects this paper improved three segmentation algorithms, for the damaged tablets propose a local edge detection algorithm that based on grey level difference, for the irregular-shaped tablets adopts the ellipse detection algorithm, which based on Hough transform technology, for the Capsule has air uses the multi-scale Canny edge detection operator. Finally, adopt Chain code contour tracking algorithm and Three-point method calculate the area of the damaged tablets, determine whether the tablets meet the requirements. Experiments show that the system can detect unqualified medicine granule quickly and accurately, it is of great practical value.