The Research And Implementation Of Fastening Tape's Flaw Fast Detection Method Based On Machine Vision

The quality control is a very important in the production of fastening tape, it often relates the question of the enterprise life. But in the present, the way of flaw examination was still completed by artificial. Not only the low efficiency, moreover it easy to make wrong examines with the omission, and brings the hidden danger for the product quality. Therefore, this article make the research about the visible defect examination of the fastening tape based on the machine vision, has emphasis studied system's hardware build structure and software's choice. In view of the image algorithm in fastening tape, proposed and develops a set of the fastening tape's visible defect examination machine.First, this article is in the view of specific breadth of the fastening tape and the examination request of the high accuracy and velocity. It was full considering fastening tape's characteristics in image gathering, aspect and so on flaw division, classification. This article Proposed each kind of hardware's choice, special in the aspect of illumination and the gathering, have guaranteed the system frame is stably reliable.Make the analysis the unique textural property and the majority noises of fastening tape, this article studies auto-adapted filter based on the average value. It full displayed the respective merit of the linear filter and the non-linear filter, could solution the noise question in image gathering process effectively.In order to raise the system's efficiency and reduce the process time of each picture, this article designs a method to judge the flaw if existence, and only carries the following processing to the image which with flaw, the rest of the image as the part of surface demonstration. By saved lot of time in system's operation, and it's suit in the online examination system.In view of fastening tape unique breadth, this article proposed an algorithm based on the first-order characteristic value. Reject the image not to be interested, and operand big algorithm processing on region which must be examined. The experiment indicated that the operation time be saved, moreover the edge is more precise. Because fastening tape's both sides of the textural property are different, therefore this article proposes an examination algorithm to the texture of fastening tape surface.Then, to use the BP neural network to divide the flaw which carries on classified processing, and hit sign processing according to the factory request. Later, make the detail explanation to the manufacture of lower position machine.Finally, in the foundation of the above research the method, this article developed a set of fastening tape visible defect examination machine based on the machine vision. The experiment indicated that this equipment online economy examination speed can achieve 100m/min, the highest examination speed can achieve 200m/min, the highest examination precision is 0.5mm, and have highly flaw recognition rate, and it could meet the actual production needs well.