Condom production often encounters such problems:
1. When the outer packaging is sealed, due to the deviation of the position of the condom in the package, the edge sealing machine may press the condom in the package into the edge sealing, causing the condom to break.
2. In order to facilitate tearing during use, condoms’ outer packaging often uses aluminum foil or PE. This makes the outer packaging easy to break during production, causing the personal lubricant inside to leak.
3. The sealing edge is not tightly sealed.
4. Personal lubricants are contaminated
Among them, the first and second problems have existed in the industry for a long time and have not been perfectly solved for a long time.
Difficulties in machine vision inspection:
1. To identify the position of the condom wrapped in the outer packaging, distinguish the condom stuck in the edge seal.
2. The outer packaging of condoms is often made of reflective silver material, and the light source requirements are very high when collecting images.
Now use deep learning to solve this series of appearance inspection problems. The industrial camera will transmit the captured image information to the trained AI model for analysis. The AI model will make judgments based on the raised contours, shadows, wrinkles, etc. of the condom on the outer packaging, and accurately identify the specific location through the outer packaging.
It is usually difficult to collect image information on the reflective outer packaging surface. These reflected lights will seriously interfere with the quality of imaging and reduce detection accuracy. To solve this problem, use optimized light sources and illuminate from different angles. In addition, adjust the angle between the industrial camera and the inspected object to avoid specular reflection of reflected light. In addition, optimize image processing and machine deep learning algorithms based on actual conditions.
