In the automated production process of toothbrushes, the manufacturing of toothbrushes is done by machinery, it is inevitable that there will be various defects caused during production. These defects will directly cause the product quality to fall and reduce the overall production quality, therefore the detection of these defects is particularly important.
Although the technical requirement for unqualification of thin bristle detection is 7 bristles per root, but 2-3 bristles need to be judged as unqualified during actual acceptance. The recognition of bristles and qualified products is not high, and the surface is very irregular. The defect location is not fixed, and the space environment does not allow multiple light sources to be set up
With the advantage of VisionBank’s artificial intelligence vision system with learning algorithms, our solution provides customers with intelligent visual inspection solutions for toothbrush bristles that have successfully met the detection accuracy of less than 0.1mm. The missed detection rate is less than 0.2%, and the false detection rate is less than 4 % the requirements.
In the project delivery phase, we tested 500 products, and the detection rate and false detection rate reached the customer’s acceptance requirements. The detection accuracy was increased from the customer’s requirements of 0.1mm to 0.05mm.