Automated Visual Inspection Model For Screw Detection on The Moving Objects In Industrial Quality Control

Modern industrial automated visual inspection models provide new methods and solutions in industrial visual quality control application. In this paper, we present automated visual inspection (AVI) model based on the normalized cross-correlation template matching algorithm for realtime screw detection on a moving object. The model can recognize products and detect screws on moving objects. We present the implementation challenges and provide guidelines for successful industrial application. We present experimental results, and discuss model constraints, and implementation parameters settings, showing that the proposed model is very sensitive both to object distance from the camera and a small rotation of the object during template detection on moving objects. The problem of vibrations in real industrial environments and our proposed solution that signi_cantly improves model accuracy is also presented. Our findings show that the model can outperform humans in visual quality control process.