Inspection of Soldering Anomalies on Surface Mount Devices Using Image Processing & ANN

The quality of production in a company is a major guarantee. It needed a human resources, assistance company to control the product quality. However, human errors causing product failures are still found in many cases. Some disadvantages of using human power for the inspection process are fatigued in the eye when carrying out inspections using manual equipment such as magnifying glasses and microscopes. So in this study a system was designed to detect and identify anomalies on the soldering of ICSOIC on surface mount devices using image processing and neural networks. To separate the soldering colour from the SOIC IC, the HSL method is used while the identification of the type of anomaly in the SOIC IC used a backpropagation neural network. From the results of the system testing carried out, the success rate reached 92.72% from 11 trials.