Automatic Screening system for Mechanical parts using Harris Corner Detection algorithm based Artificial Neural Networks

In recent years, the field of machine vision for automatic screening system is a growing technology of mechanical automation model. However, due to the technological developments in smart factories, the mechanical component inspection and measurement requires efficient precision and accuracy. Therefore, Harris Corner Detection Algorithm based Artificial Neural Networks (HCDA-ANN) are recommended in this research in order to construct an intelligent system. It integrates machine vision and artificial intelligence networks to automatically screening for mechanical components. Using the system calibration, the camera pixel is first calibrated to the actual distance of the object is evaluated. The next step in the measurement and inspection process is to take images and use the Vision 2D model to analyze. Finally, the recommended HCDA-ANN is employed in an automated screening process for mechanical parts to determine whether the metals is non-defective or defective parts. According to the results, it illustrated that the suggested HCDA-ANN method achieved better performance in terms of Mean Square Error (MSE) of 1.55 and Mean Absolute Error (MAE) of 2.32 which is comparatively higher than existing models namely Design of Experiments (DOE), Zero Defect Manufacturing (ZDM) and Automatic Testing Model (ATM).