An Intelligent Visual Defect Detection Mechanism for Enhancing Product Quality in Industries using Multi-scale EfficientNet B7 Model

for industrial products, artificial visual analysis is a conventional approach for performing quality control. The machine vision highly enhances the reliability, quality, and effectiveness of the defect identification. In visual analysis, the smart optimal illumination modules and appropriate image collection hardware are significant for acquiring high-quality images. The image evaluation and processing are primary mechanisms in attaining defect data, while deep learning is highly impacted with the image evaluation section. In the high-tech industries, the defect identification issue is highly important and is highly utilized for automated industrial quality control models. Therefore, a new visual defect detection model for industrial automation with the assist of neural network learning is introduced. At first, the data’s for visual defect detection are collected from benchmark sites. A Multi-scale EfficientNet B7 (MENetB7) mechanism is used to process the collected images and recognize the defects in the manufactured products. The multi-scale model analyzes the subtle defect patterns from background textures even if the products are in irregular shapes. The developed MENetB7 model does not increase the computational issues even when handling huge volumes of images. Finally, performance comparisons are conveyed out to check an effectiveness of the developed method in identifying the defects.