IDENTIFICATION OF CASTING PRODUCT SURFACE QUALITY USING DL

COMPUTER SCIENCE ENGINEERING

The evaluation of casting objects' quality is crucial for ensuring their dependability and effectiveness. Traditional methods for evaluating surface quality often utilize manual examination, which is time-consuming and unreliable. Recent research shows that deep learning (DL) methods have great promise for automating the detection and classification of various manufacturing process flaws. In this study, a unique method for assessing the surface quality of casting products using DL is presented. The suggested technique analyzes photos of casting surfaces and categorizes them using a convolutional neural network (CNN) architecture. A sizable dataset of annotated images covering a variety of flaws, including cracks, porosity, and surface roughness, is used to train the CNN model. Data augmentation techniques are used to expand the diversity of the training dataset and improve the accuracy of the DL model. Additionally, transfer learning is used to benefit from previously trained models and enhance the network's generalization capabilities. The performance of the trained DL model is then tested on a different validation dataset to assess how well it can detect surface flaws. The results of the experiments show that the suggested DL-based approach is highly accurate at determining the surface quality of casting products. The model outperforms conventional manual inspection techniques in terms of resilience in identifying and categorizing various types of flaws. For manufacturers in the casting industry, the automated aspect of the DL technique considerably decreases the time and effort needed for quality inspection. This study advances the field of quality control in manufacturing by demonstrating how DL approaches can be used to detect casting product surface flaws. The suggested method can be enhanced with real-time monitoring devices to provide continuous quality evaluation during the casting process. This research emphasizes how DL can increase the effectiveness and precision of surface quality evaluation in the casting industry.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.