Abstract :

Industrial Components Are Continuously Exposed To Demanding Operational Environments, Making Them Vulnerable To Defects Such As Cracks, Corrosion, Weld Imperfections, Overheating, And Surface Degradation. Early And Accurate Defect Detection Is Essential For Ensuring Operational Reliability, Improving Safety, And Reducing Maintenance Costs. Traditional Inspection Methods Rely Heavily On Manual Assessment, Which Is Often Time-consuming, Labor-intensive, And Prone To Inconsistencies. To Address These Limitations, The Proposed Framework Introduces An Intelligent Industrial Defect Detection System That Combines Deep Learning (DL) And Machine Learning (ML) Techniques For Automated Defect Identification And Classification. The Framework Initially Employs A Hybrid ConvLogiDefect (CLD) Model That Integrates Convolutional Neural Networks (CNN) For Deep Feature Extraction And Logistic Regression (LR) For Defect Classification. In Addition, Conventional ML Models, Including K Nearest Neighbors (KNN) And Decision Tree (DT), Are Implemented To Provide Comparative Performance Evaluation. To Further Enhance Detection Accuracy, The Framework Incorporates An Advanced TransForestDefectNet (TFD-Net) Model Based On Data-efficient Image Transformer (DeiT) And Random Forest Classifier (RFC). The DeiT Architecture Captures High-level Transformer-based Visual Representations, While RFC Improves Classification Robustness And Multi-class Defect Prediction. The System Is Implemented Using A Tkinter-based Graphical User Interface (GUI) That Supports Dataset Upload, Preprocessing, Model Training, Visualization, And Real-time Prediction. Furthermore, A Secure Authentication Mechanism Is Integrated To Restrict Unauthorized Access. Experimental Results Demonstrate That The Proposed TFD-Net Model Significantly Outperforms CLD, KNN, And DT Models, Providing A Scalable, Accurate, And Efficient Solution For Automated Industrial Defect Detection In Smart Manufacturing Environments

Published:

31-5-2026

Issue:

Vol. 26 No. 5 (2026)

Page Nos:

1866 -1877

Section:

Articles

License:

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

How to Cite