Scribd
Upload
6 views21 pages

Final PPT For Review

The document outlines a research project focused on developing an automated system for detecting defects in ceramic tiles using digital image processing techniques. The proposed methodology includes image acquisition, enhancement, defect detection, and comparative analysis to improve product quality and reduce waste in the tiles industry. The study aims to enhance existing defect detection methods and emphasizes the need for further work on defect classification.

Uploaded by

chintu.kalita.adtu
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
6 views21 pages

Final PPT For Review

The document outlines a research project focused on developing an automated system for detecting defects in ceramic tiles using digital image processing techniques. The proposed methodology includes image acquisition, enhancement, defect detection, and comparative analysis to improve product quality and reduce waste in the tiles industry. The study aims to enhance existing defect detection methods and emphasizes the need for further work on defect classification.

Uploaded by

chintu.kalita.adtu
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 21

CERAMIC TILES DEFECT DETECTION

SYSTEM

Guided
qss
By:
Dr. Aniruddha Deka
Associate Professor

Presenter:
Presented by:
Chintu Kalita
Vivek Kalita
Amandip Singh
Content
• Introduction • Training
• Literature Review • Performance Evaluation
• Problem Statement • Frontend
• Objectives • Sample Code
• Gantt Chart • Testing
• Proposed Methodology • Conclusion
• Types of defects • Reference
• Flow Chart
• Dataset
• Pre-Processing Images
• Labelling of Database
Introduction

• The manual method is labour intensive,slow and reduces market competitiveness.

• Human judgement is influenced by prior knowledge and often subjective.

• The objective of our research is to propose an efficient automated ceramic tiles defect detection
and classification technique.

• To address the quality standards and have a sustainable approach to the tiles industry.

• Will facilitate market competitiveness of Ceramic tiles industry.


Literature Review
PROBEM STATEMENT

Defects such as cracks and surface irregularities can occur, impacting the quality and usability of
tiles.Manual inspection is time consuming and labour intensive and prone to various factors.
Objective

Develop a reliable and automated system using Digital image processing techniques to detect
different types of defects in ceramic tiles during the production process ensuring product quality and
reducing waste
Gantt Chart
Proposed Methodology

1. IMAGE ACQUISITION & PREPROCESSING


- OBTAIN IMAGE OF CERAMIC TILES AND PREPROCESS THEM

2. IMAGE ENHANCEMENT
- DESIRED IMAGE FEATURE BECOME VISIBLE

3. DEFECTS DETECTION
- EDGES,BLOB,CRACK,SCRATCH, HOLE
ddddasasss
4. COMPARITIVE ANALYSIS
- DEFECTIVE TILE
- NON DEFECTIVE TILE

5. RESULTS
TYPES OF DEFECTS

ddddasasss

Edge-Chipping Hole Line


[[[[[

Flow Chart
Flow Chart Image Acquisition and pre
processing

Lighting Correction

edge, Line, and Hole Detection


CERAMIC TILES

Segmentation

Feature Extraction

Classification

Defective Non Defective


DEFECT DETECTED
SCRATCH / CRACK
/ BLOB

Comparitive
analysis NO DEFECT DETECTED

Results
Dataset

The database will be of tiles images, with manual annotations of the objects we acquired
after feature extraction.
It will have train and validation directory with images and txt files which contains the labels
Here is a overview how the dataset will look like -
Pre Processing Image

Images are resized to 640 x 640 and normalized with noise reduction.
Augmentations like rotations were applied too.
Labelling of Database -

After Preprocessing and feature extraction tiles were manually labelled.


Training
For training the model we used YOLO.
YOLO (You Only Look Once) is a state-of-the-art, real-time object detection system that predicts bounding boxes
and class probabilities directly from full images in a single evaluation.
The model has been trained for 50 epochs, resulting in a final validation on the weights saved as best.pt
This YOLOv8 model, utilizing a Tesla T4 GPU, comprises 268 layers with a total of 68,126,457 parameters
and operates at 257.4 GFLOPs
Performance Evaluation
FRONTEND
SAMPLE CODE
Testing
Conclusion

▪ WE ARE STUDYING EXISTING DEFECT DETECTION METHOD FOR FURTHER ENHACEMENT.

▪ MORE WORK NEEDS TO BE DONE ON DEFECT CLASSIFICATION.


References

• A Machine Learning Method for Detection of Surface Defects on Ceramic Tiles Using Convolutional Neural
Networks (2023) - R. Zhu et al., Sensors

• Ceramic tile surface defect detection based on deep learning (2022) - Y. Zhang et al., Procedia Computer Science

• Detection of Defects in Ceramic Tiles Using Deep Neural Network in Artificial Intelligence (2020) - S. Boovaneswari
et al., International Conference on Research and Innovation in Information Systems

• Improved Canny Edge Detector in Ceramic Tiles Defect Detection (2018) - M. Zareei et al., International Journal of
Engineering and Technology
Thank You.
Questions?

You might also like