Automated Ceramic Tile Defects Inspection System Based on Texture Analysis
No Thumbnail Available
Date
2013
Journal Title
Journal ISSN
Volume Title
Publisher
Uva Wellassa University of Sri Lanka
Abstract
The ceramic tile manufacturing process has now been completely automated. When it comes to
the final stage of the production, there are different types of surface defects such as holes, dirty
spots, color bleeding, etc. Some of the defects are difficult to distinguish with the traditional
techniques such as visual inspection. (Boukouvalas et al., 1994)
The proposed system is to develop an integrated system for the detection of defects on ceramic
tiles based on image processing and embedded system.
Methodology
Instead of human vision, computer vision system is used to classify the tile. First of all a picture
of the tile is taken and image processing steps are applied to image. This stage includes pre-
processing, image enhancement and feature comparison techniques. After getting the data
consisting of the frequencies of masks, RULES-3 algorithm (Aksoy, 1993) is employed to
determine the level of defect of the tile based on the features extracted and then pattern
recognition is performed.
Based on the output produced by the image processing system, the microcontroller and the
camera are controlled. In this process camera takes a picture of the tile and send it to the image
processing program. After analyzing the image, relevant signal is generated by the system and
passed to the microcontroller to control the production line.
Due to the availability of pre-defined libraries for the implementation, Visual Studio framework
has been used for implementation and 16F877A microcontroller is used as the controlling
component.
Result and Discussion
The above automatic inspection procedures have been implemented and tested on a number of
tiles using synthetic and real defects. Accuracy of color object recognition part depends on light
condition of the environment and accuracy of the results of image processing program depends
on the quality of the image. Further, the micro-controlling process (hardware component: figure
1) depends on the results produced by the computer system (application software: figure 2). The
results suggest that the performance is adequate to provide a basis for a viable commercial
visual inspection system.
Description
Keywords
Computer Science, Science and Technology, Technology, System, Software Developing