Chapter 1IntroductionAt the end of the twentieth century, industries started to automatize their processes with the help of computation and automation 1.
These two areas are commonly employed in visual inspection of products. The allocation of humans to perform the visual inspection is therefore unnecessary if a proper computer vision system is implemented. The ceramic industry has a highly automated production system. The quality control, however, is still performed by humans, which limits its speed and precision. This work proposes a complete veri?cation system for ceramic tiles based on image processing and machine learning. The system has four steps: image acquisition, pre-processing, binarization and crack detection. The ceramic tiles inspection needs to be done by professionals.
However, opinions about the presence or absence of defects can diverge. The human capacity depends on training, knowledge and experience 3. The fact that quality control is done by humans brings several problems. People, in general, can work in this activity during some limited time and get tired in a couple of hours. Thus, the judgment is affected by fatigue 3. Another negative factor is the output rate of the quality control being lower than the production rate. To overcome this, automated system is developed.
Chapter 2Literature survey? 1, Boukouvalas et al.1 applied techniques based on a set of separable line filters, through textured tile crack detector based on the Wigner distribution and a conjoint spatial frequency representation of texture. Again, they applied winger distribution for crack detector and a novel conjoint spatial-spatial frequency representation for textured tiles. In terms of colour textured tiles, this type of detection algorithm which looks for abnormalities both in chromatic and structural properties. However, use of separate filtering technique for identifying distinct defect is not a good practice.
? The Wigner method was used by Song et al. 2 for crack detection in texture image. High computational time is taken while we are to handle a large number of operations during production time. It also proceeds with visual defect classification with human intervention. Se Ho Choi et al. applied a real time mechanism for surface flaw detection of steel coil and bar in high speed production environment.
? 3, Elbehiery et al. divided their method into two parts. In the first part, existing method consisted with the captured images of tiles as input.
As the output, they showed the intensity adjusted or histogram equalized image. First portion of this method consist with the captured image of tiles as input and output of this portion is histogram equalized image with intensity adjustment. After that, they use the output of first portion as input for the second portion. Furthermore, second portion also comprises with different complementary image processing operations so as to identify and to classify a variety of surface and structural defects. Their proposed system is not automated rather it emphasizes on the human visual inspection of defect classification in industrial environment. ? 4, Song et al.
Presented Chromato-Structural Defect Detection technique 372 was developed to detect both color and texture-formation defects in randomly textured ceramic tiles. They used a scheme named “edge preservation” for noise cutback and performance improvement. In addition, they used ” second derivative Laplacian ” filter to differentiate grey scale images from each other.
Finally, they applied “double thresholding” technique to formulate binary images. Still, this type of technique is unable to find the orientation of the edge of surface, because they use “second derivative laplacian ” filter which malfunctions for corner and curves flaw detection as well.? Merazi-Meksen et al. 5 have proposed a methodology for the automation of the ultrasonic image interpretation using the NDT technique called the Time of Diffraction technique which aid in decision making. In the context of pattern recognition, the signatures of regular patterns can be fairly easily isolated in either the spatial or spatial frequency domain.
Spatial frequency analysis is often preferred as it both decomposes the image into individual frequency components and establishes the relative energy of each component. Thus noise effects are also more easily separated.? Yiyang et al. 6 have proposed a crack detection algorithm based on digital image processing technology. By pre-processing, image segmentation and feature extraction, they have obtained the information about the crack image. We use the co-joint spatial and spatial frequency representation of the Wigner Distribution 4. This enhances pattern separability as the patterns’ signatures have disjoint support regions in the co-joint representation. ? Alam et al.
7 have proposed a detection technique by the combination of the digital image correlation and acoustic emission. The former method gives a very precise measurement of surface displacements, thus crack openings and crack spacing were determined.? Alvarez-Taboada et al.
8 In particular, the goal of SVM is to estimate an unknown continuous-valued function based on a finite number of noisy samples. Humans are able to find such defects without prior knowledge of the defect-free pattern. Defects are viewed as in-homogeneities in regularity and orientation fields. Two distinct but conceptually related approaches are presented. The first one defines structural defects as regions of abruptly falling regularity, the second one as perturbations in the dominant orientation.
? Akande et all 9 Basically, it makes use of the structural minimization principle which is known to have accurate generalization performance for different datasets size as contrasted to empirical risk minimization employed by other approaches like ANN. They applied techniques for pinhole and crack detectors for plane tiles based on a set of separable line filters, through textured tile crack detector based on the Wigner distribution and a novel conjoint spatial-spatial frequency representation of texture, to a colour texture tile defect detection algorithm which looks for abnormalities both in chromatic and structural properties of texture tiles. But, using separate filtering techniques for different types of defects is not a good idea at all, because in such case high computational time is a major issue for applying a large number of operations. Again, their procedure is an automated visual inspection system where they only show the defect.
Chapter 3Proposed SystemThis work proposes a complete veri?cation system for ceramic tiles based on image processing and machine learning. The system has four steps: image acquisition, pre-processing, binarization and crack detection.This work proposes a complete control quality system for ceramic tiles based on image processing and machine learning.
The image is read, and the enhancement is done. Later the image is converted to binary image. Binarization is the process of converting a pixel image to a binary image. Using for loop and matrix, the crack in image is detected. This system can detect corner damages, edge damages and middle cracks on the surface of the tile with high accuracy and efficiency.the result is represented in the form of graph and the accuracy is calculated. The proposed system is more reliable and takes less time for processing.
Block Diagram SummaryThis work presents a complete system for automatic defect detection of ceramic tiles using image processing and machine learning. For the classi?cation step, Support Vector Machine will be used. Feature extraction will be explored and appropriate feature extraction will be selected.
This system can be used to the current market ceramic tile industry as substitution to the manual inspection system to achieve high accuracy, efficiency and to overcome the other drawbacks of manual inspection system.References1 S. Arivazhagan, R. N. Shebiah, J. S.
Magdalene, and G. Sushmitha, “Railway track derailment inspection system using segmentation based fractal texture analysis.” ICTACT Journal on Image ; Video Processing, vol. 6, no.
1, 2015.2 S. H.
Hanzaei, A. Afshar, and F. Barazandeh, “Automatic detection and classi?cation of the ceramic tiles’ surface defects,” Pattern Recognition, vol. 66, pp. 174–189, 2017.3 ?Z. Hocenski, T. Mati´c, and I.
Vidovi´c, “Technology transfer of computer vision defect detection to ceramic tiles industry,” in Smart Systems and Technologies (SST), International Conference on. IEEE, 2016,4 G. Jacob, R. Shenbagavalli, and S.
Karthika, “Detection of surface defects on ceramic tiles based on morphological techniques,” arXiv.org – Cornell University Library, 2016.5 S.
Arivazhagan, R. N. Shebiah, J.
S. Magdalene, and G. Sushmitha, “Railway track derailment inspection system using segmentation based fractal texture analysis.” ICTACT Journal on Image ; Video Processing, vol. 6, no.
1, 201.
