Using image analysis methods in quality control of ready-made products

The quality control of the final product in the ready-made industry, is made by operators with their own subjective methods. The operators are trained experts who are quite experienced to detect quality violations and fixed the ones when possible or mark them to be corrected by the related party. The decision about the quality of the product is given according to the practical limits those imposed by customer, or the upper limits specified based on experiences.Whereas operators as the human beings has some weaknesses which could cause faults that could lead to lack of quality and waste of time when they detect the defects and decide about quality control violations. The factors such as gradually loose of attention, fatigue and stress could affect the efficiency of quality control which is made by human beings who might be subjected to these effects easily. Based on these while considering costs, reliability and international acceptance of the quality controls, it becomes necessary to develop and use of obcetive, quick and economic new systems and mechanisms to establish mutual trust between producers and customers. Today, in textile and ready-made industry image recognition and analysis systems are being used as fast and effective methods to control of the products. Studies highlight the effectiveness of this methods for fiber, yarn and fabric surfaces that are being controlled by these image recognition and analysis systems By using image analysis methods in quality control of ready-made products it will be possible to avoid the human defects and provide the required standards of viiquality. Quality control process will be done in less time; cost will be reduced by ensuring the quality levels and decreasing number of operators.The purpose of this study is to investigate and specify the defects in textile and measurements by image processing/analysis in the last phase of quality control of ready-made products.The study is mainly based on trial methods. System developed in the laboratory and final test and trials were conducted in the ready-made industry facility. Preliminary studies were done in the X Giyim with t-shirt samples and research studies is conducted in the Y Giyim where 43 random basic, white, 38 size t-shirt of their production were chosen and inspected. An image processing software developed specially for this application was used in both phases. The software is developed according to specifications and quality control parameters of the sample groups.For the data acqusition, 640x480 / 1024x768 resolution digital camera was used and the image is prosessed using Matlab® software on a computer with MS Windows® operating system.Quality control results of the image prosessing system is investigated and compared with the quality control which is done by human operator?s eye inspection. The Wilcoxon Mann-Whitney Test, is one of the most powerful of the nonparametric tests for comparing two populations, is conducted to find the differences between the two quality control methots.Results indicate that image processing software system is quite effective in the quality control of t-shirts which is one of the major ready-made industry products. Acquired data shows that the deviations in the quailty control is within the tolerances when the image processing ssytem is used. This is an important indicator for the image processing system?s effectiveness. In the case of examining the fabric defects in the sample groups, the image processing system is also seen to be quite successful and the defects were identified. When working with sample groups, it is seen that the resolution of the digital camera and the light illumination is very important for correct results.viiiAs a final result, throughout the works already conducted within the literature, this study shows clearly that the image analysis methods can be used in the quality control of products of the ready-made clothing industry. Therefore this study will take its place in the literature as a new quality control method and will make a basis for further studies in this field.Keywords: Ready-madeProducts, Image Analysis Methods, Quality Control