Performing Post-Production Defect Detection of Pet Preforms With Image Processing Technology

Öz

With the speed and cost advantage provided by the developing technology, image processing technology has started to become widespread especially in the production sector. In this study, the defects in pet preform production were examined and in this direction, it was aimed to eliminate the errors by using an image processing system. Preforms found to be faulty by image processing are distinguished with great precision. For this purpose, the design and manufacture of the mechanical test station was made. In this test device, preforms can be adjusted according to product dimensions, products that pass the test can be moved to the stock box. In this study, features including statistical, textural and morphological features of the patterns obtained after segmentation of the separated product image were determined. It has been proven that the preforms show classification success with 97% accuracy in terms of classification precision. Based on Tables 1 and 2, an enterprise producing 30 million units monthly experienced a loss of 8.5 million units due to adverse conditions. This resulted in significant inefficiencies in energy use, labor, raw materials, and machinery. To address these issues, an image processing test device was developed and implemented in 2021.

Anahtar Kelimeler

• Keywords: Image processing
• Test device
• Preform Defects

Görüntü işleme Teknolojisi ile Pet Preformlarının Üretim Sonrası Kusur Tespitinin Yapılması

Öz

Gelişen teknolojinin sağladığı hız ve maliyet avantajı sayesinde, görüntü işleme teknolojisi özellikle üretim sektöründe yaygınlaşmaya başlamıştır. Bu çalışmada, PET preform üretimindeki kusurlar incelenmiş ve bu doğrultuda, görüntü işleme sistemi kullanılarak hataların giderilmesi amaçlanmıştır. Görüntü işleme ile hatalı olduğu tespit edilen preformlar yüksek hassasiyetle ayrılabilmektedir. Bu amaçla, mekanik test istasyonunun tasarımı ve üretimi gerçekleştirilmiştir. Bu test cihazında, preformlar ürün boyutlarına göre ayarlanabilir ve testi geçen ürünler stok kutusuna yönlendirilebilir. Çalışmada, ayrıştırılan ürün görüntüsünün bölütlenmesinden sonra elde edilen desenlerin istatistiksel, dokusal ve morforlojik özellikleri belirlenmiştir. Preformların sınıflandırma doğruluğu açısından %97 başarı oranı ile sınıflandırma gerçekleştirdiği kanıtlanmıştır. Tablo 1 ve 2'ye göre, aylık 30 milyon birim üretim yapan bir işletmede olumsuz koşullar nedeniyle 8,5 milyon birim kayıp yaşanmıştır. Bu durum, enerji kullanımı, iş gücü, hammadde ve makinelerde ciddi verimsizliklere yol açmıştır. Bu sorunları gidermek amacıyla 2021 yılında bir görüntü işleme test cihazı geliştirilip uygulanmıştır.

Anahtar Kelimeler

• Görüntü işleme
• Test cihazı
• Preform kusurları

Kaynakça

1. Malesa M, Rajkiewicz P. Quality control of pet bottles caps with dedicated image calibration and deep neural networks. Sensors 2021;21(2):501. doi:10.3390/s21020501.
2. Balcı M, Altun AA, Taşdemir Ş. Classification of napoleon type cherries by using image processing techniques. Selcuk Technical Journal 2016;15(3):221-37.
3. Samtaş G, Gülesin M. Digital image processing and its applications in different fields. Electronic Journal of Vocational Colleges 2021;2(1):85-97.
4. PAGEV. Available from: https://pagev.org/ [Accessed 27 May 2025].
5. Bonnot N, Seulin R, Merienne F. Machine vision system for surface inspection on brushed industrial parts. Machine Vision Applications in Industrial Inspection 2004;5303:64-72.
6. Chen T, Wang Y, Xiao C, Wu QM. A machine vision apparatus and method for can end inspection. Transactions on Instrumentation and Measurement 2016;18(56):1-11.
7. Singh S, Kaur M. Machine vision system for automated visual inspection of tile’s surface quality. IOSR Journal of Engineering 2012;2(3):429-32.
8. Nashat S, Abdullah A, Abdullah MZ. A Machine machine vision for crack inspection of biscuits featuring pyramid detection scheme. Journal of Food Engineering 2014;120:233-47. doi:10.1016/j.jfoodeng.2013.08.006.

9. Pithadiya KJ, Modi CK, Chauhan JD. Comparison of optimal edge detection algorithms for liquid level inspection in bottles. In: Proceedings of the 2nd International Conference on Emerging Trends in Engineering and Technology (ICETET '09), India; 2009, p. 447–52.
10. Seulin R, Merienne F, Gorria P. Machine vision system for specular surface inspection, Use of simulation process as a tool for design and optimization. Regional Council of Burgundy; 2001.
11. Li Y, Li F, Wang Y, Tan M. Measurement and defect detection of the weld bead based on online vision inspection. Transactions On Instrumentation And Measurement 2010;59(7):1841-9.
12. Zhang H, Li X, Zhong H, Yang Y. Automated machine vision system for liquid particle inspection of pharmaceutical injection. Transactions on Instrumentation and Measurement 2018;18:1-20. doi:10.1109/TIM.2018.2800258.
13. Degtyareva KV, Nikolaev SV, Nelyub VA, Tynchenko VS, Borodulin AS. Automatic monitoring system designed to detect defects in PET preforms. Web of Conferences 458 2023;02002. doi:10.1051/e3sconf/202345802002.
14. Choi YW, Lee SW. Development of a Defect Diagnosis Algorithm for Blow-molded Transparent Plastic Bottles based on Convolutional Neural Networks CNN. Int J Precis Eng Manuf-Smart Tech 2025. doi:10.57062/ijpem-st.2024.00122.
15. Laucka A, Andriukaitis D, Valinevicius A, Navikas D. Computer Vision System for Defects Detection in PET Preform. In: 21st International Conference on Methods and Models in Automation and Robotics (MMAR); 2016. doi:10.1109/MMAR.2016.7575323.
16. Jingni P, Shujuan L, Yan L. A real-time surface defects detection model via dual-branch feature extraction and dynamic multi-scale fusion attention. Digital Signal Processing 2024;152:104582. doi:10.1016/j.dsp.2024.104582.
17. Adin MŞ, Kam M. An Overview of Post-Processing of Fused Deposition Modelling 3D Printed Products. 1st ed. CRP Press; 2024, p. 10. doi:10.1201/9781032665351-1.