Real-time quality analysis of baked goods using advanced technologies

• Published in: Journal of food engineering 2025-03, Vol.388, p.112359, Article 112359
• Main Author: Görgülü, Ahmet
• Format: Article
• Language: English
• Subjects: Bakery products; Color control; Geometrical definition; Image processing; Water activity
• ISSN: 0260-8774
• DOI: 10.1016/j.jfoodeng.2024.112359

Some critical quality variables of baked goods baked in tunnel ovens cannot be monitored in real-time and quickly after baking. Generally, these variables are measured individually and manually. Analysis of some variables can be done in laboratory environments. It is very difficult to continuously and accurately read the quality variables of products, especially at the exits of industrial tunnel ovens operating at high speeds. With current systems, important quality variables of the products are assessed through widely spaced sampling, and most of the time the measurements made do not represent the entire batch. Additionally, interventions in the process may be delayed due to the time required for analysis, allowing defective products to reach consumers. In this study, experiments were carried out to enable real-time and high-speed identification of important quality variables such as product geometry, surface color, image features, and moisture. The developed system aims to monitor products such as cookies, cakes, biscuits, crackers and bread. Through this system, the size of the products that can represent the lot or even the entire batch can be analyzed when desired. Thus, quality oscillations can be kept within very narrow ranges. The developed hybrid system is applied to the exit of tunnel ovens, and the data read through sensors and cameras are transferred to computers using specially developed algorithms through interfaces. It has been observed that process control signals can also be generated by considering the lower and upper control limits of quality variables. With the developed software, changes can be instantly seen on the screens in the form of graphics, audio, and visual warnings can be created, historical data can be recorded, and reports can be obtained. At the end of the study, it was found that important quality variables such as circularity, diameter, thickness, color, image area, and moisture could be read with a sample size of up to 100%. As a result of the experiments, it was seen that the change in product diameters could be defined within the range of 40.30±0.5 mm. The circularity ratio between horizontal and vertical diameters can be monitored within 1.008 ± 0.018%. The change in moisture within the range of 2.741±0.164 g/100g. The light-dark image area ratios within the range of 0.486±0.033%, and the product thickness within the range of 11.56±0.178 mm. Resultly, all critical quality parameters taken into account can be detected