نوع مقاله : مقاله کامل علمی پژوهشی
عنوان مقاله English
نویسندگان English
Quantitative ash evaluation is considered an essential and standardized criterion for the accurate determination of bran content in wheat flour. Bran, as one of the valuable by-products of the white flour milling process, is a rich source of dietary fiber and is widely used in the enrichment of flour-based products, particularly in the production of whole-grain and high–nutritional-value breads. Precise control of this parameter plays a decisive role in the final product quality as well as in compliance with production standards. In this study, an intelligent model based on hyperspectral data processing was developed to classify different ash levels in wheat flour. To this end, following the acquisition and preprocessing of hyperspectral images, the extracted spectral data were analyzed under two separate frameworks: with optimal feature selection and without feature selection. Classification in both cases was performed using an artificial neural network (ANN) algorithm. The results demonstrate the remarkable effectiveness of the proposed model, achieving a classification accuracy of 98.3% without feature selection and 100% with feature selection. These findings not only confirm the high operational accuracy of the model but also highlight the inherent capability of artificial neural networks to identify complex spectral patterns and their strong generalization performance under varying experimental conditions. Consequently, it can be inferred that the integration of hyperspectral imaging technology with machine learning provides a novel, accurate, and non-destructive approach for monitoring and optimizing ash content in wheat flour. The application of this approach directly on production lines can lead to improved flour quality and, ultimately, the production of superior products with higher nutritional standards.
Quantitative ash evaluation is considered an essential and standardized criterion for the accurate determination of bran content in wheat flour. Bran, as one of the valuable by-products of the white flour milling process, is a rich source of dietary fiber and is widely used in the enrichment of flour-based products, particularly in the production of whole-grain and high–nutritional-value breads. Precise control of this parameter plays a decisive role in the final product quality as well as in compliance with production standards. In this study, an intelligent model based on hyperspectral data processing was developed to classify different ash levels in wheat flour. To this end, following the acquisition and preprocessing of hyperspectral images, the extracted spectral data were analyzed under two separate frameworks: with optimal feature selection and without feature selection. Classification in both cases was performed using an artificial neural network (ANN) algorithm. The results demonstrate the remarkable effectiveness of the proposed model, achieving a classification accuracy of 98.3% without feature selection and 100% with feature selection. These findings not only confirm the high operational accuracy of the model but also highlight the inherent capability of artificial neural networks to identify complex spectral patterns and their strong generalization performance under varying experimental conditions. Consequently, it can be inferred that the integration of hyperspectral imaging technology with machine learning provides a novel, accurate, and non-destructive approach for monitoring and optimizing ash content in wheat flour. The application of this approach directly on production lines can lead to improved flour quality and, ultimately, the production of superior products with higher nutritional standards.
کلیدواژهها English