AI-Assisted Hyperspectral Analysis for Accurate and Fast Sucrose Estimation in Sugar Beets
sensAIfood Final International Conference – Application of AI Methods for Processing Spectral Data in Feed and Food Chain, Tirane, Arnavutluk, 8 - 10 Ekim 2025, ss.48, (Özet Bildiri)
- Yayın Türü: Bildiri / Özet Bildiri
- Basıldığı Şehir: Tirane
- Basıldığı Ülke: Arnavutluk
- Sayfa Sayıları: ss.48
- Orta Doğu Teknik Üniversitesi Adresli: Evet
Özet
As factory capacities and consumer awareness of product quality increase, there is a growing demand for rapid and non-destructive quality control methods in the sugar industry. Conventional techniques such as High-Performance Liquid Chromatography (HPLC), although accurate, are costly and time-consuming. In this study, Hyperspectral Imaging (HSI) was combined with Artificial Intelligence (AI) and chemometric approaches to estimate sucrose content in sugar beets. Spectral data were collected using a line-scan HSI system operating in the 400–1000 nm range. To reduce noise and enhance signal quality, preprocessing techniques including Gaussian Smoothing, Savitzky-Golay filtering, and Standard Normal Variate (SNV) were applied. Sugar beets were segmented into top, middle, and bottom parts to investigate spatial variability in sucrose distribution. The middle section yielded the most accurate predictions, likely due to its more uniform spectral profile. AI models such as Support Vector Machines (SVM) and k-Nearest Neighbors (kNN) were applied, while classical methods like Partial Least Squares Regression (PLSR) and Linear Regression served as benchmarks. Among these, SVM achieved the best performance, with an R² value of 0.934 and RMSE of 1.158. HPLC was used as the reference method for model validation. These results demonstrate that combining HSI with AI provides a fast, non-invasive, and reliable approach for sucrose prediction, offering a valuable tool for real-time quality assurance in sugar beet processing.