用高光谱成像与机器学习自动检测蜂蜜掺糖，准确率达96.39%。
Honey Adulteration Detection using Hyperspectral Imaging and Machine Learning
- 先用LDA提取特征，再用KNN分类花源与掺假程度。
- 跨验证准确率96.39%，可替代传统化学检测方法。
- 适合食品质检、农产品安全领域研究人员使用。
本文旨在基于蜂蜜高光谱成像数据，构建一种基于机器学习的自动蜂蜜掺糖检测系统。首先通过植物来源识别子系统对蜂蜜花源进行分类；随后，利用掺假检测子系统识别糖浆掺假并量化浓度。两个子系统均包含两步：第一步使用线性判别分析（LDA）提取相关特征；第二步采用K近邻（KNN）模型分别完成花源分类和掺假水平判定。在公开的蜂蜜高光谱图像数据集上评估性能，结果表明该系统整体交叉验证准确率达96.39%，可作为当前化学检测方法的可行替代方案。
原文摘要 · Abstract (English)
This paper aims to develop a machine learning-based system for automatically detecting honey adulteration with sugar syrup, based on honey hyperspectral imaging data. First, the floral source of a honey sample is classified by a botanical origin identification subsystem. Then, the sugar syrup adulteration is identified, and its concentration is quantified by an adulteration detection subsystem. Both subsystems consist of two steps. The first step involves extracting relevant features from the honey sample using Linear Discriminant Analysis (LDA). In the second step, we utilize the K-Nearest Neighbors (KNN) model to classify the honey botanical origin in the first subsystem and identify the adulteration level in the second subsystem. We assess the proposed system performance on a public honey hyperspectral image dataset. The result indicates that the proposed system can detect adulteration in honey with an overall cross-validation accuracy of 96.39%, making it an appropriate alternative to the current chemical-based detection methods.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。