发表机构
Universidade Estadual de Campinas (UNICAMP)（坎皮纳斯州立大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对农业食品研究中合成多光谱/高光谱数据实验，提出基于文献的分类体系及遍历程序，连接证据、约束与评估设计，以表征实际测试的泛化形式。
AI 中文摘要
合成多光谱和高光谱数据越来越多地被用于解决光谱成像研究中测量数据、标签和采集资源有限的问题。然而，这些实验所产生的证据范围不仅取决于合成方法，还取决于传感条件、样本变异性、数据来源、评估设计以及下游应用。我们提出了一种基于文献的分类体系，用于通过光谱重建和/或数据增强进行合成MSI/HSI数据实验，并特别关注农业食品研究。该分类体系与一种遍历程序相配套，该程序将可用证据、传感约束、实验选择和评估设计与实际测试的泛化形式联系起来，生成紧凑的实验表征记录。虽然该框架主要面向已有研究问题的研究，但当最初仅定义了研究对象、数据集或传感设置时，同样的结构也可支持早期阶段的规划。该框架是通过问题驱动的阅读、实际研究经验以及针对其他近期研究的迭代完善而开发的，并作为进一步使用和完善的基础。
英文摘要
Synthetic multispectral and hyperspectral data are increasingly used to address limited measurements, labels, and acquisition resources in spectral-imaging research. Yet the scope of the evidence produced by these experiments depends not only on the synthesis method, but also on sensing conditions, sample variability, data provenance, evaluation design, and downstream use. We present a literature-informed taxonomy for experiments using synthetic MSI/HSI data through spectral reconstruction and/or data augmentation, with a deliberate focus on agri-food research. The taxonomy is paired with a traversal procedure that connects available evidence, sensing constraints, experimental choices, and evaluation design to the forms of generalization that were actually tested, producing a compact experiment characterization record. Although primarily intended for studies with an existing research question, the same structure can also support earlier-stage planning when only an object of study, dataset, or sensing setup is initially defined. The framework was developed through question-driven reading, practical research experience, and iterative refinement against additional recent studies, and is offered as a basis for further use and refinement.
Comments21 pages
