Evaluating multi-resolution remote sensing data - A knowledge-driven approach to geological remote sensing in a high-sulphidation epithermal system

This PhD research develops and tests a resolution-first, knowledge-based framework for characterising a high-sulphidation epithermal system using geological remote sensing. The study focuses on a single lithocap environment and uses laboratory, airborne and spaceborne datasets spanning the VNIR, SWIR, and LWIR wavelength ranges. The aim is to determine (i) what information each spatial and spectral resolution can independently provide about a mineral system and (ii) the possible multiplier effect obtained when combining resolutions. The work targets four classes of geological products that can be derived from a remote or proximal sensing source: hydrothermal alteration mineralogy, rock-forming mineralogy, fluid composition, and fluid emplacement.Methodologically, the thesis applies a consistent, physics-based image analysis workflow across all resolutions. Laboratory hyperspectral measurements in the SWIR and LWIR ranges are first used to define and validate mineral spectra, textural relationships, wavelength shifts and kaolin-group subtypes. These spectra provide the basis for interpreting imaging datasets and serve as the reference for validation. Airborne hyperspectral SWIR and LWIR data are then processed to deposit scale using imaging processing techniques, such as band ratios, spectral indices, and absorption feature/emissivity minima mapping. The same indices and ratios are applied to spaceborne hyperspectral SWIR and multispectral VNIR-SWIR-TIR data to generate directly comparable products, ranging from regional to deposit scales. All interpretations are made using a knowledge-based approach, anchored in mineral spectroscopy and epithermal system models, rather than data-driven classification alone.The study uses only geological and remotely sensed data; no personal data are involved. Laboratory measurements were conducted on rock and mineral samples under standard procedures. Airborne and spaceborne datasets were obtained from public and commercial providers under their licence conditions and applicable national regulations for airborne surveys. As such, no specific ethical review was required, and all data collection adhered to legal and regulatory requirements. The combination of clearly documented processing steps, physics-based indices, and cross-scale validation is designed to make the workflow transparent and reproducible, enabling its application to other mineral systems and survey designs.

本博士研究开发并测试了一套以分辨率为核心、基于知识的框架，用于借助地质遥感技术刻画高硫化浅成热液系统（high-sulphidation epithermal system）。
本研究聚焦单一岩帽（lithocap）环境，采用覆盖可见光近红外（Visible and Near-Infrared, VNIR）、短波红外（Short-Wave Infrared, SWIR）及长波红外（Long-Wave Infrared, LWIR）波段范围的实验室、航空及星载数据集。
本研究旨在达成两大目标：其一，明确不同空间分辨率与光谱分辨率可独立提供的矿物系统相关信息；其二，探究融合多种分辨率后可获得的潜在倍增效应。本研究针对四类可通过遥感或近距传感获取的地质产物展开：热液蚀变矿物学、造岩矿物学、流体成分及流体侵位。
在方法论层面，本研究针对所有分辨率类型采用统一的基于物理的图像分析工作流。首先，通过短波红外与长波红外波段的实验室高光谱测量，定义并验证矿物光谱、结构关系、波长偏移及高岭石族（kaolin-group）亚型。这些光谱数据为成像数据集解译提供基础，并作为验证参考。随后，借助波段比值法、光谱指数及吸收特征/发射率最小值制图等图像处理技术，对航空高光谱短波红外与长波红外数据进行矿床尺度的处理。将相同的指数与比值方法应用于星载高光谱短波红外数据及多光谱可见光近红外-短波红外-热红外（VNIR-SWIR-TIR）数据，以生成从区域尺度到矿床尺度的可直接对比的产物。所有解译均采用基于知识的方法，锚定矿物光谱学与浅成热液系统模型，而非仅依赖数据驱动的分类。
本研究仅使用地质与遥感数据，未涉及任何个人数据。实验室测量严格遵循标准流程对岩石与矿物样品开展。航空与星载数据集均从公共及商业提供商处获取，并遵循其许可条款与国家相关法规。因此，本研究无需开展专门的伦理审查，所有数据收集均符合法律与监管要求。
本研究通过清晰记录的处理步骤、基于物理的指数方法与跨尺度验证，旨在实现工作流的透明化与可复现性，使其可推广应用于其他矿物系统与调查设计方案中。