Reflectance Imaging Spectroscopy, often referred to as hyperspectral imaging, is an imaging technique that enables the simultaneous capture of spatial and spectral information from a scene without physical contact and in a non-invasive manner. These desirable features make it especially well-suited for applications in Cultural Heritage analysis, where the investigation of historical artifacts should avoid causing irreversible damage. This thesis is about the revisiting of the imaging pipeline from data acquisition to the processing steps that fuse two independent hyperspectral images captured in separate spectral ranges. The need to address this topic comes from the fact that Visible Near-Infrared (VNIR) and Short-Wave Infrared (SWIR) imaging spectroscopy are being consistently deployed in the field of Cultural Heritage to conduct a series of research tasks including but not limited to analyzing the basic components of historical artifacts (pigments, dyes, binding media, mordants, fiber, etc.), long-term artifact monitoring, assessment during conservation treatments, component mapping, and revealing of hidden patterns not discernible to the human eye. However, VNIR and SWIR hyperspectral images of the same scene are often analyzed independently because of the intrinsic differences present at the image sensor level, which makes data fusion a challenging problem. The first goal of this thesis is to develop an appropriate imaging setup for the simultaneous acquisition of VNIR-SWIR hyperspectral data with the twofold aim of obtaining high-quality data while preserving the integrity of the studied artifact. Secondly, the spatio-spectral alignment of the two hyperspectral images is addressed. Since the problem of spatial image registration has been extensively studied in the literature, we focus on the factors that may influence its performance in this context. For the spectral alignment, we propose a novel \textit{splicing} correction that smoothly connects hyperspectral images with adjacent or overlapping spectral ranges. We then explore the application of image sharpening (e.g. pansharpening) techniques originally developed for remote sensing on proximally-sensed historical artifacts, proposing a discussion focused on the negative impact that some algorithms have on subsequent analysis processes such as the classification of spectral signals. Finally, from the hypothesis of having to capture complex artifacts such as glossy paintings, we address the integration of polarimetric imaging in the fusion pipeline, developing an acquisition paradigm for the acquisition of VNIR-SWIR spectral Stokes images that allows the study of spectro-polarimetric quantities such as the correlation between the reflectance and the linear degree of polarization. In the initial hypothesis, the joint analysis of VNIR and SWIR Reflectance Imaging Spectroscopy data can be thought of as more powerful than the individual analyses conducted separately. However, this hypothesis could not be fully verified within this thesis, and some open questions are left for future explorations regarding its validity.