This thesis develops optical techniques for hyperspectral seafloor mapping, focusing on geometric and radiometric processing. It explores using a conventional camera and a hyperspectral line scanner together, leveraging advancements in digital camera technology to enhance accuracy. The technique aligns hyperspectral scan lines with camera images, using a 3D model to simulate light beam paths and define the seafloor origin of scan lines. Initially demonstrated on a deep-water coral reef, the method was adapted for shallow water mapping from an unmanned aerial vehicle, accounting for navigation and refraction effects. The study also proposed a semi-analytical model to describe the relationship between seabed reflectance and measured spectra, using geometric parameters and optical properties of the water column. This model was validated with deep-water field data, highlighting the potential for accurate multi-scale mapping of the seafloor.