John A. Howell

Terrestrial lidar and hyperspectral data fusion products for geological outcrop analysis

Close-range hyperspectral imaging is an emerging technique for remotely mapping mineral content and distributions in inaccessible geological outcrop surfaces, allowing subtle chemical variations to be identified with high resolution and accuracy. Terrestrial laser scanning (lidar) is an established method for rapidly obtaining three-dimensional geometry, with unparalleled point density and precision. The combination of these highly complementary data types - 3D topography and surface properties - enables the production of value-added photorealistic outcrop models, adding new information that can be used for solving geological problems. This paper assesses the benefits of merging lidar and hyperspectral imaging, and presents qualitative and quantitative means of analysing the fused datasets. The integration requires an accurate co-registration, so that the 2D hyperspectral classification products can be given real measurement units. This stage is reliant on using a model that correctly describes the imaging geometry of the hyperspectral instrument, allowing image pixels and 3D points in the lidar model to be related. Increased quantitative analysis is then possible, as areas and spatial relationships can be examined by projecting classified material boundaries into 3D space. The combined data can be interpreted in a very visual manner, by colouring and texturing the lidar geometry with hyperspectral mineral maps. Because hyperspectral processing often results in several image products and classifications, these can be difficult to analyse simultaneously. A novel visualisation method is presented, where photorealistic lidar models are superimposed with multiple texture-mapped layers, allowing blending between conventional and hyperspectral imaging products to assist with interpretation and validation. The advantages and potential of the data fusion are illustrated with example outcrop data.

Close-range hyperspectral imaging for geological field studies: workflow and methods

Close-range hyperspectral imaging is a new method for geological research, in which imaging spectrometry is applied from the ground, allowing the mineralogy and lithology in near-vertical cliff sections to be studied in detail. Contemporary outcrop studies often make use of photorealistic three-dimensional (3D) models, derived from terrestrial laser scanning (lidar), that facilitate geological interpretation of geometric features. Hyperspectral imaging provides complementary geochemical information that can be combined with lidar models, enhancing quantitative and qualitative analyses. This article describes a complete workflow for applying close-range hyperspectral imaging, from planning the optimal scan conditions and data acquisition, through pre-processing the hyperspectral imagery and spectral mapping, integration with lidar photorealistic 3D models, and analysis of the geological results. Pre-processing of the hyperspectral images involves the reduction of scanner artefacts and image discontinuities, as well as relative reflectance calibration using empirical line correction, based on two calibrated reflection targets. Signal-to-noise ratios better than 70:1 are achieved for materials with 50% reflectance. The lidar-based models are textured with products such as hyperspectral classification maps. Examples from carbonate and siliciclastic geological environments are presented, with results showing that spectrally similar material, such as different dolomite types or sandstone and siltstone, can be distinguished and spectrally mapped. This workflow offers a novel and flexible technique for applications, in which a close-range instrument setup is required and the spatial distribution of minerals or chemical variations is valuable.

Computer-aided image geometry analysis and subset selection for optimizing texture quality in photorealistic models

Photorealistic 3D models are used for visualization, interpretation and spatial measurement in many disciplines, such as cultural heritage, archaeology and geoscience. Using modern image- and laser-based 3D modelling techniques, it is normal to acquire more data than is finally used for 3D model texturing, as images may be acquired from multiple positions, with large overlap, or with different cameras and lenses. Such redundant image sets require sorting to restrict the number of images, increasing the processing efficiency and realism of models. However, selection of image subsets optimized for texturing purposes is an example of complex spatial analysis. Manual selection may be challenging and time-consuming, especially for models of rugose topography, where the user must account for occlusions and ensure coverage of all relevant model triangles. To address this, this paper presents a framework for computer-aided image geometry analysis and subset selection for optimizing texture quality in photorealistic models. The framework was created to offer algorithms for candidate image subset selection, whilst supporting refinement of subsets in an intuitive and visual manner. Automatic image sorting was implemented using algorithms originating in computer science and information theory, and variants of these were compared using multiple 3D models and covering image sets, collected for geological applications. The image subsets provided by the automatic procedures were compared to manually selected sets and their suitability for 3D model texturing was assessed. Results indicate that the automatic sorting algorithms are a promising alternative to manual methods. An algorithm based on a greedy solution to the weighted set-cover problem provided image sets closest to the quality and size of the manually selected sets. The improved automation and more reliable quality indicators make the photorealistic model creation workflow more accessible for application experts, increasing the users confidence in the final textured model completeness.