Geospatial data acquisition at global to regional scales has wide acceptance, and tools such as Google Earth have been instrumental in extending Earth visualisation far beyond specialist users of GIS and satellite imagery. At the outcrop scale, the number of industry and academic geoscientists adopting digital technologies to gather field data is steadily increasing. When integrated with traditional field skills, these technologies offer two fundamental advantages: firstly, outcrop geology can now be recorded with very high detail and precision; secondly, observations and data are precisely georeferenced, which is a prerequisite for 2D and 3D spatial analysis. Digital outcrop data are being used in a wide variety of ways, many of which can be characterised in terms of two end members. Firstly, using methods such as terrestrial laser scanning and digital photogrammetry, it is possible to create highly realistic virtual copies of the outcrop. These virtual outcrop models can be used to great effect to enhance teaching, to provide virtual field-trips (most effective in conjunction with a real visit to the outcrop), to promote group discussion and interpretation, or as part of Health & Safety briefing. Secondly, digital outcrop data is also being used to derive quantitative attribute measurements from specific geological features. Here the emphasis is not on capturing a photo-realistic copy of the outcrop, but rather on gathering the relevant types of data at the most appropriate resolution and geospatial precision for the type of analysis undertaken. In addition to laser scanning, useful technologies include dGPS, laser range-finding, and Total Station surveying. Examples of this kind of quantitative analysis include fault curvature, roughness, branch-line geometry, spatial variation in fault displacement, fracture spacing and 3D spatial clustering, fold curvature, sedimentary channel morphology, lateral and vertical facies variations, and geomorphological analysis of terrace offsets. These kinds of studies are giving new insights into geological processes, and provide real-world constraints for validation and calibration in geological modelling. Because of the versatility of terrestrial laser scanning, a raw lidar dataset can often be used both to generate realistic virtual outcrop models, and to carry out detailed quantitative analysis. The main difference lies in the way the raw data are processed. This is reflected in the challenges that we are now facing. To make it easier to generate and use photo-realistic virtual outcrop models we need more efficient methods to render very large datasets (hundreds of millions of points or polygons). This requires better ways to smooth and filter point data, more dynamic meshing of points to recreate the outcrop surface, improved texture mapping of higher quality photos onto the mesh, and more seamless "Level of Detail" functionality so that progressively more detail is shown as the user zooms in on the outcrop. To improve quantitative analysis of geological features from lidar data, we need better colour imagery, and better co-registration of images with the raw point cloud. This can help to improve methods for automatic picking of features and removal of vegetation. More relevant methods for 3D spatial analysis of geological data need to be developed and tested. Further work is needed to compare different types of laser scanning equipment and improve the quantification of errors.