Cavity Shape from Parallel Linear Shading: A Low-cost Technique to Collect Data for an Image Mining Oriented Geoprocessing System

Abstract

An Image Mining oriented Geoprocessing depends essentially on spatially defined information from images. So it is possible to support decisionmaking quite suitably in several areas, such as environment management, urban management and health care. When large-scale use of image capturing and interpretation devices becomes possible, it seems attractive to have low cost additional infrastructures. This paper discusses the extraction of geometric features of cavities, primarily motivated by the need for monitoring patients with wounds called pressure ulcers (PU). An image with linear shadows on the cavity of a model is generated in order to enable the measurement of its deformation caused by depth. This yields maximum depth and volume in an experimental model that are compared with measurements made previously in a conventional manner. Differences with conventional measurements are partially satisfactory and suggest further improvements in image capturing device and computational procedures.