Image content descriptors-the detection stage

This paper addresses the difficulties involved in the automatic generation of image content descriptors. These descriptors can be used for the indexing, matching and searching of data in an image database. The current stage of the work has concentrated on the development/collation of a set of feature detectors covering points (corners), lines (edges, ridges) and regions which will operate on a variety of image sensors (eg SPOT, IRLS - Infra-Red LineScan, dd5 - Russian satellite data). One of the difficulties that has existed in computer vision is that researchers have had a tendency to concentrate on developing single algorithms (eg edge and region detectors) for specific types images. It is only recently that papers are now appearing which consider the questions of combining detectors to improve their performance and reliability on a wider variety of images. All subsequent stages such as searching or matching of descriptions depend on a fundamentally reliable detection stage. For this reason the initial objective of our work was not to depend purely upon just one algorithm but to have several algorithms which may then be combined into a composite estimate of the feature(s) that are required