Intelligent pattern detection and compression. An application to very low bit rate transmission of ship wake aerial images

In land or sea surveillance systems, images taken from satellites or aerial platforms must be compressed and transmitted to a ground station for subsequent processing. If the employed platform is an airborne HAAP whose main purpose is to provide mobile/personal communication services over sea or impervious areas, major problems could arise for the transmission of images on radio-mobile compatible (e.g. GSM) channels, due to the very low available bit rate. A first stage detection algorithm could be recommended, to be run on board in order to transmit only those images which are likely to contain the patterns of interest. Moreover, this first stage detection could help performing an intelligent compression, so as to accurately preserve the objects of interest at the expenses of an impaired quality of the recovered background, which may be of negligible interest. In this paper, we address two joint detection-compression algorithms suitable for the recognition of ship wakes in sea images. The algorithms have proven to yield very high compression ratios while preserving the objects of interest. (C) 1999 Elsevier Science B.V. All rights reserved.