M. Govorov1, G. Gienko2, B. Maguire1

1 - Malaspina University-College, Nanaimo, Canada

2 - University of the South Pacific, School of Geography, Fiji


The paper introduces an idea of the open-source GeoTruth engine a Web-based, worldwide community-open participatory mechanism for collection, storing and distribution of geo-referenced ground truth terrestrial photographs and corresponding metadata to support interpretation of aerial and satellite imagery.

The concept of GeoTruth system is based on the assumption that natural observations of the environmental landscape could be substituted by terrestrial photographs. Geo-referenced, time-stamped and annotated terrestrial landscape photographs could be uploaded to a community-open geo-media engine and then be made accessible to the wider community for identification of ground features during the interpretation of remotely sensed imagery.

The proposed concept goes beyond an ordinary data warehouse as apart from storing terrestrial photographs and their metadata which could be used in the client-based approach (i.e. when the user can download all data from the server and process it on the local computer). The GeoTruth system could provide to the users some additional services such as server-based supervised multi-spectral image classification (using GeoTruths database of terrestrial photographs) and generation corresponding land cover maps using openly accessible remotely sensed images.

The paper explores the technological and implementation issues of the GeoTruth system, including graphical user interface mechanisms for geo-compatible exchange, analysis and visualization of annotated ground truth photographs and their metadata. Visualization tools are clearly a key component because they provide a common language to which all community members can relate. Similar, design GUI and implementing of an initial set of tools for GeoTruth System are very important task to attract members and support an integrated knowledge environment.

The system should be designed to accommodate users with different professional background and to provide corresponding functional components, accessible through several levels of User Interface (UI): "Photo", Contribution and Remote Sensing". The Photo UI is concerned with the search, visualization and retrieval of terrestrial photographs. Contribution UI will work for georeferencing the terrestrial photographs, metadata input and processing, the calculated fields of view etc. The terrestrial photographs would be drawn on top of the satellite image, showing the camera field of view. The Remote Sensing UI would incorporate all of the above features, but include further server and client tools for image analysis.