HIGH RESOLUTION LANDCOVER MODELLING WITH PLÉIADES IMAGERY AND DEM DATA IN SUPPORT OF FINE SCALE LANDSCAPE THERMAL MODELLING

Abstract. In the evaluation of air-borne thermal infrared imaging sensors, the use of simulated spectral infrared scenery is a cost-effective way to provide input to the sensor. The benefit of simulated scenes includes control over parameters governing the spectral and related thermal behaviour of the terrain as well as atmospheric conditions. Such scenes need to have a high degree of radiometric and geometric accuracy, as well as high resolution to account for small objects having different spectral and associated thermal properties. In support of this, innovative use of tri-stereo, ultra-high resolution Pléiades satellite imagery is being used to generated high detail, small scale quantitative terrain surface data to compliment comparable optical data in order to produce detailed urban and rural landscape datasets representative of different landscape features, within which spectrally defined characteristics can be subsequently matched to thermal signatures. Pléiades tri-stereo mode, acquired from the same orbit during the same pass, is particularly favourable for reaching the required metric accuracy because images are radiometrically and geometrically very homogeneous, which allows a very good radiometric matching for relief computation. The tri-stereo approach reduces noise and allows significantly enhanced relief description in landscapes where simple stereo imaging cannot see features, such as in dense urban areas or valley bottoms in steep, mountainous areas. This paper describes the datasets that have been generated for DENEL over the Hartebeespoort Dam region, west of Pretoria, South Africa. The final terrain datasets are generated by integrated modelling of both height and spectral surface characteristics within an object-based modelling environment. This approach provides an operational framework for rapid and highly accurate mapping of building and vegetation structure of wide areas, as is required in support of the evaluation of thermal imaging sensors.