The large amount of data that can be acquired by spacecraft and rover instruments exceeds the bandwidth capacity of the transmission link to Earth. In order to maximize the rate of transmitted data given the capacity constraints of current downlink channels, we propose to make use of the onboard information processing capabilities for prioritizing data transmission and condensing the representation of the imaged surface through an efficient representation of its geometry and texture description. This will drastically reduce the size of the transmitted data and will allow a more accurate geological classification of the surfaces imaged by the rover. The main topic addressed here is the inference of an accurate geometric description of the geological surfaces being imaged by the rover and the definition of a prioritization scheme for optimal transmission to Earth. This prioritization scheme will be based on the geometric properties of the reconstructed surface. However, augmented description of the geological structure based on the image texture properties will be studied. This will represent a first step towards a 3D-texture classification of the objects. We will present a method to reconstruct aggregate geometric models of the surface imaged by the spacecraft.
Also I will briefly present Siggraph'00 papers related to mesh description and compression.