Comparative properties of four airborne sensors and their applicability to land surface interpretation

Abstract

Digital aerial photography is a widely used tool for mapping and monitoring biotopes and land use types. As technology develops new and improved sensors (e.g. higher radiometric resolution and dynamic range) become increasingly available. Comprehensive knowledge of the radiometric properties of digital imagery sensors is of special importance to adequately evaluate its applicability to land surface interpretation and mapping. This knowledge permits the processing of information in a more rigorous and controlled manner. The objective of this work was to evaluate and compare the performance of four aerial photography sensors (ADS40-1 and - 2, UltracamD and DMC) in terms of their applicability to biotope type mapping in North Germany. Different biotopes and terrain features were systematically sampled and analyzed in terms of their spectral and radiometric characteristics. Qualitative and quantitative approaches were used to compare the sensor performance on different terrain features. A preliminary assessment of the camera's relative adequacy to more automated interpretation approaches, using object oriented analyses was performed. Results showed that most of the differences seem to be attributable to channel width and radiometric DN range. Some of the challenges of using object based approaches are discussed. From this study it is concluded that there is a series of factors that seem crucial to be taken into account when trying to assess interpretability in the context of LU/LC surveys. Some of them, such as edge sharpness and contrast, are already accounted for in image interpretability measures. Others, mostly related to spectral properties, such as multispectral contrast between objects and background or between neighbouring objects, need to be considered and integrated into a quantitative and more rigorous indicator. © 2008 International Society for Photogrammetry and Remote Sensing. All rights reserved.