Using image statistics for automated quality assessment of urban geospatial data

In this paper, a framework is proposed to assess the quality of urban geospatial data using image information. An important aspect is the emphasis on accuracy and reliability of the system. The quality of the image information is quantified by characterizing the ridge detection performance in terms of detection rate. Based on statistics of a typical road and its immediate surroundings, a prediction of the detection performance is made and the parameter set for the optimal performance of the ridge detection is derived. The prediction is used to define a displacement quality measure. Experiments were conducted on IKONOS panchromatic and Quickbird multispectral satellite images.