TEMPORAL CORRELATION OF METADATA ERRORS FOR COMMERCIAL SATELLITE IMAGES: REPRESENTATION AND EFFECTS ON STEREO EXTRACTION ACCURACY

Abstract

Abstract. Extraction of ground points using a basic stereo pair of commercial satellite electro-optical images typically yields vertical errors much smaller than expected. In particular, the magnitude of vertical errors relative to horizontal errors is significantly smaller than expected based on imaging geometry (convergence angle, etc.) alone. This paper suggests that temporal correlation or similarity of metadata (sensor position, attitude) errors between two same-pass images is the major cause of this phenomenon. It discusses the sources of temporal correlation, how it can be represented, and how an optimal ground point extraction algorithm detailed in the paper uses this representation in order to provide the best possible 3D location and corresponding 3 x 3 error covariance for reliable predicted solution accuracy. This paper also provides an estimate for temporal correlation, approximately 0.70 (70%), and explains how this value was derived based on the ratio of measured 0.9 p vertical errors to measured 0.9 p horizontal errors compiled over many stereo pairs and ground truth points as described in various papers in the literature. As demonstrated in this paper, based on simulation and error propagation for typical stereo geometry, if this correlation is not accounted for, predicted 0.9 p vertical error is approximately 60% too large. Knowledge of temporal correlation is essential for reliable stereo accuracy prediction as well as proper modeling of a priori metadata uncertainty in the support of metadata adjustment in a value-added process, such as registration to sparse control or a block adjustment.