The geometric accuracy of high resolution satellite imagery, such as QuickBird or Ikonos, was evaluated in several previous works also carried out by the University of Bologna. The common finding of these researches is that such orthoimages, under specific conditions, could be used as medium - large scale raster maps. Therefore, to this purpose, they must adequately processed, and their accuracy is depending on many factor, in particular on the geometric correction model adopted for the orthorectification, on the number and distribution of Ground Control Points (GCPs) over the scene and on the accuracy of Digital Terrain Model. The evaluation of for such orthoimages was always carried out by means of a certain number of sparse Check Points (CPs), evenly distributed over the whole scene, whose coordinates were obtained both from maps and from GPS ground surveys. In the present work we were interested in evaluating the possibility that these orthoimages could be used not only as products in themselves, but also as an information source both for the creation of new 1:5000 or 1:10000 numerical maps, and for map updating. For this reason, geometric accuracy, even if evaluated over a large number of CPs, is not sufficient to guarantee a correct detection and interpretation of objects in the image and, therefore, it is not sufficient to assess the accuracy over a limited number of sparse points, but it is advisable to vectorize objects from orthoimages and to evaluate their accuracy with ground surveys. The image quality and its radiometric characteristics are critical aspects for the vectorisation, because they could support or impede the correct photointerpretation of objects in the scene. For this reason, our main purpose was the evaluation of the effective use and potential of orthoimages for creation or updating of large scale structured GeoDataBase, and, particularly, the work was addressed to building detection. Regione Emilia Romagna is interested in this application in order to create a GeoDataBase starting from the updating of analogical maps, using high resolution satellite images. This sperimentation was possible thanks to materials (maps and orthoimages) supplied by Regione Emilia Romagna
Barbarella M., Zanni M (2006). Metric and qualitative evaluation of quickbird orthoimages for a large scale geodatabase creation. REPORTS ON GEODESY, 2/77, 255-262.
Metric and qualitative evaluation of quickbird orthoimages for a large scale geodatabase creation
BARBARELLA, MAURIZIO;ZANNI, MICHELA
2006
Abstract
The geometric accuracy of high resolution satellite imagery, such as QuickBird or Ikonos, was evaluated in several previous works also carried out by the University of Bologna. The common finding of these researches is that such orthoimages, under specific conditions, could be used as medium - large scale raster maps. Therefore, to this purpose, they must adequately processed, and their accuracy is depending on many factor, in particular on the geometric correction model adopted for the orthorectification, on the number and distribution of Ground Control Points (GCPs) over the scene and on the accuracy of Digital Terrain Model. The evaluation of for such orthoimages was always carried out by means of a certain number of sparse Check Points (CPs), evenly distributed over the whole scene, whose coordinates were obtained both from maps and from GPS ground surveys. In the present work we were interested in evaluating the possibility that these orthoimages could be used not only as products in themselves, but also as an information source both for the creation of new 1:5000 or 1:10000 numerical maps, and for map updating. For this reason, geometric accuracy, even if evaluated over a large number of CPs, is not sufficient to guarantee a correct detection and interpretation of objects in the image and, therefore, it is not sufficient to assess the accuracy over a limited number of sparse points, but it is advisable to vectorize objects from orthoimages and to evaluate their accuracy with ground surveys. The image quality and its radiometric characteristics are critical aspects for the vectorisation, because they could support or impede the correct photointerpretation of objects in the scene. For this reason, our main purpose was the evaluation of the effective use and potential of orthoimages for creation or updating of large scale structured GeoDataBase, and, particularly, the work was addressed to building detection. Regione Emilia Romagna is interested in this application in order to create a GeoDataBase starting from the updating of analogical maps, using high resolution satellite images. This sperimentation was possible thanks to materials (maps and orthoimages) supplied by Regione Emilia RomagnaI documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
