.

Geo Informatics 2-2008
article minimum ground controls physical models because it enables users having little familiarity with the ikonos sensor to perform a geometric correction without gcps; only a digital elevation model (dem) is required. since biases or errors still exist after computing the rpcs, the results can be post-processed with a translation and several precise gcps; alternatively, the original rpc parameters can be refined with linear equations and precise gcps. several recent articles and papers addressing ikonos data showed good results for small areas by using rpcs together with a few gcps to apply a complementary first order polynomial adjustment to the data. more details about the rpc method can be found in the paper by grodecki and dial, block adjustment of high-resolution satellite images described by rational functions , photogrammetric engineering & remote sensing, january 2003. usually the initial rpcs provided with the image data do not have very high accuracy due to the limitation of the onboard system accuracy of the satellite. the accuracy improves after several refinements are applied by the satellite data vendors. this is true for ikonos images—the accuracy of their rpc data has significantly improved recently due to refinements in the geometric calibration of the sensor. for example, the horizontal positional accuracy is 15m ce90 for the geoortho kit product, and better than 2.0m ce90 for the precisionplus product. the latest version of pci geomatics orthoengine software was used for this testing. this software supports reading of the data, manual or automatic gcp/tie point (tp) collection, geometric modeling of different satellites using toutin s rigorous model or the rpc method, automatic dem generation and editing, orthorectification, and either manual or automatic mosaicking. orthoengine s rpc method is based on the block adjustment method developed by grodecki and dial and was certified by space imaging (www.pcigeomatics.com/ support_center/tech_papers/rpc_pci_cert.pdf). the method computes the polynomial adjustment math model for each image. figure 2a: multispectral image at 4m resolution figure 2b: panchromatic image at 1m resolution figure 2c: pan-sharpened image at 1m resolution ∆p = a0 + as ∙ sample + al ∙ line + asl ∙ sample ∙ line + ∆r = b0 + bs ∙ sample + bl ∙ line + bsl ∙ sample ∙ line + latest news? visit www.geoinformatics.com 53 march 2008
Powered by fluidbook tm