.

GeoInformatics-7-2009
article the rms gcp residuals were about 2.0m in x and 1.9m in y with a maximum residual of 5.4m in x and 4.4m in y. when all the gcps were changed to icps, the rms icp errors were about 3.7m in x and 4.6 m in y with a maximum error of 6.6m in x and 9.5m in y. mated system. multiple computers can also be used to speed up the processes. the fully automated process means that it is easy to generate rapideye orthos/mosaics for many critical applications for quick turnaround. modern computer hardware such as multi-core in the case of scene processors and graphiacquired over koprivcal processing units nice region 30 gcps (gpu) have been found were collected. the to improve the speed figure 8: automatic mosaicked rapideye image of phoenix rms gcp residuals of computation-bound were about 2.6m in x and 2.3m in y with a maximum residual of 5.8m processes. the model for gpu computing is to use a cpu and gpu in x and 5.4m in y. when all the gcps were changed to icps, the rms together in a heterogeneous computing model. the sequential part of icp errors were about 5.1m in x and 3.9m in y with a maximum error the application runs on the cpu and the computationally-intensive part of 10.5m in x and 8.6m in y. runs on the gpu. from the user s perspective, the application simply runs faster because it is using the high-performance of the gpu to boost hence, both data have rms error within the resolution of the sensor performance. pci geoimaging accelerator (gxl) has taken advantage when no gcps were used. figure 6 and 7 show the orthorectified images of this modern computer hardware by integrating the use of gpu comusing no gcps overlaid with google earth. puting to perform intense computation tasks such as pansharpening, orthorectification and automatic mosaicking. it provides speed imautomatic mosaicking provements of approximately 6 times for pansharpening, 10 times for the successful generation of high accuracy rapideye orthos means that orthorectification and 5 times for automatic mosaicking. these improveit is possible to create seamless mosaics of rapideye data without ments in processing speed will help the user to obtain results much gcps. however, mosaicking and color balancing are usually extremely faster without any change in accuracy. time consuming processes. the pci automatic cutline searching, conclusions mosaicking and color balance tools could be used to perform the entire it is possible to generate high accuracy orthos and mosaics of rapideye process automatically. no human intervention would be required durdata without ground control points for quick turnaround. the test results ing the process. to test the automatic mosaicking of rapideye data, show rms errors consistently around one resolution of the data. the the block of three phoenix data were used. the mosaic file has a size fact the gcps are not required for rapideye geometric correction of approximately 5.6 gigabytes. figure 8 shows the overview of the translates to very significant cost and time savings for the user. in mosaic image and figure 9 shows a full resolution of the mosaic image addition, automated batch processing to generate a large quantity of overlaid with the cutline in red color. it can be seen from figure 9 that rapideye orthos/mosaics is now possible using single or multiple the roads are aligned to each other perfectly at the cutline between computers. the use of gpu computing can improve the speed of the two images. orthorectification to 10 times faster and automatic mosaicking up to automated batch processing using gpu 5 times faster. since high accuracy rapideye orthos and mosaics can be generated dr. philip cheng cheng@pcigeomatics.com automatically without gcps, it is possible to integrate all the processes is a senior scientist at pci geomatics. in a fully automated batch system. mr. jiri sustera jiri.sustera@gisat.gz is the batch programs required to perremote sensing expert at gisat. form all the steps are available inside pci software. it can be run the authors would like to acknowlthrough python or pci easi scripts. edge rapideye company and jrc european commission for the advantages of automated proproviding the test data. cessing are that it will: (1) maximize production, (2) automate repetitive time-consuming tasks to produce consistent results, (3) improve operating efficiencies, (4) reduce labor costs, and (5) shorten throughput time for the delivery cycle. the generation of a large quantity of high accuracy orthos or mosaics, such as a mosaic of an entire country, can figure 9: full resolution automatic mosaicked rapideye image of phoenix overbe generated easily with the autolaid with cutlines 40 october/november 2009
Powered by fluidbook tm