Satellite surveying accuracy discussion
PhotoSat surveying accuracy:
The PhotoSat surveying accuracy varies between 15cm and 30cm RMSE. The parameters affecting the accuracy and the measurement and verification of the accuracy are discussed below.
The PhotoSat elevation surveying accuracy is measured on areas of bare ground with slopes of less than 20% grade. The measurement points must be at least 10m from above ground features such as buildings and trees.
Proprietary PhotoSat models of the systematic distortions in stereo satellite photos
PhotoSat has developed proprietary models of the systematic distortions in the photos for most types of commercial stereo satellite systems. We have measured and characterized these systematic distortions in our test areas. The use of these models in our satellite processing is an essential component of the PhotoSat satellite surveying accuracy.
Accuracy test area in Eritrea
We have a test area in Eritrea where we have more than 50,000 ground survey points accurate to 2cm in elevation. We use this test area to develop the satellite photo distortion models used in our processing and to demonstrate the accuracy of our satellite surveying. For more information on the Eritrea test area click here.
Accuracy test areas in California
We have a test area in California where we use LiDAR grids accurate to 5cm in elevation. We use this test area to develop the satellite photo distortion models used in our processing and to demonstrate the accuracy of our satellite surveying. For more information on the California test area click here.
Repeated surveying of oil sands mines
Frequent surveying of the oil sands projects in Northern Alberta using stereo WorldView satellite photos has enabled us to develop processing systems to efficiently identify and attenuate distortions in the PhotoSat satellite elevation survey grids.
Adjusting systematic distortions in the stereo satellite photos
To achieve the best possible stereo satellite elevation surveying accuracy PhotoSat identifies and adjusts the distortions in the stereo photos. We apply our models of the systematic distortions of each type of satellite photo as follows:
- Identify and attenuate distortions in individual pairs of satellite photos
- use overlapping areas of stereo satellite photos to identify and attenuate distortions in each of the photo pairs
- use ground survey points to identify and attenuate specific types of distortions not attenuated in steps 1 and 2.
Verification of PhotoSat satellite surveying accuracy
We have verified our satellite surveying accuracy with numerous accuracy studies over our test areas in Asmara, Eritrea and the Garlock Fault in California. Our satellite surveying accuracy has also been verified by many accuracy studies conducted independently by PhotoSat customers. A few of these companies have generously allowed PhotoSat to publish their accuracy results.
PhotoSat surveying project accuracy reports
On every project with sufficient ground survey data or very accurate LiDAR survey data, PhotoSat produces an accuracy report. The report shows the relationship between the PhotoSat surveying and the other available survey data. In a significant number of projects these reports identify errors in the existing project survey data.
30cm relative surveying elevation accuracy, 3m global accuracy, with no ground survey control
Using WorldView and Pleiades stereo satellite photos PhotoSat can usually achieve 30cm elevation accuracy and 3m global accuracy with no ground survey control. The relative accuracy is defined as the accuracy between points on areas of bare ground sloping at less than 20% grade separated by up to 5km.
30cm surveying elevation accuracy, 3 ground survey points per 100km2
PhotoSat can usually achieve better than 30cm elevation surveying accuracy with a minimum of three evenly distributed photo recognizable ground survey points for areas up to 100km2. For larger areas one survey point is required for each additional 100km2. The elevation survey accuracy is measured on areas of bare ground sloping at less than 20% grade.
20cm surveying elevation accuracy, 10 ground survey points per 100km2
PhotoSat can usually achieve 20cm elevation surveying accuracy with a minimum of ten evenly distributed photo recognizable ground survey points for areas up to 100km2. For larger areas five survey points are required for each additional 100km2. The elevation survey accuracy is measured on areas of bare ground sloping at less than 20% grade.
15cm surveying elevation accuracy
Requires WorldView stereo satellite photos and extensive existing survey data
To reliably achieve 15cm elevation survey accuracy PhotoSat must identify and adjust small scale systematic distortions in the stereo WorldView satellite photos. To do this we must first identify areas where the ground has not changed between the date of the as built surveys and the dates of the satellite photos.
We measure elevation differences at these points. Adjustments are then made to the satellite elevation grid using proprietary models of characteristic distortions of the WorldView stereo satellite photos. This requires extensive existing survey data such as the following:
- Mine site “as built” surveys
- Oil & Gas 2D and 3D seismic source and receiver locations
- Good quality LiDAR surveys
Stereo satellite photos for 15cm surveying elevation accuracy
Satellite surveys accurate to 15cm in elevation are only produced from inline WorldView satellite stereo pairs. These stereo photos are taken about one minute apart by the following satellites:
PhotoSat produces satellite surveys accurate to better than 15cm in elevation for the Suncor Steepbank and Millennium oil sands mines. We deliver the satellite survey data for these mines one to two times per month.