As an example, I'm looking at the 60 arc-second (1 degree) data from the SRTM mission, where each 1-degree tile of data contains 1201 x 1201 data points for elevation. I'm assuming that this was the RAW sensor data resolution.
The data is published in geographic coordinates (equiangular coordinates, lat/lon) meaning each point of data ends up representing a trapezoidal chunk of land. With geographic coordinates, as you get closer to the poles, this chunk of land shrinks in surface area.
If I've understood this correctly, the result is that at some point as you proceed towards a pole, the RAW sensor data will become coarser than the published data and therefore a single RAW data sample would end up covering two published data points. This results in multiple published data points containing the same data.
I'm assuming this means the data was as some point over-sampled.
My questions are the following:
أكثر...
The data is published in geographic coordinates (equiangular coordinates, lat/lon) meaning each point of data ends up representing a trapezoidal chunk of land. With geographic coordinates, as you get closer to the poles, this chunk of land shrinks in surface area.
If I've understood this correctly, the result is that at some point as you proceed towards a pole, the RAW sensor data will become coarser than the published data and therefore a single RAW data sample would end up covering two published data points. This results in multiple published data points containing the same data.
I'm assuming this means the data was as some point over-sampled.
My questions are the following:
- In the SRTM data case, what method was used to over-sample the data?
- Is oversampling a common practice with these types of data sets? For instance, the ASTER project?
- Is there a better way to publish geographic data without oversampling?
أكثر...