Automatic checking (and/or) refinement of DTMs without reference data (2005)
The need for Digital Elevation Model (DEM) checking becomes apparent as more and more data need updating and validation. Often automatic DEM extraction techniques prove to be inadequate and extensive manual correction is necessary. Since the use of reference data is expensive, therefore cannot be used in all cases, a novel method for DTM checking through orthophotographs created from the two photographs of the stereopair has been introduced. The mathematical model is loosely based on Norvelle’s Iterative Orthophoto Refinement method (1994). The underlying concept is based on the fact that two orthoimages created from different aerial images should be identical. Any differences between these orthoimages can be attributed to DTM errors, as DTM is considered to be the main source of error. The proposed methodology uses image matching to collect same points on orthophotos and translates their discrepancies into DTM deviations from the truth (real surface). These deviations from the truth and can be used for quality assessment or update/correction. The mathematical model is rigorous and straightforward, it does not approximate, and therefore there is no need for iterations. Results over artificial data, test data and final report over EuroSDR's project "Automated Quality Control of DTMs", proved that the method works very well, even when testing DTMs from LiDAR data.
This is a by product of the PhD thesis (in Greek, available upon request).
Tests are included in EuroSDR publication #51, “Checking and Improving of DTMs”.
For more info please check references.
This is a by product of the PhD thesis (in Greek, available upon request).
Tests are included in EuroSDR publication #51, “Checking and Improving of DTMs”.
For more info please check references.
