For my research purpose I need to construct urban terrain model. But In india, LiDAR data are not affordable enough as well as hard to find for a researcher with middle class background. There are some papers where researchers used that model. But they did not provide detail procedure of making that. One of them modified of DEM by using of ground clearance among buildings. The process itself is not clear enough to follow.(DOI: 10.1007/s11442-011-0844-7)
Generation of DEM for urban areas with reasonable accuracy using stereo pair images (either satellite or airborne) and Interferometric Synthetic Aperture Radar (InSAR) image.
given the tight legal restrictions in India, using drones to acquire the data you need will likely not work. Best bet are indeed satellite data. Here India is very well equipped: the Cartosat stereo satellites provide very good data for elevation extraction. While the first Cartosat generation still has 2.5 m spatial resolution, the latest one (Cartosat-2C, launched in 2016) bring that down to about 0.6 m. Cartosat 3 is expected to bring this down further to about 25 cm in the panchromatic.
good luck, Norman
If you have aerial images you can try a SfM approach for orient the image block and a dense image matching for extract the DSM. You can use open source photogrammetric software such as MicMac for this purpose.
Good luck, Gil
I agree with Volney Junior Borges de Bitencourt. You can use a drone in order to have the aerial photos you need and then a software like photoscan by agisoft in order to create what you want.
Kleomenis
Mr. Sandip,
I have faced similar problem and came with a logical and cost effective solution. Firstly I am mentioning that the process can't assure remarkable accuracy but can be cost effective in both subjective and objective matter. The process is mentioned step step by step below:
- Download free DEM/DSM data fro your target location (https://gisgeography.com/free-global-dem-data-sources/)
- Classify your urban area based on height of the building, and do a stratified random sampling survey to find out average building height for each individual urban class.
- Build raster pyramid (DEm/DSM) and develop raster attribute table showing the elevation value for each pixel
- Select and classify the pixels based on elevation values e.g. low, high, medium
- Subtract pixel values from the field values (stratified sampling survey)considering relevant classes.
- Interpolate new values to construct urban DTM model for your target city.
- The process can be reapplied for validation and more accurate outcome.
Hope that it will solve your problem to some extent.
- Badges
- Science topic
- Similar topics
- Geoscience
- Atmospheric Sciences
- Climate Science
- Climate Change