Automatic Woody crop and forest plantation inventory and statistic data extraction with GIS and drone (UAV) technology

Usage of small dimensions RPAS ( drones ) has a real and practical benefit on forest, environment and agriculture management. Drones can obtain Digital Surface Models and pseudo-NDVI orthomosaics that can be manipulated by Geographic Information Systems (hereinafter GIS) allowing professional and rigorous studies along with technically supported decisions.

We provide here an example of woody crop inventory done with our drones and GIS (QGIS) Software. It can be noted how powerful this information can be to properly manage and take decisions on precision agriculture and forest plantations.

In the following example droning used one of our drones (specifically the DE820) and a NDVI camera to survey the whole field. We did it in one flight at height of 100m above ground. Flight took 10 minutes.

The field was mainly populated by olive trees in the south of Spain.

Once on the ground, we were able to generate the Orthomosaic and both DTM (Digital Terrain Model) and DSM (Digital Surface Model). By using these three sources of information and the raster and vector calculator on the GIS, we could detect and classify every olive tree as an independent entity.

Having every olive tree entity located in the GIS database, a lot of valuable information can be extracted:

  • GPS Location
  • Height
  • Estimated volume
  • Perimeter
  • NDVI index
  • Surface area
  • etc

Woody crop was identified from surface and other vegetation by means of geographical calculations taking into account a series of parameters like shape, height, area and elevation gradient. At the same time NDVIb color disambiguation was applied to better refine the tree selection.

In this example we were able to differentiate every olive tree from the rest of the terrain and other vegetation present in the study zone.

 

GIS and drone technology is a game changer in Precision Agriculture and Environment management. This information provides an efficient, massive and rigorous way to manage field crops, forests, and any other geographic inventory.

Data obtained in this way can create massive lists/catalogs of fine geolocated entities allowing categorization, historical comparisons, etc.

The fact that every olive tree has been categorized, geolocated and found their characteristics is of a great importance now and even more in the future. Almost every new technology related to automatic crop harvest or automatic  fertilizer application will have to be supported by these data.

Moreover, historical data across the years can give precious information about growth patterns and diseases that can me data crossed against production rates per tree etc.

Full post here.

Views: 1740


Moderator
Comment by MarioSpeedwagon on August 31, 2015 at 11:06pm
Jesus,
Have you had much luck with other crops, such as apples or grapes? How important is consistent altitude?
Comment by Jesus A on September 1, 2015 at 6:25am

Hi Mario,

There should be not much difference with apples. Grapes may be a different story as the are grown in rows. But I guess there shouldn't either

If you use GCPs on the terrain, there is no problem as long as the flight level is more or less stable


Moderator
Comment by MarioSpeedwagon on September 1, 2015 at 2:02pm
The olives in your example image have a substantial amount of buffer room between them. I don't know of any crop around here that has that much space between them, but I would love to try.
Comment by Stanley Anak Suab on September 1, 2015 at 8:35pm

Good Job Jesus. Thanks for the sharing. By the way your article here is more into agriculture plantation. I think in your case the DTM from aerial photos is good where as you have areas with good bare earth coverage and no dense forest cover. I would call it "the ideal condition". My work predominantly deal with forest plantations and tropical natural forest I rather use accurate SARadar DTM like the 1m IFSAR for the terrain works. Im interested with the crown canopy shape derivation from aerial photo you did there :-D 

Comment by David on September 11, 2015 at 2:52pm

Excellent work there. I've been trying to understand how this could be used to track noxious weeds in similar situations. What sort of camera / lens / resolution are you needing to get high quality object recognition? Have you tried setups that were insufficient for the task?

I was thinking of using spectral analysis and noise patterns to differentiate regions of noxious grass from the native / pasture grasses and trees, but I have only limited experience with this? For my application, are you able to see if there would be a benefit to using binocular vision (better understanding of the shape of an object), or just take twice as many photos (more samples)?

Thanks

Comment by Mauricio García on March 15, 2016 at 10:54am

Hi. It looks realy good..

What camera did you use on that flight ?? I's an NDVI converted camera or is a branded one, MicaSense or similar. ??

thanks

Comment

You need to be a member of DIY Drones to add comments!

Join DIY Drones

© 2019   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service