Plant asessment and counting in PIX4Dfields! 🌳

Great performance of Agriculture over a citrus grove planted in a square layout!

Objective :dart: : Tree assessment and counting using the Magic Tool.

Conclusion:
:chart_with_downwards_trend: Low-performing trees identified.
:evergreen_tree::evergreen_tree: 600 trees counted, 1% error.

1- Orthomosaic from high-resolution RGB drone imagery (4cm/pixel)

2 - Magic Tool in action, grid showing automatic classification, after some refinement by hand, selection done.

3 - Select Conver to “Area annotation”. Magic Tool provides an area annotation for each tree detected.
image

Afterward, separate it into single polygons. Click Edit, and Create single polygons.

4 - After exporting TGI index statistics per tree, low index trees can be identified.

Try it yourself! Download Pix4D software
Learn how to use the Magic tool: https://www.youtube.com/watch?v=2NQW4YAKRVY&t=4s

2 Likes

How do I export to get the TGI Index statistics per tree.
I check all tutorials I cannot find one to explain this.

je n’arrive pas à obtenir un polygone unique par plante mais des pixels par plantes qui ne me facilite pas le comptage, comment faire?

  1. You need to convert the MagicTool annotation into single polygons:

  2. Copy them to the TGI index layer if they are not already present.

  3. Export → Statistics

I I want you data ,
I want to study
how can I get it :joy:

Regarding single polygons, how many can the software handle? I’ve tried within the range of 900-2000 and it’s unsurprisingly causes a noticable drop in performance. Are there any tips on managing it? The polygons are imported as a .shp file from other software.

This dataset is part of our training course PIX4Dfields certification

The low performance is expected. There is no fixed number and it will depend on the computer specs.

Try cleaning the shapefile if it has more than one column of information. Where is this shp from?

Fair enough, the .shp will be from other software I’m using DF Scanner for AI mapping tree crowns.