Odm: Support for crop/canopy height calculation (DSM - DTM)

[pierotofy]

This is present in the lidar2dems code, should be fairly simple to implement by subtracting DTM from DSM. It was suggested on gitter.

[yjmenezes]

@pierotofy

Point classification from RGB photos is a hard task.
I plan try lidar2dems. http://applied-geosolutions.github.io/lidar2dems/

I upload to my git a small urban area. 10 photos + LiDAR.
I have rural areas too. Some flight.

I hope it can be usefull for tests.

[bipulneupane]

@pierotofy thank you for your help and apologies for late gratitude.

@yjmenezes Thank you for your help. Can i make CHM out of the DSM and DTM images that are created not by using l2d? I want to get CHM from the outputs of WebODM? Is there any limitation while doing so?

[yjmenezes]

Hi @bipulneupane

As you, i am looking for .

I am note sure if @hobu PDAL has a filter to produce bare earth DTM from ODM point cloud (XYZ+RGB).
https://www.pdal.io/stages/filters.html

The problem is shown here:
https://rapidlasso.com/2015/09/21/creating-dtms-from-dense-matched-points-of-uav-imagery-from-senseflys-ebee/

I upload a LiDAR + RGB data set in a hope that someone could use it and help to make ODM even better. ODM is a great work. Thank you ODM developers.

[bipulneupane]

@yjmenezes
Hi!

I wanted to know

1. if it is always necessary to classify the point cloud first
2. Error while density:
   $ l2d_dems density georeferenced_model.las

Error creating density : No LAS files found
l2d_dems density completed (.) in 0:00:00.000208

I am getting error on the first step of classifying.

[yjmenezes]

@bipulneupane

I think you must point the directory and not a las file.
l2d_dems density /path_to_las/

i have never used l2d_dems, i am sorry.
I think it's primary use

is for LiDAR data.

btw, I upload a LAS LiDAR dataset . It is not carefully classified but has all laser return pulses.
Class (2) would be bare earth.
https://github.com/yjmenezes/church_dataset

[GeospatialDaryl]

LiDAR data (like in the referenced diagram) includes an intensity value;
which is the response peaks on the graph. SfM PC will not have that
information.

To develop a CSM, you will need to develop a Ground Surface model, and then
use that to classify your point cloud, then calculate separation of those
surfaces.

On Thu, Feb 8, 2018 at 10:48 AM, julio cesar de menezes <
[email protected]> wrote:

@bipulneupane https://github.com/bipulneupane

I think you must point the directory and not a las file.
l2d_dems density /path_to_las/

i have never used l2d_dems, i am sorry.
I think it's primary use
[image: lidar_las]
https://user-images.githubusercontent.com/4775645/35991794-77611dce-0cef-11e8-8970-bc2209398810.png
is for LiDAR data.

btw, I upload a LAS LiDAR dataset . It is not carefully classified but has
all laser return pulses.
Class (2) would be bare earth.
https://github.com/yjmenezes/church_dataset

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
https://github.com/OpenDroneMap/OpenDroneMap/issues/752#issuecomment-364210327,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AFOkJZmaK5o1I5Nv5nihM4z2idBLkuaWks5tS0F4gaJpZM4RuU3H
.

--

---

Daryl Van Dyke
GIS Analyst
Klamath Strategic Habitat Conservation Team
US Fish & Wildlife Service - AFWO, R8
[phone]
https://github.com/GeospatialDaryl

[dakotabenjamin]

I can see this being implemented as a contrib/ module

[lucadelu]

Here you can find Python code using PDAL (probably it should be updated) to calculate CHM from input LAS file and DTM.

Do you think could be useful? I could help to adapt to OpenDroneMap with some suggestions

[pierotofy]

@lucadelu that would be super-useful! Here's some things that stand out to me:

• PDAL code will need to use JSON, not XML (as XML is being phased out in PDAL)
• I would explore to see if it would be better to calculate the CHM based on the existing rasters (DSM and DTM) rather than using PDAL's python filter. Just a thought.

[lucadelu]

@pierotofy thanks for your answer, yes we should use to use JSON, this was old code..

About CHM calculation we could have several way to calculate:

• from existing rasters but at this point we could do this using existing software like GRASS GIS through mapcalc operation
• from an existing DTM and the calculated DSM from OpenDroneMap (still using GRASS GIS)
• from an existing DTM and the OpenDroneMap point cloud (using PDAL)