thanks for the video. It has been very helpful for obtaining a clean CHM. Only thing left would have been to include audio. Ideas are better expressed with words ( printed text or with audio). Great job!
Yes it is a simple approach, and that makes them so senseful :-) But keep attention with SfM-based pointclouds. These clouds tend under some conditions to 'bend' the pointcloud like a soup tureen (german: Suppenschüssel). It mean it is deformed vertically. If the scene is large enough u can get false Z-coordinates in a range of 30 to 50 meter! U have always to proof the coherency of the data against an aerial rerefence like an dGPS-net or ALS-data. To perform this u have to apply a statistical methods. Check James et al. 2014, 2017 and 2020 for these issue. 1) www.researchgate.net/publication/262773972_Mitigating_systematic_error_in_topographic_models_derived_from_UAV_and_ground-based_image_networks 2) DOI: 10.1002/esp.4125 3) onlinelibrary.wiley.com/doi/epdf/10.1002/esp.4878
U mean structural damages based on 3d-lidar data? Actual i have not the time, but the way is very simple: Use 2 clouds (before and after) and use distance-calculation, best on M3C2 algorithm, and 'compare' both clouds on the vertical (Z) direction and projected on the 'old' pointcloud to see, which trees have gone and how high they were.. That's the kind of calculation, CloudCOMPARE has his name. Direct comparison of 3d point clouds. M3C2 is already implemented in CC. But keep attention - it may cost a lot of calculatng time. The way we're gone on forestry was exactly that way. Afterwards u can export 2d-Raster via 2d-raster-export-function in CC.
I also have a question, how can I obtain ground points in steep terrain? I have used the CSF, but it got very few ground points. I guess that this method does not satisfy the ground point separation under this condition. So, Do you have any good suggestions？ Thanks a lot.
Hi, i don't know which sensor was used. The data came from my professor. I think it is probably a Riegl ALS-Sensor. The data was not collected by us directly, i think it was a commercial campaign. For details to the data ask Prof. Dr. Bodo Bookhagen from Institute of Geoscience of the University Potsdam, Germany. Terminus is 'el pilar, guatemala'
Actually i am working on a self calibrating drought monitor based on phenology and sentinel-2 time-series. In near future i would like to test tri-stereo sat-images for point-cloud generation. But these data is very expensive - too expensive for me as a private man.
I am not the author of these data. It came from my Prof @ University Potsdam. Please ask the remote-sensing Professor Dr. Bookhagen for the data. www.geo.uni-potsdam.de/mitarbeiterdetails/show/524/Bodo_Bookhagen.html
Hmm, in which format comes the data? Cloudcompare is able to handle lot of binary and cleartype formats. A simple way is to bring the data in ASCII-Text form an then modify the header row in where u change X and Z (like u would change the head of an excel table. But u need an editor who can handle these amount of data. R-studio is a good way to handle simple ASCII data import. @ all: SORRY FOR ADs - i had marked the video an 'NON-COMMERCIAL' but TH-cam SENDS STILL ADs in front of my video - Impudence !!!!!!!!! :-(
Actually i am working on an individual tree-detection (loss) approach on my job in municipal forest authority. Exported CHM-Raster can be used for that.
@@ojumle very good. I map the water catchment area in the dam. There is an area of 70 hectares to will be waterproofed. There is a lot of maquis vegetation and I was looking for a solution that could best model the ground.
How can I find the ground line in it by taking sections. Because this technique did not work well for me. It has a lot of bursts of color and it's very difficult to layer.
@@ojumle many thanks. I 'm using the beta version (2.6.3) and CSF Filter not exist at all. Doesn't exist a way to import the plug in CSF in this version? The newest version 2.11.3 of CloudCompare doesn't run in my pc :( I don't know why : it appears CC has stopped working).
Hi, a simple way is to use the Cloudcompare integrated Raster-function with some interpolation-options. It is senseful to use not too small cellsize (maybe 0.5 to 1m) , otherwise the output is not suitable for CHM-modeling. Another way is to calculate rastered ground-signal and rastered maximum height signal and substract it from another to get a CHM-Raster.
@@ojumle Thanks for your response.I'd like to ask a few questions, if you don't mind. Can I export the attribute table of the CHM results? I want to get more detail on each of these points and analyze them. I mean not converting to raster data.
Hi, i think no, because these data came from my Professor and i am not sure about copyrights. It was a georeferenced LAZ-file including elevation information - a Standard LAZ-file. The key is to calculate a ground-model and perform a Cloud-2-Cloud distance-calculation in where u project the differences in the non-ground point-cloud an apply a color-visualidation.
@@ojumle Hola!! muchas gracias por tu respuesta! ok trataré de buscar algunos datos alternativos y ver si puedo realizar el ejercicio! cualquier cosa te pregunto por aqui o no se si por correo electronico? Muchas gracias, Espero tengas un año nuevo lleno de paz, alegrias, mucha salud para poder disfrutar cada momento. Que todo te salga bien y feliz año nuevo para ti y tu familia! exitos!!!
Hi, i can send u a segment of our 3d-camus model, on what i do my master-thesis at this time. I can send it, because i am the author of this dataset. We used a Riegl terrestrial laserscanner to scan the complete campus, some areas have more wood, some areas are typical for light urban side and we have a pond and a train-station including street-tunnel. I have two options: Option1 is a RGB colored and strong filtered dataset but with not the best geo-referencing an some overlays of different scan-positions. My work was to reassamble the whole dataset to rise the inner accuracy. Now i perform a height-analysis of the new dataset with dGPS-GCPs. If u want, i can send u the GPS dataset too. The new dataset is not filtered and have no RGB data (an option is to use googleearth to color the points afterwards). We used thes dataset in the lectures for 2 years and have many work-experience with Lidar. Check remote-sensing section of our University ( www.uni-potsdam.de/en/mnfakul/study-and-teaching/master/remote-sensing-geoinformation-and-visualization.html) contact B. Bookhagen for el-pilar data. Greetings and a new year!
thanks for the video. It has been very helpful for obtaining a clean CHM. Only thing left would have been to include audio. Ideas are better expressed with words ( printed text or with audio). Great job!
Thank you so much for the video, despite having no sound it was still a great tutorial !
Thanks for the tutorial! I tried it with a point cloud of a UAV SFM and got good results.
Yes it is a simple approach, and that makes them so senseful :-) But keep attention with SfM-based pointclouds. These clouds tend under some conditions to 'bend' the pointcloud like a soup tureen (german: Suppenschüssel). It mean it is deformed vertically. If the scene is large enough u can get false Z-coordinates in a range of 30 to 50 meter! U have always to proof the coherency of the data against an aerial rerefence like an dGPS-net or ALS-data. To perform this u have to apply a statistical methods.
Check James et al. 2014, 2017 and 2020 for these issue.
1) www.researchgate.net/publication/262773972_Mitigating_systematic_error_in_topographic_models_derived_from_UAV_and_ground-based_image_networks
2) DOI: 10.1002/esp.4125
3) onlinelibrary.wiley.com/doi/epdf/10.1002/esp.4878
Thanks for these articles! They have an important conceptual framework that it will be very good for me to learn. Greetings! 🥣🥣🥣
Amazing tool!. Gear job!
Can you please make a video on snow damages by using lider data ?
U mean structural damages based on 3d-lidar data? Actual i have not the time, but the way is very simple:
Use 2 clouds (before and after) and use distance-calculation, best on M3C2 algorithm, and 'compare' both clouds on the vertical (Z) direction and projected on the 'old' pointcloud to see, which trees have gone and how high they were.. That's the kind of calculation, CloudCOMPARE has his name. Direct comparison of 3d point clouds. M3C2 is already implemented in CC. But keep attention - it may cost a lot of calculatng time.
The way we're gone on forestry was exactly that way. Afterwards u can export 2d-Raster via 2d-raster-export-function in CC.
amazing, thanks for the tutorial
I didn't find the filter by values (1' 40'') in my version. I I am curious where is that tool. Looking forward to your reply，thanks
Edit->Scalar Field->Filter By Value
@@ojumle thanks a lot
I also have a question, how can I obtain ground points in steep terrain? I have used the CSF, but it got very few ground points. I guess that this method does not satisfy the ground point separation under this condition. So, Do you have any good suggestions？ Thanks a lot.
Hi, great vid. What sensor or package do you use?
Hi, i don't know which sensor was used. The data came from my professor. I think it is probably a Riegl ALS-Sensor. The data was not collected by us directly, i think it was a commercial campaign. For details to the data ask Prof. Dr. Bodo Bookhagen from Institute of Geoscience of the University Potsdam, Germany. Terminus is 'el pilar, guatemala'
Actually i am working on a self calibrating drought monitor based on phenology and sentinel-2 time-series. In near future i would like to test tri-stereo sat-images for point-cloud generation. But these data is very expensive - too expensive for me as a private man.
Buonasera dove trovo il plugin lidar file maker in formato zip da installare in Qgis. GRAZIE
I don't know, i think there was a Rapidlasso LasTools-Interface for pointclouds in QGIS. Otherwise via Python-Interface in QGIS.
Is this data available to download?
I am not the author of these data. It came from my Prof @ University Potsdam. Please ask the remote-sensing Professor Dr. Bookhagen for the data.
www.geo.uni-potsdam.de/mitarbeiterdetails/show/524/Bodo_Bookhagen.html
Hi, When I'm trying to do canopy height model, x coordinate is coming as height. How to solve this problem. Kindly help me
Hmm, in which format comes the data? Cloudcompare is able to handle lot of binary and cleartype formats. A simple way is to bring the data in ASCII-Text form an then modify the header row in where u change X and Z (like u would change the head of an excel table. But u need an editor who can handle these amount of data. R-studio is a good way to handle simple ASCII data import.
@ all: SORRY FOR ADs - i had marked the video an 'NON-COMMERCIAL' but TH-cam SENDS STILL ADs in front of my video - Impudence !!!!!!!!! :-(
im shearching this video, thanx.
Actually i am working on an individual tree-detection (loss) approach on my job in municipal forest authority. Exported CHM-Raster can be used for that.
@@ojumle very good. I map the water catchment area in the dam. There is an area of 70 hectares to will be waterproofed. There is a lot of maquis vegetation and I was looking for a solution that could best model the ground.
How can I find the ground line in it by taking sections. Because this technique did not work well for me. It has a lot of bursts of color and it's very difficult to layer.
@@nedretbozkurt8526 I don't understand the question.
I didn't found Csf in my version..
There was a new release of CC some time ago. The button is on the right vertical orientated button bar called "CSF Filter"
@@ojumle many thanks. I 'm using the beta version (2.6.3) and CSF Filter not exist at all.
Doesn't exist a way to import the plug in CSF in this version?
The newest version 2.11.3 of CloudCompare doesn't run in my pc :( I don't know why : it appears CC has stopped working).
@@antonionuzzi7911 I cannot give u an answer to this question, i am not a programmer, i am a geoscientist ;-)
How to export this CHM?
Hi, a simple way is to use the Cloudcompare integrated Raster-function with some interpolation-options. It is senseful to use not too small cellsize (maybe 0.5 to 1m) , otherwise the output is not suitable for CHM-modeling. Another way is to calculate rastered ground-signal and rastered maximum height signal and substract it from another to get a CHM-Raster.
@@ojumle Thanks for your response.I'd like to ask a few questions, if you don't mind. Can I export the attribute table of the CHM results? I want to get more detail on each of these points and analyze them. I mean not converting to raster data.
Hola!! puedes poner los datos? me gustaria hacer el ejercicio paso a paso, gracias
Hi, i think no, because these data came from my Professor and i am not sure about copyrights. It was a georeferenced LAZ-file including elevation information - a Standard LAZ-file.
The key is to calculate a ground-model and perform a Cloud-2-Cloud distance-calculation in where u project the differences in the non-ground point-cloud an apply a color-visualidation.
@@ojumle Hola!! muchas gracias por tu respuesta! ok trataré de buscar algunos datos alternativos y ver si puedo realizar el ejercicio! cualquier cosa te pregunto por aqui o no se si por correo electronico? Muchas gracias, Espero tengas un año nuevo lleno de paz, alegrias, mucha salud para poder disfrutar cada momento. Que todo te salga bien y feliz año nuevo para ti y tu familia! exitos!!!
Hi, i can send u a segment of our 3d-camus model, on what i do my master-thesis at this time. I can send it, because i am the author of this dataset. We used a Riegl terrestrial laserscanner to scan the complete campus, some areas have more wood, some areas are typical for light urban side and we have a pond and a train-station including street-tunnel. I have two options: Option1 is a RGB colored and strong filtered dataset but with not the best geo-referencing an some overlays of different scan-positions. My work was to reassamble the whole dataset to rise the inner accuracy. Now i perform a height-analysis of the new dataset with dGPS-GCPs. If u want, i can send u the GPS dataset too. The new dataset is not filtered and have no RGB data (an option is to use googleearth to color the points afterwards). We used thes dataset in the lectures for 2 years and have many work-experience with Lidar. Check remote-sensing section of our University ( www.uni-potsdam.de/en/mnfakul/study-and-teaching/master/remote-sensing-geoinformation-and-visualization.html) contact B. Bookhagen for el-pilar data.
Greetings and a new year!
Are u from argentia?
@@ojumle no, soy de Costa Rica!😊