Automated point cloud vectorization
-
- V.I.P Member
- Posts: 1037
- Joined: Tue Mar 29, 2011 7:39 pm
- 13
- Full Name: Scott Page
- Company Details: Scott Page Design- Architectural service
- Company Position Title: Owner
- Country: USA
- Linkedin Profile: No
- Location: Berkeley, CA USA
- Has thanked: 205 times
- Been thanked: 78 times
- Contact:
Automated point cloud vectorization
https://www.pythagoras.net/news/pythago ... y-released -video link
https:/player.vimeo.com/videos/pythagoras
- link to all videos
A new release from Pythagoras:
Some of the new features:
Automated point cloud vectorization
Classification of point cloud
Enhanced DTM performance
Updated profiling tools
General performance upgrade
There are now several software packages with the ability to vectorize of point clouds, but I've yet to see a comprehensive review of these products, comparing tools, features, price, and performance.
https:/player.vimeo.com/videos/pythagoras
- link to all videos
A new release from Pythagoras:
Some of the new features:
Automated point cloud vectorization
Classification of point cloud
Enhanced DTM performance
Updated profiling tools
General performance upgrade
There are now several software packages with the ability to vectorize of point clouds, but I've yet to see a comprehensive review of these products, comparing tools, features, price, and performance.
-
- Honorary Member
- Posts: 1047
- Joined: Mon Oct 06, 2008 8:31 pm
- 15
- Full Name: Matthew McCarter
- Company Details: Costain
- Company Position Title: BIM Manager
- Country: England
- Linkedin Profile: Yes
- Has thanked: 5 times
- Been thanked: 4 times
Re: Automated point cloud vectorization
The classify ground looks like it did a good job, but this is possible in lots of software.
The vectorise edges from an image of road markings had a lot of extra lines created that would need clean up. Creating those manually would be quicker overall I reckon.
As for the vectorising of the building facade from point cloud data, there was even more extra lines created where no detail exists.
Unfortunately, it still looks like the cleanup time is too much in this sort of product that a well skilled CAD monkey can't beat and probably with tidier CAD practice. (not lots of unneeded nodes in the linework etc).
The vectorise edges from an image of road markings had a lot of extra lines created that would need clean up. Creating those manually would be quicker overall I reckon.
As for the vectorising of the building facade from point cloud data, there was even more extra lines created where no detail exists.
Unfortunately, it still looks like the cleanup time is too much in this sort of product that a well skilled CAD monkey can't beat and probably with tidier CAD practice. (not lots of unneeded nodes in the linework etc).
- smacl
- Global Moderator
- Posts: 1409
- Joined: Tue Jan 25, 2011 5:12 pm
- 13
- Full Name: Shane MacLaughlin
- Company Details: Atlas Computers Ltd
- Company Position Title: Managing Director
- Country: Ireland
- Linkedin Profile: Yes
- Location: Ireland
- Has thanked: 627 times
- Been thanked: 657 times
- Contact:
Re: Automated point cloud vectorization
We've done quite a bit of work on automatic vector extraction in SCC, a couple of videos below show some of this.
Automatic kerb line extraction
youtu.be/kgcV5rt3R84
Automatic rail (theoretical running edge) extraction
youtu.be/T9HPA9_aGKQ
Automatic extraction of line work based on section or selection
youtu.be/L1ApPht9Pfw
How well these techniques work in my opinion come down to the quality of data and noise removal / cleaning techniques employed. Mobile LIDAR and good quality UAV data have the added advantage of reasonably evenly spaced points which make statistical analysis of potential features easier.
Automatic kerb line extraction
youtu.be/kgcV5rt3R84
Automatic rail (theoretical running edge) extraction
youtu.be/T9HPA9_aGKQ
Automatic extraction of line work based on section or selection
youtu.be/L1ApPht9Pfw
How well these techniques work in my opinion come down to the quality of data and noise removal / cleaning techniques employed. Mobile LIDAR and good quality UAV data have the added advantage of reasonably evenly spaced points which make statistical analysis of potential features easier.
-
- V.I.P Member
- Posts: 1037
- Joined: Tue Mar 29, 2011 7:39 pm
- 13
- Full Name: Scott Page
- Company Details: Scott Page Design- Architectural service
- Company Position Title: Owner
- Country: USA
- Linkedin Profile: No
- Location: Berkeley, CA USA
- Has thanked: 205 times
- Been thanked: 78 times
- Contact:
Re: Automated point cloud vectorization
+1 Nice demo. I wish all the major platforms included such tools by default: quality data > noise removal > cleaning techniques > happy clients > a cold beer!smacl wrote: ↑Wed May 16, 2018 8:06 am We've done quite a bit of work on automatic vector extraction in SCC, a couple of videos below show some of this.
Re: Automatic extraction of line work based on section or selection.
How well these techniques work in my opinion come down to the quality of data and noise removal / cleaning techniques employed. Mobile LIDAR and good quality UAV data have the added advantage of reasonably evenly spaced points which make statistical analysis of potential features easier.
- smacl
- Global Moderator
- Posts: 1409
- Joined: Tue Jan 25, 2011 5:12 pm
- 13
- Full Name: Shane MacLaughlin
- Company Details: Atlas Computers Ltd
- Company Position Title: Managing Director
- Country: Ireland
- Linkedin Profile: Yes
- Location: Ireland
- Has thanked: 627 times
- Been thanked: 657 times
- Contact:
Re: Automated point cloud vectorization
Thanks Scott, any piece of technology that brings that cold beer closer to hand is a step forward in my book! Are you going to Geo-Business next week?Scott wrote: ↑Wed May 16, 2018 4:00 pm+1 Nice demo. I wish all the major platforms included such tools by default: quality data > noise removal > cleaning techniques > happy clients > a cold beer!smacl wrote: ↑Wed May 16, 2018 8:06 am We've done quite a bit of work on automatic vector extraction in SCC, a couple of videos below show some of this.
Re: Automatic extraction of line work based on section or selection.
How well these techniques work in my opinion come down to the quality of data and noise removal / cleaning techniques employed. Mobile LIDAR and good quality UAV data have the added advantage of reasonably evenly spaced points which make statistical analysis of potential features easier.
- RobM
- I have made 100> posts
- Posts: 101
- Joined: Mon Aug 22, 2016 9:55 am
- 7
- Full Name: Robin Miller
- Company Details: My own boss
- Company Position Title: Managing Director
- Country: Germany
- Linkedin Profile: No
- Been thanked: 14 times
Re: Automated point cloud vectorization
Wow, SCC looks like its handling those tasks really well. I've never seen this software before. Is there a german reseller I could contract about a trial / prices?
- smacl
- Global Moderator
- Posts: 1409
- Joined: Tue Jan 25, 2011 5:12 pm
- 13
- Full Name: Shane MacLaughlin
- Company Details: Atlas Computers Ltd
- Company Position Title: Managing Director
- Country: Ireland
- Linkedin Profile: Yes
- Location: Ireland
- Has thanked: 627 times
- Been thanked: 657 times
- Contact:
Re: Automated point cloud vectorization
Hi Rob, we don't currently have a reseller in Germany but you can download a trial license here. We'll be launching SCC r13 next week at Geo Business in London which has some more automation tools, I'll post an update once it goes on general release and PM a price list.
- smacl
- Global Moderator
- Posts: 1409
- Joined: Tue Jan 25, 2011 5:12 pm
- 13
- Full Name: Shane MacLaughlin
- Company Details: Atlas Computers Ltd
- Company Position Title: Managing Director
- Country: Ireland
- Linkedin Profile: Yes
- Location: Ireland
- Has thanked: 627 times
- Been thanked: 657 times
- Contact:
Re: Automated point cloud vectorization
As part of our drive to come up with faster and highly efficient workflows for processing mobile LIDAR in SCC, we've just completed our initial implementation of an automated white line extraction tool. This is designed to allow the rapid and accurate extraction of dashed and continuous white lines across large mobile LIDAR data sets.
youtu.be/vaR_jwa7QtY
In the above example we show the model creation and extraction of all white lines for an 18km section of motorway made up of ~620 million points where the total processing time on a modest PC including model creation, extraction and editing was ~40 minutes which we're pretty pleased with. Next up is automated extraction of crash barriers and other hard features.
youtu.be/vaR_jwa7QtY
In the above example we show the model creation and extraction of all white lines for an 18km section of motorway made up of ~620 million points where the total processing time on a modest PC including model creation, extraction and editing was ~40 minutes which we're pretty pleased with. Next up is automated extraction of crash barriers and other hard features.
- smacl
- Global Moderator
- Posts: 1409
- Joined: Tue Jan 25, 2011 5:12 pm
- 13
- Full Name: Shane MacLaughlin
- Company Details: Atlas Computers Ltd
- Company Position Title: Managing Director
- Country: Ireland
- Linkedin Profile: Yes
- Location: Ireland
- Has thanked: 627 times
- Been thanked: 657 times
- Contact:
Re: Automated point cloud vectorization
Another example of some automated extraction based on static scanning this time, with semi-automated collection of rail lines (theoretical running edges) and overhead lines. The object here was to get the line height and stagger information relative to plane of rail. While we have fully automated power line extraction in SCC, the semi automated approach works better for static scanning where the position of the line is prone to move between setups (particularly in windy conditions). This leads to 'ghosting', i.e. multiple occurences of the same line from different setups.
youtu.be/c5QorfYnw2g
youtu.be/c5QorfYnw2g
- GCFdesign
- V.I.P Member
- Posts: 308
- Joined: Wed Sep 04, 2019 7:23 am
- 4
- Full Name: Gary Fellows
- Company Details: GCF Design Limited
- Company Position Title: Director
- Country: United Kingdom
- Linkedin Profile: Yes
- Has thanked: 59 times
- Been thanked: 31 times
Re: Automated point cloud vectorization
smacl wrote: ↑Wed May 16, 2018 8:06 am We've done quite a bit of work on automatic vector extraction in SCC, a couple of videos below show some of this.
Automatic kerb line extraction
youtu.be/kgcV5rt3R84
Automatic rail (theoretical running edge) extraction
youtu.be/T9HPA9_aGKQ
Automatic extraction of line work based on section or selection
youtu.be/L1ApPht9Pfw
How well these techniques work in my opinion come down to the quality of data and noise removal / cleaning techniques employed. Mobile LIDAR and good quality UAV data have the added advantage of reasonably evenly spaced points which make statistical analysis of potential features easier.
Please inform me how you have managed to get your YouTube video to be visible and play on here? When l put my YouTube link in, all l get is a link, nothing else.
Cheers in advance.