Autodesk - ReCap Future

Discuss all Autodesk ReCap Studio related issues here.
jamesworrell
V.I.P Member
V.I.P Member
Posts: 634
Joined: Mon Jun 16, 2014 1:45 pm
Full Name: James Worrell
Company Details: Bennett and Francis
Company Position Title: Director
Country: Australia
Linkedin Profile: Yes
Location: Brisbane, Queensland, Australia
Been thanked: 11 times
Contact:

Re: Autodesk - ReCap Future

Post by jamesworrell »

VXGrid wrote:
Fri Feb 07, 2020 2:14 pm
The question is: Is the functionality sufficient to do the job it is needed for. As long as there is no out of the world approach for registering scans, active development is not necessary, as long as the product is maintained (import of scan data to new release formats is updated, bugs are getting fixed, ...).
This is the crux of it .. why develop Recap? What further functionality would the current technology really offer beyond a few UI/UX tweaks?

The environment effectively needs to be re-factored to server-based/centralised delivery - be it cloud or on-premise. Jetstream+Cloudworx is in effect the future delivery model - it just needs more competition to keep it honest.

AutoCAD and Revit et. al. have "provider models" for point cloud data - look at Cloudworx - again it needs more competition.

Euclideon - get a plug-in for Revit to draw data from the Vault - a viewer is of no use in and of itself - and I know Jetstream has Euclideon tech in there. There is no point having point clouds siloed for viewing only in isolation - the point clouds need to be where other data is used so you can then visualise, clash, model, extract yada yada. Pointerra are rumouredly working on same. Heard Aerometrex are playing around with the idea, but not sure of that one either.

Once you have a server-based point cloud stream that is effective - the game changes. And if you have a few providers that can do it, pricing will become reasonable.

Server-side also makes things centralised processing engines a piece of cake. Workflows like meshing, AI/ML object recognition, automated extract - all win, rather than having copies of the same TB-sized point clouds all over the place with processing taking place in dribs and drabs. The whole lot needs a consistent API so it can be automated and scripted.

User avatar
smacl
Global Moderator
Global Moderator
Posts: 588
Joined: Tue Jan 25, 2011 5:12 pm
Full Name: Shane MacLaughlin
Company Details: Atlas Computers Ltd
Company Position Title: Managing Director
Country: Ireland
Linkedin Profile: Yes
Location: Ireland
Has thanked: 49 times
Been thanked: 55 times
Contact:

Re: Autodesk - ReCap Future

Post by smacl »

jamesworrell wrote:
Sat Feb 08, 2020 1:36 am
The environment effectively needs to be re-factored to server-based/centralised delivery - be it cloud or on-premise.
Depends to some extent on the job. If the site doesn't have access to a high speed internet connection and is in any way remote, we'd often see people doing initial processing work back in the site office or hotel room after collection to check for gross errors and omissions. I see huge value in having the registered well structured point cloud accessible from a shared platform for processing purposes, but also see similar value with all the QA happening without the overhead of unnecessary data transfer. Not sure whether sharing data that hasn't been through the QA stages is a great idea and I tend to advocate for the person with site knowledge doing the QA rather than separating collection and QA into different functions done by different people.

Also not sure about the relative processing cost of cloud versus local workstations with all other costs factored in? Maybe Denis would know a bit more here.

jamesworrell
V.I.P Member
V.I.P Member
Posts: 634
Joined: Mon Jun 16, 2014 1:45 pm
Full Name: James Worrell
Company Details: Bennett and Francis
Company Position Title: Director
Country: Australia
Linkedin Profile: Yes
Location: Brisbane, Queensland, Australia
Been thanked: 11 times
Contact:

Re: Autodesk - ReCap Future

Post by jamesworrell »

smacl wrote:
Sat Feb 08, 2020 10:54 am
jamesworrell wrote:
Sat Feb 08, 2020 1:36 am
The environment effectively needs to be re-factored to server-based/centralised delivery - be it cloud or on-premise.
Depends to some extent on the job. If the site doesn't have access to a high speed internet connection and is in any way remote,
Hence the bit “on-premise” .. a client-server arrangement can be on the same box if necessary.


Personally cloud is always too expensive. On prem for the normal
loads .. cloud is for the extreme peaks.

VXGrid
Forum Supporter
Forum Supporter
Posts: 204
Joined: Fri Feb 24, 2017 10:47 am
Full Name: Martin Graner
Company Details: PointCab GmbH
Company Position Title: Research and Development
Country: Germany
Linkedin Profile: No
Has thanked: 6 times
Been thanked: 13 times

Re: Autodesk - ReCap Future

Post by VXGrid »

smacl wrote:
Fri Feb 07, 2020 3:45 pm

With point cloud sizes continuously increasing and many related 3d algorithms having greater than linear complexity, e.g. > O(n), a point cloud engine that isn't continuously improved upon is likely to become an unacceptable bottleneck to workflows reasonably quickly. I think laser scanning and point cloud processing is such a technologically dynamic area at the moment, software solutions that don't address change with added value will become obsolete.
I don't think that all cloud engines gets constant improvement (why do that if there are no new algorithms). You get it working and if the sizes are so dramatically rising, you need to have a completely other approach anyway.
ReCap down samples the point cloud to 1mm cubicles, as long as the area size doesn't change, the engine does not get a challenge, even if the scanner is producing 20 times the points per scan position.
And speedwise? It's RCP it will take forever anyway...

jamesworrell wrote:
Sat Feb 08, 2020 1:36 am

This is the crux of it .. why develop Recap? What further functionality would the current technology really offer beyond a few UI/UX tweaks?

[...]
The reason why every big company is producing software solutions around their existing products I guess. To generate a complete workflow in their respective software packages so the customer never needs to leave the big company software bubble.
Getting the Drawing community to start their workflow one step prior with the scanning + registration.
Perfect fit for small applications, low accuracy applications, short training periods.
jamesworrell wrote:
Sat Feb 08, 2020 1:36 am

[...]

The environment effectively needs to be re-factored to server-based/centralised delivery - be it cloud or on-premise. Jetstream+Cloudworx is in effect the future delivery model - it just needs more competition to keep it honest.

[...]

Once you have a server-based point cloud stream that is effective - the game changes. And if you have a few providers that can do it, pricing will become reasonable.

Server-side also makes things centralised processing engines a piece of cake. Workflows like meshing, AI/ML object recognition, automated extract - all win, rather than having copies of the same TB-sized point clouds all over the place with processing taking place in dribs and drabs. The whole lot needs a consistent API so it can be automated and scripted.

I don't think that point cloud server streaming options are going to be the future.
The sizes are not handlable when uploading, streaming the cloud back to your computer is only working when connected to high speed internet, the point cloud is merly a intermediate product.

In my opinion in the future there will be automatic object creation out of the cloud (no meshes, real objects) and direct deletion of the points used for these objects, and all points which can't be assigned are left in there, so a small part of the point cloud is left, for human interpretation.
Object attributes can be extracted by only the images, no need for the point cloud.

yankoch
V.I.P Member
V.I.P Member
Posts: 675
Joined: Wed Oct 05, 2011 10:14 am
Full Name: Yan Koch
Company Details: TagLabs
Company Position Title: President
Country: France
Linkedin Profile: Yes
Has thanked: 10 times
Been thanked: 25 times

Re: Autodesk - ReCap Future

Post by yankoch »

VXGrid wrote:
Sat Feb 15, 2020 10:32 am
I don't think that point cloud server streaming options are going to be the future.
The sizes are not handlable when uploading, streaming the cloud back to your computer is only working when connected to high speed internet, the point cloud is merly a intermediate product.
+1
Moerover, we must not forget that streaming Gigabytes of pointclouds is an absolute disaster in terms of bandwidth and data center consumption.
Many people "play" with the pointcloud online because it's fun, ignoring it is a Netflix².
VXGrid wrote:
Sat Feb 15, 2020 10:32 am
In my opinion in the future there will be automatic object creation out of the cloud (no meshes, real objects) and direct deletion of the points used for these objects, and all points which can't be assigned are left in there, so a small part of the point cloud is left, for human interpretation.
Object attributes can be extracted by only the images, no need for the point cloud.
Yes, extracting useful objects - 3D solids + information - is also what I see as the future of pointclouds. The pointcloud will still be used to show the global environment, and also to do accurate measurements.
Kind regards,

Yan Koch
[email protected]

Post Reply

Return to “Autodesk ReCap Studio”