TOOL LIDAR data - available for free for England
- Thread starter assettocorsamods
- Start date
Pixelchaser
Well-Known Member
if the data is right, this is right, I assure you on that. and I doubt there is a way to get closer than an expensive ground scan.
Mr Whippy
Active Member
The data isn't right.
That's why you're having to process it
I get what you mean, I'm just being the devils advocate with regards to what the data can really offer.
Like I said before, sometimes you're better doing it 'by eye' like lilski said he did.
With datasets like these your eye is going to be as good as any untested smoothing algorithm.
It's slow, I've spent days going around track edges by eye on 20km of mobile laser scanned road, but it was the only way to gauge the real floor due to hedgerows, ditches, grasses etc at the roadsides.
Of course I'm sure what you've done works ok etc.
This is just more reference for others reading the thread and getting the data.
Don't rush for smoothers when the mk1 eyeball and human brain is still a super powerful toolset
That's why you're having to process it
I get what you mean, I'm just being the devils advocate with regards to what the data can really offer.
Like I said before, sometimes you're better doing it 'by eye' like lilski said he did.
With datasets like these your eye is going to be as good as any untested smoothing algorithm.
It's slow, I've spent days going around track edges by eye on 20km of mobile laser scanned road, but it was the only way to gauge the real floor due to hedgerows, ditches, grasses etc at the roadsides.
Of course I'm sure what you've done works ok etc.
This is just more reference for others reading the thread and getting the data.
Don't rush for smoothers when the mk1 eyeball and human brain is still a super powerful toolset
Pixelchaser
Well-Known Member
the data is correct. processing it for game world is essential. no other way. I just never expected it to be so process able but ive got better at everything lately. my point is I have perfectly processed it. this data is the most accurate data ive ever seen of a real world location. and ive measured ever change over the processed raw with the raw simultaneously ,quite literally you can over think all of this. I havn`t and I have it right here, I cant see where i can go wrong. ive strived for this level of data for years today I have it. just track making skillz now and my best tool always has been my eye. happy days because it doesn't get better than this, correct or incorrect. its always 0.7m incorrect and nobodys eye is going to get 5km x 5km more correct. the game doesn't require it.
higher rez internal area ill merge next.
if anyone wants an area processing ill happily oblige.
today I have it. just track making skillz now and my best tool always has been my eye. happy days because it doesn't get better than this, correct or incorrect. its always 0.7m incorrect and nobodys eye is going to get 5km x 5km more correct. the game doesn't require it.higher rez internal area ill merge next.
if anyone wants an area processing ill happily oblige.
Last edited:
Mr Whippy
Active Member
Sorry Pixelchaser I thought you were talking about the track surface data, not terrain.
Is the track surface quality comparable to the Oulton Park mesh I uploaded the other year?
Or is it more like the lidar data that was posted from the Highways Agency from their on-van scanners?
Yes you can end up over thinking what the data tells you.
That's why mk1 eyeball is better for the track surface than 'algorithms' you just fiddle with without base reference.
I've worked from 50m grids down to sub 5mm noise floor laser scans on public roads.
I'd *always* go with my eye over any other process for the track shape itself, if you can't directly use the dataset to get the road form correct.
Quite often bits of cars not removed from the scans properly, or maybe even a couple of points on a cone or something, throw results off.
Like you, you'd prefer to paint in PS over have textures automated, and so it stands here with track surfacing.
For the very best track surface with non-perfect data, the old slow way is still the best way
What is the track btw? Care to share the laz file?
PS, the U.K. OS lidar, dtm, dsm data is now also joined by ortho photo tiles.
They're an odd compressed format but should convert ok in xnview 32bit with plugin I think.
So they'll be nice to drape over dsm/dtm or underlay as references perfectly aligned to your dataset!
I'm still thinking for a semi-pro track that a decent drone pass + photogrammetry is best.
You get lidar like quality for less price, with fantastic colour data baked in by design. So you can perfectly model over it and see track edges etc.
Hmmmm.
Is the track surface quality comparable to the Oulton Park mesh I uploaded the other year?
Or is it more like the lidar data that was posted from the Highways Agency from their on-van scanners?
Yes you can end up over thinking what the data tells you.
That's why mk1 eyeball is better for the track surface than 'algorithms' you just fiddle with without base reference.
I've worked from 50m grids down to sub 5mm noise floor laser scans on public roads.
I'd *always* go with my eye over any other process for the track shape itself, if you can't directly use the dataset to get the road form correct.
Quite often bits of cars not removed from the scans properly, or maybe even a couple of points on a cone or something, throw results off.
Like you, you'd prefer to paint in PS over have textures automated, and so it stands here with track surfacing.
For the very best track surface with non-perfect data, the old slow way is still the best way
What is the track btw? Care to share the laz file?
PS, the U.K. OS lidar, dtm, dsm data is now also joined by ortho photo tiles.
They're an odd compressed format but should convert ok in xnview 32bit with plugin I think.
So they'll be nice to drape over dsm/dtm or underlay as references perfectly aligned to your dataset!
I'm still thinking for a semi-pro track that a decent drone pass + photogrammetry is best.
You get lidar like quality for less price, with fantastic colour data baked in by design. So you can perfectly model over it and see track edges etc.
Hmmmm.
Pixelchaser
Well-Known Member
well its better than that oulton park imo. its just the terrain. nothing else, obviously its heavily edited at their end. there is some issues with the side barriers. they are classed as terrain and it changes the height over the sides, but nothing that cant be fixed by eye. its Lime Rock Park. I have a post dedicated to it here. http://assettocorsamods.net/threads/lime-rock.887/#post-4232
pm sent.
pm sent.
Last edited:
Pixelchaser
Well-Known Member
btw lilski you referred to obtaining a track on a single tile. I realise my file when captured as a large area was covering 2 tiles. but when I order through here, they email me a link to the exact selection I want.
https://coast.noaa.gov/dataviewer/#/lidar/search/
https://coast.noaa.gov/dataviewer/#/lidar/search/
Pixelchaser
Well-Known Member
2016 UK data is now live.
http://environment.data.gov.uk/ds/partners/#/partners/login. Over 18,000 sq km of new LIDAR data has been added since the 2014 Composite and there is now coverage of 60% of England at 1m resolution or higher
http://environment.data.gov.uk/ds/partners/#/partners/login. Over 18,000 sq km of new LIDAR data has been added since the 2014 Composite and there is now coverage of 60% of England at 1m resolution or higher
Pixelchaser
Well-Known Member
posted walkthrough video of what i do to get from website to game reference mesh. over on my lime rock thread but it might be usefull here. regards grabbing data data and turning it into reference meshes.
Lime Rock
so...having learned a bit recently ,I thought id have a look again at my favourite british track and its lidar. oulton park. and after understanding the necessity of the intensity image looking at it now it begs to be done.
but the difference with this data is there is no intensity map and you must generate it in cloud compare. however when you have the intensity layer selected you can not render to raster or image. the only option I see is fricken screen grab. @LilSKi might you know about this ? and maybe how to generate a proper intenisity map from cloud compare itself. I know it fits. on screengrab style mode thingy "capture" it has a yellow frame and using it seems to fit as in the gif. but is there a way to render it at a decent size ?
Lime Rock
so...having learned a bit recently ,I thought id have a look again at my favourite british track and its lidar. oulton park. and after understanding the necessity of the intensity image looking at it now it begs to be done
.
but the difference with this data is there is no intensity map and you must generate it in cloud compare. however when you have the intensity layer selected you can not render to raster or image. the only option I see is fricken screen grab. @LilSKi might you know about this ? and maybe how to generate a proper intenisity map from cloud compare itself. I know it fits. on screengrab style mode thingy "capture" it has a yellow frame and using it seems to fit as in the gif. but is there a way to render it at a decent size ?
Last edited:
LilSKi
Well-Known Member
I just did a screen grab in cloud compare and then cropped it to the yellow border. I then put it on a plane with the same aspect ratio as the image and scaled the plane to the data. You can use a few far away points like buildings to line it up with the data. I then move the intensity plane up and down on the Z axis through the data and check alignment at multiple points. Might be a better way but that is what I did for NJMP.
Pixelchaser
Well-Known Member
yeah thats what i did and its ok as above ... the NOAA data and its intensity tiff file that they give you when you order the explicit document by selection area was perfectly aligned for the model and a decent size too. 1pixel per metre. this method must be half the resolution.
I've not done any track modelling before, but keen to try and make at least one of the tracks I'm competing at next year. Here's the calendar with a mix of sprints, hillclimbs and AutoSolos (coned circuits on old airfields):
24th March Great Western Sprint, Castle Combe Circuit
8th April Kemble AutoSolo
12th May Llandow Sprint
26th May Westonzoyland AutoSolo
22nd July Clay Pigeon Sprint
28th July **** Mayo Sprint, Castle Combe Circuit
5th Aug Kemble AutoSolo
27th Aug Wessex Sprint, Lyneham
9th Sep Westonzoyland AutoSolo
22nd Sep Manor Farm Hillclimb
23rd Sep Manor Farm Hillclimb
One difficulty is that whilst all the events are timed point-to-point, some of the courses (Great Western, Llandow and Clay Pigeon) are 1.75 laps, so you pass the finish line twice. This is a nightmare with datalogging software and probably with AC too...
24th March Great Western Sprint, Castle Combe Circuit
8th April Kemble AutoSolo
12th May Llandow Sprint
26th May Westonzoyland AutoSolo
22nd July Clay Pigeon Sprint
28th July **** Mayo Sprint, Castle Combe Circuit
5th Aug Kemble AutoSolo
27th Aug Wessex Sprint, Lyneham
9th Sep Westonzoyland AutoSolo
22nd Sep Manor Farm Hillclimb
23rd Sep Manor Farm Hillclimb
One difficulty is that whilst all the events are timed point-to-point, some of the courses (Great Western, Llandow and Clay Pigeon) are 1.75 laps, so you pass the finish line twice. This is a nightmare with datalogging software and probably with AC too...
Pixelchaser
Well-Known Member
Westonzoyland is the only one with data.
Pixelchaser
Well-Known Member
yeah full data. I have castle coombe layout fully built to be honest, but its flat, waiting for lidar.
Pixelchaser
Well-Known Member
one can convert that gps to suitable csv text file, . but it doesn't capture the actual track per say.
Pixelchaser
Well-Known Member
edited as you posted there. yeah send it over ill have a look at it. there are options for csv, comma separated values. if you have the option send 1 with commas and 1 with no commas but spaces separating... if you can
Having a thumb through my old data and most of my loggers (I've had a few) don't do altitude. However, I can nip down with a GPS dongle and get some altitude data, but it won't be any time soon as they don't let people walk on the track much.
The EA data includes Hammerdown to Westway and that's it. That section is one of the biggest height change areas though, so will be a huge improvement. From Quarry to Hammerdown it's basically flat and from Westway to Camp is also flat. I've cycled many miles around there on a bicycle and you can really feel the gradient!
The EA data includes Hammerdown to Westway and that's it. That section is one of the biggest height change areas though, so will be a huge improvement. From Quarry to Hammerdown it's basically flat and from Westway to Camp is also flat. I've cycled many miles around there on a bicycle and you can really feel the gradient!
Share this page
Support the site
