RIEGL VZ600i

[dhirota]

In the picture with Joshua it seems there is enough lighting although the pointcloud seems a bit darkish. Any reason for that Dennis?

Any chance of a comparison with the blk360g2 data?

The DIH+Joshua photo was done by Ray with my iPhone in auto ISO mode in a very dark SAM's Club. The Riegl VZ600i needed ISO=2000, shutter=1/200
to receive the images shown. We could have raised the ISO higher, but decided to keep it at that level of brightness.

Joshua will send over his BLK360G2 scans after he processes them.

Did you take pictures while scannign or after? I've personally never fiddled with the picture settings on my vz400i.

We took several trial photographic images of the 36 for several initial scan positions to try and figure out the correct ISO and shutter speed. Once we decided after the first 3 or 4 scan positions, we just used identical values without checking each scan position. We are trying to determine if there are any sensors available to qualify the light intensity at the cart height from the floor and x,y position on the store floor?

[jedfrechette]

We are trying to determine if there are any sensors available to qualify the light intensity at the cart height from the floor and x,y position on the store floor?

Do you mean like a light meter with datalogging? Something like this

https://www.bhphotovideo.com/c/product/ ... light.html

would give you camera exposure values and Lux measurements. Do you just want those for reference or do you actually want to be able to log those values for each scan position?

[TommyMaddox]

It was great to visit with Dennis in Honolulu, Hawaii is a truly wonderful place and I would recommend it to anyone on the fence about making time to go.

It's always taken us a lot of effort and tuning to get great color results out of the D850 on the VZ400i. I will share some results of this in the post. We were only able to do so when moving away from the recommended settings and using the auto exposure, auto iso metering per shot on the camera with some other custom settings. You would run into issues when using the 'capture images while scanning' mode if one section of the area you're scanning in (for example scanning near sunrise/sunset with the sun is low on your side) is notably darker or brighter (and thus requires longer exposure time) than possible for the camera to capture while the scanner is in motion. It would have been wonderful if that mode actually paused the horizontal traverse and laser capture momentarily for whatever image capture duration was required as determined by the D850 so that no section of the scan had inconsistent exposure and the motion blur could be eliminated. This is much more noticeable in industrial or commercial environments than it is on civil or mining scopes because the text/markings and sharp lines/edges jump out at you.

It's particularly interesting on the 600i's triple camera internal setup. The FARO units we have will give consistently useful color every day (with the exception of super bright areas which we use the LIHDR processing for) but requires some time to do exposure metering in order to accomplish this. I can't remember if the 600i offered the ability to set the camera properties for each sensor (as they'll each be aiming in a different direction with distinctly different lighting for each one), but I don't remember there being time allocated to automated exposure metering at the time I did testing.

I often wonder if it would be possible to have these sensors perform picture/video capture while scanning to determine which exposure/iso parameters to apply for the optimal performance in actual image capture mode.

That said, the image alignment, field of view, parallax, and resolution was really quite good. I don't know that I would even want to bother with adding the external camera to the 600i unless I was doing a project where very long range image resolution and a corresponding camera/lens package would be required. The impact for us in not having to run a fairly complex field and office workflow to compensate for the alignment issues on the VZ400i/D850 system would be amazing in both reducing the difficulty in operating the system for newer users, and for delivering a much more consistent backend effort level.

Castrol Track Riegl Signage 1.jpg

[dhirota]

Too many other things happening to respond to Jed's and Tommy's responses, so I will try later when I have additional time.

We have completed 6 projects in the HNL urban core using the Riegl VZ600i with RTK for registration. This is for our staff to become familiar with the field and office processing of the Riegl data. We exported the 6 projects (596 scan positions) in E-57 and merged them into Recap 2024 for a total point cloud of 4.15 BillionPts as shown below:

RIEGL VZ600i 598 SCAN POS 4.15Bpts..jpg

We are attending the Pacific SAME forum on September 13, so we needed something to show the Generals and Admirals during breakfast and lunch. It was decided to use Veesus' Zappcha to visualize some of our scanning work. Many people have not seen a 4.15 Billion Pt color lidar scan without people and cars on an iPhone with a gross measuring capability.

CAPITOL-to-THOMAS-SQ.jpg

H1-WARD-to-VINEYARD.jpg

[dhirota]

I should have added these few pieces of information that were forgotten in the hurry to generate the composite RECAP PRO.

We used the exported scan position KMZ of each VZ600i project to generate the location of the scanned projects which were dragged into GoogleEarth Pro. The KMZs were easy to generate since we were using RTK for registering each scan position.

6-SCAN POSITION PROJECTS KMZ.jpg

To give you a scale of the field time walking along the street side walks, executing a "stop-and-go" data capture, I took the date-time stamp of the beginning and last scan to compute the project field time. This did not include the travel time from our office to the start location which varied.

The total VZ600i scan time was 666.627 minutes for the 596 scan positions, including scanning (15million points average-maximum range 300 meters), 36 internal images (12Mpixel) captured, and movement from to each new scan position.

This averages out to 67 seconds per scan position. Using our parameters, the actual timed start up to end data store time is 31 seconds.

[dhirota]

I forgot to mention that the original RECAP file is 83.6GB was reduced to 22.9GB by ZAPPCHA.

[dhirota]

Here is a view of 300 high level people attending the SAME PACIFIC forum, including the Director of Programs for one military branch, who liked the ZAPPCHA very much, since he had not seen something like this before. I used the ZAPPCHA images to show approximately 20 people most of them had not seen point clouds of any size on a iPhone. The people from outside of Hawaii are interested in our company being included on their team for civil engineering, land surveying and 3D scanning from as far away as a US Tribal Indian nation and a company from Dubai today.

SAME PACIFIC FORUM 01.jpg

SAME PACIFIC FORUM 02.jpg

They are planning to spend over US$2Billion in the Pacific Region centered around Hawaii as the headquarters for Army, Navy, Marine, AirForce, and CoastGuard, including work in Australia and NewZealand. One of the important projects they indicated to scan are the rock fall areas that are covered with vegetation, which should be a good project application.

[dhirota]

Getting back to Lieven's question about the brightness of the illumination captured.

https://www.laserscanningforum.com/foru ... 34#p107634

....
Today, Joshua Steinert was scanning Sam's Club Pearl City, Honolulu with his carbon fiber tripod + LGS BLK360 Gen2. We (Ray Sandla and I) were able to join him to complete 36 scan positions (total time=60 minutes) with our Riegl VZ600i on the inside of SAM's. This was another experiment, since we have not completed a large indoor commercial space with hundreds of shoppers walking in the space.
.....

In the picture with Joshua it seems there is enough lighting although the pointcloud seems a bit darkish. Any reason for that Dennis?

......

We are not quite sure what happened with the exposure of the images, but in case a problem like this occurs after you have the left the project and cannot gracefully return, then you need to figure out how to fix it.

HAS ANYONE WITH A VZ600i TRIED TO FIX THIS PROBLEM?

One of the major problems with internal cameras is the number of images that need to be corrected. In a simple scan project like this with 36 scan positions, you need to fix 36sp X 12 view locations/sp X 3 cameras/view location=1,296 images.

You need a batch process to automatically extract the image, correct the exposure and white balance, and place it back into the correct location without flipping the orientation. Here is our first attempt, which is not working 100%.

230906_083217_02_01_c3-ORIG.jpg

230906_083217_02_01_c3-IMPROVED.jpg

[landmeterbeuckx]

Can't you run a batch process on it therough Irfanview or other? I presume these pictures are still in the same folder at each scan (SCANPOSIMAGES) as with a vz400i? Or is the structure of files reorganized with this 600i?

[dhirota]

....
It's always taken us a lot of effort and tuning to get great color results out of the D850 on the VZ400i. I will share some results of this in the post. We were only able to do so when moving away from the recommended settings and using the auto exposure, auto iso metering per shot on the camera with some other custom settings. You would run into issues when using the 'capture images while scanning' mode if one section of the area you're scanning in (for example scanning near sunrise/sunset with the sun is low on your side) is notably darker or brighter (and thus requires longer exposure time) than possible for the camera to capture while the scanner is in motion. It would have been wonderful if that mode actually paused the horizontal traverse and laser capture momentarily for whatever image capture duration was required as determined by the D850 so that no section of the scan had inconsistent exposure and the motion blur could be eliminated. This is much more noticeable in industrial or commercial environments than it is on civil or mining scopes because the text/markings and sharp lines/edges jump out at you.

It's particularly interesting on the 600i's triple camera internal setup. The FARO units we have will give consistently useful color every day (with the exception of super bright areas which we use the LIHDR processing for) but requires some time to do exposure metering in order to accomplish this. I can't remember if the 600i offered the ability to set the camera properties for each sensor (as they'll each be aiming in a different direction with distinctly different lighting for each one), but I don't remember there being time allocated to automated exposure metering at the time I did testing.
...
That said, the image alignment, field of view, parallax, and resolution was really quite good. I don't know that I would even want to bother with adding the external camera to the 600i unless I was doing a project where very long range image resolution and a corresponding camera/lens package would be required. The impact for us in not having to run a fairly complex field and office workflow to compensate for the alignment issues on the VZ400i/D850 system would be amazing in both reducing the difficulty in operating the system for newer users, and for delivering a much more consistent backend effort level.

All of the items mentioned by Tommy for the VZ400i/Nikon 850 are similar to what is happening to the VZ600i/internal cameras with the auto exposure, auto iso metering for internal projects with the VZ600i.

When you evaluate an approach to improve the quality of color image capture by a scanning system, part of the problem is the communication of the vendor to help you figure out the configuration to produce the optimum visual product. We are still discussing this with Riegl support, especially for interior projects.

The approach of correcting an existing set of VZ600i images provides a significant set of challenges (1) Do you improve the raw images captured by the internal cameras; or (2) Improve the images to be processed by RiSCan Pro to color the point clouds. Shown below are the files displayed for a single scan position.

SCANNED IMAGE FILES.jpg

RiCAN PRO IMAGE FILES.jpg

We decided to try an improvement of the raw images

https://www.laserscanningforum.com/foru ... 54#p107754

The example shown was changed using GIMP in batch mode automatic exposure improvement. As part of this improvement, the meta data contained in the EXIF file format needs to be checked for conformity. The visualization looks fine, but importing the files, RiSCAN PRO generated errors for each file. We are evaluating other software, since we discovered that GIMP does not transfer the entire meta data file, probably the reason for the errors.

Below are screenshots of the original SAMS images and a few improved images processed by RiSCAN PRO successfully as a test, but still need to be significantly exposure improved. We will now have to batch improve all 1,296 images and process them with RiSCAN PRO to achieve a work flow that we can use to fix exposure problems in the future.

ORIGINAL.jpg

IMPROVED-UNCLEANED.jpg

IMPROVED.jpg