Ciclops Scanners - any experience?

As the title says I'm interested to hear any feedback from people who have used these scanners.

I just assembled ours but can not get it to work just yet. I posted on the Facebook page and was given some tips to try.
Watching some videos I see there is a second card added to the Arduino that ours don't have. Going to order it next Friday.
Tried the wizard calibration and the lasers fail. One of the comments I got was the wizard is had to pass.
Thinking of building the raspberry pi scanner too.

If you seach for 'cowtech scanner' you will find a company in the states that did a kickstarter for a scanner based on the Ciclop, but they have a good selection of support documents that would be worth reading through.

If you could give me some rough measurements that would be awesome.

Centre of the camera mount to the furthest part of the threaded bars that support the lasers.
From the endpoint measured in the last measure directly to the end point on the other arm.
Height from desk to the centre line of the lasers.
Length from the rear (electronics side) of the camera mount to the far side of the platform base.

Was going to print the stock parts but the builds are long, thinking about basing a copy on v-slot. Will of course openly share work.

Stand down! ;-)

I've pieced together the CAD from the GITHUB in FreeCAD and so can start to take measurements directly from the CAD. The angle on the laser arms seem about 30 degrees from the horizontal. Other values should be fairly easy to get.

I've spotted (but not tried - no scanner yet) a version of Horus on github that has been ported to python3. While you may not be interested in tweaking Horus if you can get it working from the python files you'll get free performance upgrades with each of the updates to the underlying modules:

[github.com]



Edit: As you'd expect the aim of the lasers targets an intersection at the central axis of the platform. That's enough for me to order extrusion now. Getting closer.

Edited 2 time(s). Last edit at 04/23/2019 03:22AM by WesBrooks.

I know this topic is rather old.

But I own a ciclops and can tell that it is not worth building and/or buying one.

Bether go this way:
[blog.prusaprinters.org]

Edited 1 time(s). Last edit at 10/14/2019 10:56AM by Dancer.

i have a cyclops and find that using 50ma lasers helps with scan, using more than one laser on each side can help but it is hard to align to same line. as long as the light is not at the exact same angle it does not cause interference, and even if it is the same angle the distance of the interference would be smaller than visible by the eye and the line would average out (brighter and dimmer) even before the resolution of a human eye.

the reason for more than 1 laser per side is to manage the angles of refraction. using 1 laser the limit is about 20deg or so that the camera can see so rounded and curved surfaces are hard to see even with 50ma laser that are more than 2 inches tall.

the length the laser light travels is not an issue either as the side angle (right and left) distance to the camera generates the distance calc.

this being said, if i knew what i did now, i would have gone another route. the best experience i have had with scanners was a picza probe piezo scanner years ago. it was accurate to 0.02mm took hours, but it was so reliable. it just became too old to use with modern software. ( too much headache to keep working with each update of windows and no mac support)

Edited 1 time(s). Last edit at 10/15/2019 08:46PM by jamesdanielv.

I've got it running after forcing through a load of the code to python 3 (started by someone else), compiling the latest libraries for the on-computer software, and basing it on v-slot extrusion.

It's still a fussy thing and I've not been able to spend any time on it recently as my printers have been too much of a headache. I did decide that a light box would be essential to be able to calibrate, but then shut off the lights for the actual scan.

The next change for me would be a closer focus, higher resolution camera. I would like to look at another kinematic system for it, probably a basic Cartesian. I also wonder about using the results from this and a photo based system to compliment each other.

the resolution detail normally is higher on a laser scanner than a photo collections based 3d calculated scanner. this is because a direct reference to distance is calculated from shape of laser. the shape is based on difference and this is calculated as if it is a hypotenuse of a triangle.
probably the best type of 3d scanner is light scatter where different sets of lines are projected on the image and the differences in width and shape help calculated the angle.

with images methods it uses shadow, light source, and stitching to create a 3d representation. it seems to work better on large stationary objects outside. but in my experience detail is lacking in mesh, but is good in texture quality.

in my opinion there currently is no perfect method for scanning 3d objects, the most stable and reliable experience is from contact point probe. but this has no texture data.
so each method will require several scans, at different angles to get enough of a point cloud to fill in a mesh with a detailed surface.

the usability of the cyclops scanner is limited by how bright the lasers are, the surface of the object being scanned, and the angle of the object surface. if scanning sculptures it should be reasonably ok after several scans are combined
Also using white powder like baby powder on items and covering up transparent surfaces helps . maybe a spray such as spray Marking Chalk White.


for the cyclops scanner. mine was 3d printed.
also i've noticed when not using ambient light it is better to have a lamp that has a wide angle and have it more than 3 feet above and behind camera. and to make sure that the calibration base is red, and that the checked outline not be shinny, and be laser printed, or in my case a 1/4 inch thick black foam board, with inside white, and black square laser cut and peeled out carefully to keep square shape and kerf to have size and shape as perfect squares. if nothing else then use bright white cotton paper and a laser printer to generate the checkered pattern.
also instructions state that board should go to edge of stage, but i find i have an easier time with calibration if i have it centered, but an inch more inwards.


this scanner is a pain, but when it works it does ok. at least when scanning with a single laser at a time.

Edited 4 time(s). Last edit at 10/16/2019 07:32PM by jamesdanielv.

... my accuracy numbers for camera scanning with Laser Line or Scattered Light are like 0,5% of the image width for LL or 0,2% for SL -- as ciclops uses LL, this should be the target to reach ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]