Scanning Technologies, Pros/Cons

You don't know yet, which scanning system to buy or build or if 3d scanning is for you at all? Ask here!
Post Reply
Grampa
Posts: 4
Joined: 25 Sep 2018, 06:36

Scanning Technologies, Pros/Cons

Post by Grampa »

Hi,

For some time I have been modeling in cad with point clouds and recently our company decided to take on the challenge of scanning steel forming rolls. They rented a 3D scanner which uses structured light as it's scanning technique with a tolerance of 0.1mm.
https://www.artec3d.com/portable-3d-sca ... tec-spider

However now when we are looking at the results from the scan and compare them to my hand measurements they seem a bit off, like about everything between "0.2 - 0,9". A diameter which I measured by hand should be ø120mm could become ø119,5mm and a normal dimension of 45,5 could be 45,65. They even look a little "Bent" like a banana where a surface could diff between +0,15 to -0,29 mm. The worst part is that we are unsure which scanning technology is the best and precise.

Could you name some pros and cons with the different scanning technologies and explain why you'd prefer it?
User avatar
Micr0
Posts: 586
Joined: 15 Nov 2016, 15:20
Location: New York City

Re: Scanning Technologies, Pros/Cons

Post by Micr0 »

Grampa wrote: 25 Sep 2018, 06:49 Hi,

For some time I have been modeling in cad with point clouds and recently our company decided to take on the challenge of scanning steel forming rolls. They rented a 3D scanner which uses structured light as it's scanning technique with a tolerance of 0.1mm.
https://www.artec3d.com/portable-3d-sca ... tec-spider

However now when we are looking at the results from the scan and compare them to my hand measurements they seem a bit off, like about everything between "0.2 - 0,9". A diameter which I measured by hand should be ø120mm could become ø119,5mm and a normal dimension of 45,5 could be 45,65. They even look a little "Bent" like a banana where a surface could diff between +0,15 to -0,29 mm. The worst part is that we are unsure which scanning technology is the best and precise.

Could you name some pros and cons with the different scanning technologies and explain why you'd prefer it?
That kind of deviation doesn't surprise me. That said there is a lot more to getting good scans than pointing a camera at something and yelling "say cheese". Prep can go a long way to improving accuracy. I have an Artec Eva and Love both the scanner and Artec studio 12. However I am very aware of it's limitations. With David I have refined my process to be able to fairly easily/consistently get ~.07mm and better if I'm really careful (this is a 150mm calibration scale /work envelop). However, and this held true for the Eva too, it took practice.


What are the dimensions of the rollers you are scanning?

What is to accuracy you REALLY need?

What is the surface finish etc?
µ
Grampa
Posts: 4
Joined: 25 Sep 2018, 06:36

Re: Scanning Technologies, Pros/Cons

Post by Grampa »

Hi Micr0,

Well, the rollers we scanned were about all between Ø190 - Ø80 mm. The rollers are split to make manufacturing easier and if they are combined which we have scanned them the dimension comes in even dimensions of 125, 105, 85, 65, 45 mm.

When it comes to accuracy, these rollers are forming 0.4 mm sheet metal. The accuracy would land around at least 0.02 mm. We did a calibration with the people who rented us the scanner as we didn't get any results to start with. Problem is that if the rollers are 0.1 mm to small, they practically are spent.

When i think of it, the person who scanned the rollers used hair spray and then used his nail to make stripes on the surface to make areas the scanner would recognize. How do you do when you scan reflective surfaces?

Looked at one of the drawings we've made and the desired surface roughness is about Ra 0,6.
User avatar
Micr0
Posts: 586
Joined: 15 Nov 2016, 15:20
Location: New York City

Re: Scanning Technologies, Pros/Cons

Post by Micr0 »

Grampa wrote: 27 Sep 2018, 06:45 Hi Micr0,

Well, the rollers we scanned were about all between Ø190 - Ø80 mm. The rollers are split to make manufacturing easier and if they are combined which we have scanned them the dimension comes in even dimensions of 125, 105, 85, 65, 45 mm.

When it comes to accuracy, these rollers are forming 0.4 mm sheet metal. The accuracy would land around at least 0.02 mm. We did a calibration with the people who rented us the scanner as we didn't get any results to start with. Problem is that if the rollers are 0.1 mm to small, they practically are spent.

When i think of it, the person who scanned the rollers used hair spray and then used his nail to make stripes on the surface to make areas the scanner would recognize. How do you do when you scan reflective surfaces?

Looked at one of the drawings we've made and the desired surface roughness is about Ra 0,6.
Off the top of my head I don't think the Spider is the best solution for you considering the size and the accuracy you need. It would seem to be within the specs of the scanner, and it should have given you a relatively accurate scan. You may have had to scale or clean up a bit in Studio but it should have given you something usable.

Spraying helps (I use Magnaflux developer personally) and since they did that I'd rule that out. I don't understand the scratching stripes. I mean I understand why he did it but I cant see why it would be necessary.
µ
Grampa
Posts: 4
Joined: 25 Sep 2018, 06:36

Re: Scanning Technologies, Pros/Cons

Post by Grampa »

Hi Micr0,

We are gonna make a test on Thursday where we are gonna first hand measure a couple rollers, then scan them using the artec spider again and then visit a company who got a laser scanner with camera stand, unfortunately don't remember the brand but they use it to scan sheet metal details.

It will certainly interesting to see the results.

About the hairspray. I spoke with my coworker and he said it was to make scan points for the camera to be able to locate itself and the object but I think it could have been done differently and better. Have you tried hairspray and compared to your scanning spray, I told them it might be that the spray doesn't do the job.

I attached one of the scans that turned out really bad. This was made with the said scanner and yes, hair spray.
Attachments
Scanning.PNG
User avatar
Micr0
Posts: 586
Joined: 15 Nov 2016, 15:20
Location: New York City

Re: Scanning Technologies, Pros/Cons

Post by Micr0 »

Grampa wrote: 28 Sep 2018, 06:53 Hi Micr0,

We are gonna make a test on Thursday where we are gonna first hand measure a couple rollers, then scan them using the artec spider again and then visit a company who got a laser scanner with camera stand, unfortunately don't remember the brand but they use it to scan sheet metal details.

It will certainly interesting to see the results.

About the hairspray. I spoke with my coworker and he said it was to make scan points for the camera to be able to locate itself and the object but I think it could have been done differently and better. Have you tried hairspray and compared to your scanning spray, I told them it might be that the spray doesn't do the job.

I attached one of the scans that turned out really bad. This was made with the said scanner and yes, hair spray.
That does look bad. I know the Spider is capable of much better than that. Judging from the image of what you are scanning I would say the spider should give you what you want. The spider and Artec studio don't need target points to scan properly. The spray should only be to give the projected pattern (The Spider projects an SL Pattern) a consistent non reflective surface to illuminate. The only time I use targets with the Artec is when there are large areas of little detail or curvature change. Or I have to scan very large objects in multiple files that I then have to assemble manually. I think once you get the Spider and scanning sorted out it should do what you want.

BTW where are you located?
µ
Grampa
Posts: 4
Joined: 25 Sep 2018, 06:36

Re: Scanning Technologies, Pros/Cons

Post by Grampa »

Micr0 wrote: 28 Sep 2018, 16:36
Grampa wrote: 28 Sep 2018, 06:53 Hi Micr0,

We are gonna make a test on Thursday where we are gonna first hand measure a couple rollers, then scan them using the artec spider again and then visit a company who got a laser scanner with camera stand, unfortunately don't remember the brand but they use it to scan sheet metal details.

It will certainly interesting to see the results.

About the hairspray. I spoke with my coworker and he said it was to make scan points for the camera to be able to locate itself and the object but I think it could have been done differently and better. Have you tried hairspray and compared to your scanning spray, I told them it might be that the spray doesn't do the job.

I attached one of the scans that turned out really bad. This was made with the said scanner and yes, hair spray.
That does look bad. I know the Spider is capable of much better than that. Judging from the image of what you are scanning I would say the spider should give you what you want. The spider and Artec studio don't need target points to scan properly. The spray should only be to give the projected pattern (The Spider projects an SL Pattern) a consistent non reflective surface to illuminate. The only time I use targets with the Artec is when there are large areas of little detail or curvature change. Or I have to scan very large objects in multiple files that I then have to assemble manually. I think once you get the Spider and scanning sorted out it should do what you want.

BTW where are you located?
Hi Micr0,

Well, that one was one the worse versions. I did a model after it and compared it but it gave me a strange result. We later received a drawing of the actual component and inserted into the analysis program (GOM Inspect 2018) and the result showed that the part was larger then the scan. This is what trigger the situation. Another problem we got now is that the rollers are being used so re-scanning them isn't an option as they can become worn.

We are located in Sweden btw. :3

Not to worry you too much, i attached a picture of the better scans.

When you scan with your scanner, does it happen that it loses its location, like we did with the spider. When it lost it's location, it starts alarming and sometimes we could resume the scan by moving back, other times we had to redo the entire scan as it started to build up a new model in the same scan.
Attachments
Scann 2.PNG
Post Reply