I'm NOT Stupid

[dogwitch]

34 minutes ago, Needfuldoer said:

 

 

my eyes.....

i hate you so much right now...

[LTTMobileJake]

3 hours ago, 2023 said:

You are all old boomer living in the past. /s

;( I was born in '90 mate

[dogwitch]

3 hours ago, LTTMobileJake said:

;( I was born in '90 mate

... i could be your father age.... thanks for making me feel old!

[DeerDK]

On 2/28/2023 at 2:45 AM, Eddiee said:

Keyboard mapping thoughts.

 

How accurate is the laser keyboard mapping?

• Could the laser keyboard mapping spot uneven keys?
• Key size and spacing is interesting too. What keyboards have that MacBook Pro feel! j/k.
• Could it map how concave the keys are? The keyboard I have has keycaps that are a bit more concave than I would like. Considering how extreme it feels compared to my old board, knowing how concave keys are compared to other boards out there might be something I would want to be aware of in my future keyboard purchases.
• Could be used to compare 3rd party keycaps, and which you might be able to mix and match considering how similar the dimensions are?
• The keycaps I have also do not have cutouts for backlighting on the secondary punctuation keys, which I find a bit annoying. Noting little stuff like that would be useful.

These might be other interesting bits of information people may be interested in as well that I can't think of.

Interesting questions. 

I work with 3d mapping and surveying myself - although the equipment we use is for landscape and buildings, so we usually work with a tolerance within a few mm. 

In principle, if the equipment is capable and correctly calibrated, uneven keys should be mapable and with a high point density making a top down map of the keyboard with slight colour variation dependant on the z value should also be possible. They may have to tweak the curve though for a good and readable result 

[ImorallySourcedElectrons]

6 hours ago, dogwitch said:

... i could be your father age.... thanks for making me feel old!

Lies! I've been trained by the best movies Hollywood could produce, and from those I learned that old people should always complain about kids skateboarding on the sidewalk.

 

17 minutes ago, DeerDK said:

Interesting questions. 

I work with 3d mapping and surveying myself - although the equipment we use is for landscape and buildings, so we usually work with a tolerance within a few mm. 

In principle, if the equipment is capable and correctly calibrated, uneven keys should be mapable and with a high point density making a top down map of the keyboard with slight colour variation dependant on the z value should also be possible. They may have to tweak the curve though for a good and readable result 

While I'm not familiar with the system they're using, I've worked with CMMs with laser profilers that will happily do <10 µm resolution with a fairly fast scan. Even the (by now ancient) kinect cameras were already consistently doing 1 mm depth resolution on nearby things when configured correctly. But I agree, it all depends on how they handle it from a software point of view, though key segmentation from a height map might very much be possible with a fairly simple k-means approach, after which finding the individual "islands" that make up the key should be relatively trivial. The bigger issue would probably be garbage laptop keyboards.

 

But I would also expect that they allow some user tuning of the parameters the program uses, trying to automate everything completely is often not that efficient.

[DeerDK]

11 minutes ago, ImorallySourcedElectrons said:

...

 

While I'm not familiar with the system they're using, I've worked with CMMs with laser profilers that will happily do <10 µm resolution with a fairly fast scan. Even the (by now ancient) kinect cameras were already consistently doing 1 mm depth resolution on nearby things when configured correctly. But I agree, it all depends on how they handle it from a software point of view, though key segmentation from a height map might very much be possible with a fairly simple k-means approach, after which finding the individual "islands" that make up the key should be relatively trivial. The bigger issue would probably be garbage laptop keyboards.

 

But I would also expect that they allow some user tuning of the parameters the program uses, trying to automate everything completely is often not that efficient.

Another issue is surface and reflection. Dark or shiny is generally less ideal if Lidar tech is used, as well as "steep" angles but this is less of an issue due to the limited area we are talking about.

But its generally a field of my interest. My company is working with implementing a combination of video recording (photogrammetry) combined with GNSS (GPS) xyz data. We can get some pretty good results that the workers on a dig or construction site can record themselves.

[AndrewJacksonZA]

17 hours ago, ImorallySourcedElectrons said:

I think it's important to separate the graphs for videos, presentations, and articles from the ones that go on a website, they're very different use cases.

 

I agree. For videos, one can create templates that one can insert data into, export those as images, and then have the video team use them. I've done that export for certain use cases at my previous company via an SSIS job and then emailed those exports via another job.

 

I think your "paid subscription feature set" is is brilliant. For this use case, one can create different templates (or use the same ones) but allow user interaction with it on the website.

 

One can use the same data source from within Tableau Server to enable both use cases.

 

@Plouffe or team, should you pursue this avenue or one like it, I'm happy to engage with you. There are also PLENTY of resources available online. But remember, there are many, MANY ways to skin pretty much any cat.

[AndrewJacksonZA]

-

Edited March 2, 2023 by AndrewJacksonZA
Deleting. Apologies for the accidental double post.

[AndrewJacksonZA]

On 3/1/2023 at 2:44 AM, LTTMobileJake said:

've spent over a decade working with engineering teams 

But yet you look so young, child.

[ImorallySourcedElectrons]

6 hours ago, DeerDK said:

Another issue is surface and reflection. Dark or shiny is generally less ideal if Lidar tech is used, as well as "steep" angles but this is less of an issue due to the limited area we are talking about.

Yeah, it's definitely not a one-size fits all solution like many folks make it out to be. But one advantage is that some scanners these days operate on multiple wavelengths (I've even seen a few "white light" ones), meaning they stand a reasonable chance of working out of the box without painting the surface if there's nothing transparent or highly reflective. But I'd imagine it'd still give issues for the ceramic and metal key caps that are starting to appear left and right. What we did for metal parts on the CMM was dust on some talcum powder, which was usually enough to get it to work reasonably well, but that might cause issues with the key switch mechanism. And angles are indeed an issue! Usually you had to try to keep the probe parallel to the surface, so I wonder how well this systems handles it.

 

Another one you can run into when scanning polymers is that some polymers are slightly transparent for the scanner, so you can sometimes get some signal return from what's behind the polymer part as well. Though sometimes that can be a happy accident as well!

 

6 hours ago, DeerDK said:

My company is working with implementing a combination of video recording (photogrammetry) combined with GNSS (GPS) xyz data. We can get some pretty good results that the workers on a dig or construction site can record themselves.

You are a braver person than me, pose estimation gives me a headache once you have to start calculating camera parameters from questionable smartphone data, etc.

[DeerDK]

22 minutes ago, ImorallySourcedElectrons said:

.... 

You are a braver person than me, pose estimation gives me a headache once you have to start calculating camera parameters from questionable smartphone data, etc.

Im not directly involved in the calculations. Although I did learn about it at University. We have some people working on the development. The process is mostly automated. I work with registration of the data, so you know, identifying pipes and such in the point cloud, and the conversion to GIS and CAD. We also do traditional 3d scanning and terrestrial surveying. 

[Forky the Forklift]

On 2/27/2023 at 11:57 PM, Plouffe said:

Things at the lab are going GREAT, and we can't wait to show you! Come take a tour and check out some cool new features like the metal 3D printer, submersion tank, and more before we move the writers over.

 

 

I would like to talk to HR regarding the inappropriate workplace behaviour of the CEO documented in this video (timecode: 17:52 - 17:56 )

 

Yours sincerely 

 

Forky the Forklift.

[Spotty]

The lab guy that has been in the last few lab update videos (Jake?) has done pretty good on camera. I can't wait for that all metal PC project with Alex!