Micro QR Detection about boofcv

[lessthanoptimal]

Woo a real world example of Micro QR! I was struggling to find some to test this on. Anyways, you're right that just inverting the image is the easiest way to solve that problem. To modify the code to detect regular and inverted markers efficiently would require a completely different approach. Missing the last two digits might be an actual bug. Logic for decoding logic is inside of MicroQrCodeDecoderBits <https://github.com/lessthanoptimal/BoofCV/blob/af6405afc5931c978bfedc0a2e55fb3260c3207e/main/boofcv-recognition/src/main/java/boofcv/alg/fiducial/microqr/MicroQrCodeDecoderBits.java>. You will need to find the specification document to hunt for the bug. I should have time to look at this in the next two weeks or so. I'm guessing there's a type-o for the number of data bits. Cheers, - Peter

…

On Fri, Jan 6, 2023 at 12:20 PM darren1713 ***@***.***> wrote: Issue: Recognition of inverse / negative Micro QR, and decoding issue Great library! We started using Micro QR for production of small devices, and have now realized that readers, and libraries, are very scarce for Micro QR. Compounding this complexity is our laser marking process on black plastic, that leaves white markings, created an inverse Micro QR code is even less supported. Below is an example of a Micro QR code, which contains the following: 001170081008300534063156620 [image: image] <https://user-images.githubusercontent.com/1369000/211092596-560f2d22-cf8a-4878-931a-c0521011dd6f.png> We can take a picture of the Micro QR, then invert the colors, then use Boof CV to scan it, but it's missing the last two digits, and only decodes: 0011700810083005340631566 [image: image] <https://user-images.githubusercontent.com/1369000/211092635-bde40972-bcfc-4443-9fac-6fb8aa378158.png> We are an engineering team/programmers, although we have never worked with the Boof CV repo. If you have some thoughts on this topic would be a great start. The first issue with the negative QR code might be something as simple as applying a negative filter to the image processing, but the second issue missing the last couple digits of the Micro QR code is a deeper issue. Thanks! — Reply to this email directly, view it on GitHub <#379>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAFUOV4GU5BHLA45Y5HOVHTWRB5BPANCNFSM6AAAAAATTNS4NQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

-- "Now, now my good man, this is no time for making enemies." — Voltaire (1694-1778), on his deathbed in response to a priest asking that he renounce Satan.

[jefferypalmer]

Thanks Peter!

Since you have a good grasp on this, we will gladly sponsor your project, time, etc, for working on some of these fixes. Do you have a fund this project link by chance? Or drop your email/contact and we can chat.

Cheers,

Jeff

[lessthanoptimal]

Hi Jeff,

I would look into this for free, but sponsoring it will definitely be appreciated and bump it in the priority list. Haven't set up a fund for the project and now i'm diverting engineering requests to my startup at https://www.ninox360.com/ You can fill out the form there or use any of the emails there which I'll see.

Thanks!

• Peter