OCR Recognition Return Row

Hi Guys, I asked this before but didn’t really get to the bottom of what I needed, I think there were a couple of bugs in the new OCR module.

So what I am hoping for is to search a record using OCR.
We are a vehicle repairer and we turn out some 100 + vehicles per week. I have views already set up to show vehicles in progres, due in, etc etc.

It would be fantastic to be able to scan a number (Registration) plate and return the record (row) which matches the image taken.

Has anyone done this before, is it available as an example APP? Any help would make a massive difference in our day to day operation.

Kind Regards

0 17 636
  • UX
17 REPLIES 17

Hi Aleksi, yes thanks for that. The app will extract the text and add it to a row.
But how do we use the OCR as a search tool?

Kind Regards

You could create a small Dashboard view with a quick editable field and then use the result for searching with the slice.

Can you explain this better. I would like to duplicate the same functionality. I would like to use OCR in a warehouse and scan a label to see if the other boxes with the same label have already been delivered.

It’s quite an old topic now but never got around to doing this. Would it be possible for the Appsheet guys to put togethert a simple App showing how this could be done via dashboard or something?

To scan an object ( vehicle number plate ) and return the row(s) which match the scan.
@Scott_Haaland @Aleksi

Hi @Dave_Willett

I did something a few weeks ago:

Hi @Aurelien , it looks like it could be what I’m after. I’ll play it a few times to understand it.
Thank you for the assistance.

Kind Regards

Glad to hear it.

Basically you create a real column to write down the value read by the OCR.
This column is of type ref, related to the table you want
And then, get the information you need from this with a [refKey].[DataToGet] formula.

Have a nice day, feel free to ask if you need more information.

Really appreciate that.
I do have one issue though, I have only one output field and the OCR model won’t train.

can you give it samples so that you can train it?

Got it. I only needed to use OCRTEXT:

Nice !
That works for pretty simple images though, and vehicule plate are perfect for it indeed.
In my case I had to deal with various configurations (shape, angle…), so I needed to train it with some pictures first, which include to save pictures for a while, and delete everything once my inventory job was done

I’m glad to hear that does the job for you !

It looks like its working…

Thanks again

Terrific !
You’re welcome

Hi @Aurelien
I’m getting errors. I reproduced your video ( I Think) and ref to qryInprogressAppsheet table.
Any ideas why?

Hi @Dave_Willett

Sorry for the time to answer.
In my opinion, at reading the error message you get, you are calling a column that does not exist in the ref table.
You are currently looking for a column called “RegRead” in the Registration table that the OCR column refers to.
Am I correct ? Can you provide a snapshot of the columns that are in the Ref source table ?

i got to the problem @Aurelien , it was the ref table I didn’t access the correct field. All sorted now.
Thanks again for the assistance.

Kind Regards

Top Labels in this Space