Save Form Processing Image to SharePoint list as an Image

(0) Share

Report

Posted on by Smitsky34

I have an AI Builder Model setup to read images for odometers.

I've created a PowerApp and included the control for the FormProcessor and the AI portion of it is working fine.

After the image is processed and it reads the miles I then want to insert the results into a SharePoint List.

My challenge is accessing the image in the FormProcessor Control to then Patch to the SharePoint list. I've successfully patched with and Image2 control but I don't want the user to have to select the same image twice (Add picture, Analyze in FormProcessor). I need to either, update the FormProcessor Control with the image from the Image Control, or visa versa so I can then run my Patch statement.

The goal, select the image once, analyze, set the miles read, patch and save the actual image and the miles.

Here's my patch statement:

Patch('Van Checkout - Checkin', Defaults('Van Checkout - Checkin'),
{
Title: Gallery1.Selected.Title,
Van: {'@odata.type':"#Microsoft.Azure.Connectors.SharePoint.SPListExpandedReference",
Id:Gallery1.Selected.ID,
Value:Gallery1.Selected.Title},
'Start Odometer Image': Image2.Image,
'Start Odometer Number': Value(OdometerReading.Text)
});
Notify("The data has been saved to the list");

Categories:

AI Builder

I have the same question (0)

All responses (3)

Answers (0)

CU22081303-0 150 on at

Like (0)

Report
Hi, it looks like this question is a few week old but I found a solution so I'll put it here in case it helps you or anyone else looking for a similar fix.

When we get the "Original Image" from the text recognition component, it returns a long Base64 string which we cannot save to the record's Image field because we can't put text in an image.
We can convert that base64 to an image and use that for the Image control on our page (in your example, this control is called Image2).

Here is the code:
Substitute(TextRecognizer1.OriginalImage,"data&colon;image/png;base64,","")

My repurposed test screen looks like this:

My Button control has this code:
Patch(Cases, First(Cases), {Description: Last(FirstN(Gallery3.AllItems, 5)).Text, ImageField: Image3.Image });

And clicking it causes one of my cases to get a new description, and most importantly, and image!

There is probably a nice, clean way to make the patch function do the conversion right then and there without an additional image component, but if it were me, I would just put this image control behind the Image Recognition component so the users don't have to know about it.

Hope this helps someone, thanks for the interesting problem to explore! Great job on the odometer reader too, there's a lot of extra numbers in those images to avoid. My license plate reader still reads a ton of extra text, I haven't narrowed it down to just the plate number yet.

Was this reply helpful? Yes No
Anonymous231456 2 on at

Like (0)

Report

Hi Smitsky, can you please share the steps followed for creation of the Odometer AI Reading Model. It will be very helpful.

Was this reply helpful? Yes No
John121A930 4 on at

Like (0)

Report

You can use this Free Visa Online Tool for this purpose https://visaphoto.co

Was this reply helpful? Yes No