Guidance Needed: Best Azure AI Services for Barcode Extraction, Image Quality Validation & Real-time Camera Guidance in Canvas PowerApps

Question

Guidance Needed: Best Azure AI Services for Barcode Extraction, Image Quality Validation & Real-time Camera Guidance in Canvas PowerApps

Sachin Soni 46

Hi Community,

I’m working on a PowerApps application that involves scanning barcodes, capturing images, and validating image quality before uploading to Azure Blob Storage. I want to understand that which Azure AI service best fit for my requirements.

Below are the Azure AI–specific needs of the solution:

Barcode Reading

Users will scan a 25‑digit barcode (Packing Slip / Delivery Control Slip).

**2. Image Quality Enhancement

Image Quality Check / Validation (Most Important)**

Before allowing submission, I need AI to detect:
Blurry images
Wrong angle or incomplete object
Poor lighting

Does Azure provide built-in services which can handle above three scenario effectively ? Looking for advice on above three points

Thanks in advance for any guidance!

Karnam Venkata Rajeswari 280 Reputation points Microsoft External Staff Moderator

2026-02-26T04:25:38.5866667+00:00

Hello Sachin Soni,

Did you get any chance to review the above response?

Do let me know if you have any further queries.

Thank you!
Sachin Soni 46 Reputation points

2026-02-26T06:18:34.61+00:00
Hi Karnam Venkata Rajeswari,

Thank you for the detailed explanation. I would like to share an update regarding our scope.

Image quality enhancement is no longer part of our requirements.

For the remaining use cases:

Barcode reading & image scanning → I will proceed with Azure AI Document Intelligence using the prebuilt‑barcode model.

Image quality checks (blur, lighting, angle, completeness) → I plan to use Azure AI–based image analysis, but I'm unsure which service is most appropriate.

Could you please advise which option is best suited for image quality validation?

Azure AI (Vision services in Azure AI Foundry)

Azure Custom Vision

Azure Computer Vision (Image Analysis)

Using Both Services Together in a Single Resource Group

I also want to confirm that both Document Intelligence and the chosen Vision service can run within the same Resource Group, and how they will operate together.

Your guidance will help me finalize the approach.

Thank you!
Karnam Venkata Rajeswari 280 Reputation points Microsoft External Staff Moderator

2026-02-26T08:21:45.44+00:00

Hello Sachin Soni,

Glad to hear that my earlier response was helpful!

On further query regarding finalizing the appropriate service to perform image quality checks (blur, lighting, angle, completeness) , instead of going ahead with image analysis, please consider choosing Azure AI Custom Vision for the following reasons

1.      Image Analysis does not document a direct blur score , exposure score or skew score as needed.

2.      With Azure AI Custom Vision , a classifier model can be built to train and label images as Pass / Fail (or multi-label like blurry, too dark, cut off, too skewed) which suits the use case. This is the most “workable” way to validate blur + lighting + angle + completeness together.

3.      Please train the model using a labelled dataset of good vs bad images for each quality metric

4.      As an additional step , Azure AI Custom Vision can be used primary validation check point with optional Azure AI Vision Image Analysis for supporting signals. But please ensure to create two different resources - “Computer Vision” resource for image analysis and “Azure AI Vision” for Azure AI Custom Vision related tasks as both are two different services that hit different API’s. A mismatch can result in errors.

Azure AI Vision v/s Custom Vision –

·       Azure AI Vision provides metadata like tags/objects

·       Custom Vision evaluates quality labels

As asked if both document Intelligence and the chosen Vision service can run within the same Resource Group – the answer is Yes, we can create them under the same RG.

They will operate together as follows:

1.      A resource group is explicitly defined as a logical container for related resources for a solution, and resources in a resource group can be managed together.

2.      Azure services are decoupled resources which are independent that can be called with different endpoints/keys , though they exist in the same RG.

Please refer to the below resources to have a headstart

·       Create a Foundry resource - Foundry Tools | Microsoft Learn

·       Use the Azure portal and Azure Resource Manager to Manage Resource Groups - Azure Resource Manager | Microsoft Learn

·       What is Custom Vision? - Foundry Tools | Microsoft Learn

·       Azure AI Custom Vision | Microsoft Azure

Please let me know if you have any questions.

Thank you!
Karnam Venkata Rajeswari 280 Reputation points Microsoft External Staff Moderator

2026-02-28T17:46:45.04+00:00

Hello Sachin Soni,

We haven’t heard from you on the last response. So, just checking back to see if you have a resolution yet. In case if you have any resolution, please do share the same with the community as it can be helpful to others. Otherwise, please respond with more details and we will try to help.
Sachin Soni 46 Reputation points

2026-03-02T04:59:56.0133333+00:00

Hi,

Thank you for the follow‑up. Your guidance has been very helpful. I’m closing the loop on this one, and I appreciate the support from the community.

Thanks again!
Karnam Venkata Rajeswari 280 Reputation points Microsoft External Staff Moderator

2026-03-02T05:11:46.9+00:00

Hello Sachin Soni,

Glad to hear that the responses provided were helpful.

Since I’ve converted my earlier comment into an answer, could you please take a moment to mark it as Accepted? This helps others in the community with the same question find the solution more easily.

Thank you!

Answer accepted by question author

1 additional answer

Your answer

Karnam Venkata Rajeswari 280 Reputation points Microsoft External Staff Moderator

2026-02-26T04:25:38.5866667+00:00

Hello Sachin Soni,

Did you get any chance to review the above response?

Do let me know if you have any further queries.

Thank you!
Sachin Soni 46 Reputation points

2026-02-26T06:18:34.61+00:00

Hi Karnam Venkata Rajeswari,

Thank you for the detailed explanation. I would like to share an update regarding our scope.

Image quality enhancement is no longer part of our requirements.

For the remaining use cases:

Barcode reading & image scanning → I will proceed with Azure AI Document Intelligence using the prebuilt‑barcode model.

Image quality checks (blur, lighting, angle, completeness) → I plan to use Azure AI–based image analysis, but I'm unsure which service is most appropriate.

Could you please advise which option is best suited for image quality validation?

Azure AI (Vision services in Azure AI Foundry)

Azure Custom Vision

Azure Computer Vision (Image Analysis)

Using Both Services Together in a Single Resource Group

I also want to confirm that both Document Intelligence and the chosen Vision service can run within the same Resource Group, and how they will operate together.

Your guidance will help me finalize the approach.

Thank you!
Karnam Venkata Rajeswari 280 Reputation points Microsoft External Staff Moderator

2026-02-28T17:46:45.04+00:00

Hello Sachin Soni,

We haven’t heard from you on the last response. So, just checking back to see if you have a resolution yet. In case if you have any resolution, please do share the same with the community as it can be helpful to others. Otherwise, please respond with more details and we will try to help.
Sachin Soni 46 Reputation points

2026-03-02T04:59:56.0133333+00:00

Hi,

Thank you for the follow‑up. Your guidance has been very helpful. I’m closing the loop on this one, and I appreciate the support from the community.

Thanks again!
Karnam Venkata Rajeswari 280 Reputation points Microsoft External Staff Moderator

2026-03-02T05:11:46.9+00:00

Hello Sachin Soni,

Glad to hear that the responses provided were helpful.

Since I’ve converted my earlier comment into an answer, could you please take a moment to mark it as Accepted? This helps others in the community with the same question find the solution more easily.

Thank you!

Answer 1

Hello Sachin Soni ,

Welcome to Microsoft Q&A .Thank you for reaching out with detailed case description.

From the query,it is understood that you are looking for Azure AI and Custom Vision services for the following uses cases.

1. 25 digit Bar code Detection

2. Image Quality Enhancement

3. Image Quality check / Validation

1. Using Azure AI Document Intelligence for 25 Digit Bar code Detection

· Azure AI Document Intelligence has “prebuilt- barcode” model for barcode related usecases.

· Code 128 ,Code 39 ,EAN / UPC , PDF417 and additionally QR codes are supported by this model. As 25 digit barcodes fall under the code 128 category , AI document intelligence can be used.

· The below are the supported formats

For documents - PDF (.pdf) - Single‑page and multi‑page PDFs are supported

For images -JPEG / JPG (.jpg, .jpeg), PNG (.png) ,BMP (.bmp) ,TIFF / TIF (.tiff, .tif) ,HEIF (.heif)

Please refer the below documentations to get started with Azure AI document intelligence

Add-on capabilities - Document Intelligence - Foundry Tools | Microsoft Learn

Quickstart: Document Intelligence client libraries - Foundry Tools | Microsoft Learn

The suggested is an Azure‑native way to extract barcodes and text from documents in one pass.This is how it works:

1. Create an Azure AI Document Intelligence resource.

2. Use the prebuilt‑barcode model (or layout/other models with barcode extraction enabled) in your API request.

3. Works via REST API or SDKs (Python, C#, JavaScript, Java).

4. Upload images or scanned PDFs of your packing slips.

5. The API detects and decodes barcodes whether they’re Code 128, Code 39, PDF417, etc., and includes their values in the response

6. You’ll get a JSON output where the barcodes are listed with their decoded values.

2. Image Quality Enhancement

· While there is no single shot , ready to use , direct approach to enhance image quality , Azure’s AI vision can be used to handle this use case.

The below links can be further used for detailed reference.

Use AI Enrichment With Image and Text Processing - Azure Architecture Center | Microsoft Learn

Suggested approaches would be:

a) Preprocessing the images in Azure (before Blob) by using

Azure Blob Storage Trigger → Azure Function

Use a Function that triggers on image upload, enhances it, then stores a processed version into another container.

b) Using custom ML pipelines using Azure Machine Learning for AI‑based enhancement (super‑resolution, denoising, restoration).This can be done by training or use a pre‑trained model (e.g., ESRGAN, SwinIR, Restormer) and then deploy it as a real‑time inference endpoint using one of the below

· Azure Machine Learning Endpoints

· Azure Kubernetes Service

· Azure Container Apps

Then call that endpoint before upload → blob.

The following references might be helpful for enhancing image quality

Smart-croppedthumbnails - Azure Vision in Foundry Tools - Foundry Tools | Microsoft Learn

Analyzean image - Training | Microsoft Learn

https://learn.microsoft.com/en-us/training/modules/analyze-images/3-analyze-image?pivots=csharp

Imagecaptions - Image Analysis 4.0 - Foundry Tools | Microsoft Learn

3. Image Quality check / Validation

For blurry images

Azure AI Vision does not have a dedicated “blur score” API, but low confidence + poor detections + OCR failure are expected outcomes for blurry images.

Blurry images can be indirectly detected using:

· Image Analysis confidence scores

· Object detection confidence

· OCR (Read) confidence / missing text

· Image captions mentioning blur (example: “a blurry image of…”).The image quality (blur, lighting, resolution) directly impacts accuracy and confidence.

Please check this reference guide to follow through

https://learn.microsoft.com/en-us/azure/ai-foundry/responsible-ai/computer-vision/image-analysis-characteristics-and-limitations?view=foundry-classic

· For Wrong angle or incomplete object and poor lighting the metrics for Confidence loss (tags, captions, objects) and detection failure (objects missing / partial) can be used as validations and image quality checks.

An exclusive ML model can be built leveraging Azure AI Vision for the above purpose. It would either accept or reject the incomplete images based on lighting and blur.

The model outline would be as below:

· Use Azure AI Vision – Image Analysis to extract tags, captions, objects, and confidence scores from every uploaded image.

· Treat low average tag/caption confidence as a strong signal for blurry or poorly lit images.

· Use Object Detection to verify expected objects are detected (e.g., ID card, product, face).

· Flag incomplete objects when bounding boxes are too small or cropped relative to the image size.

· Detect wrong angle when expected objects are detected with very low confidence or partial boxes.

· Run Read OCR and flag images when text is missing or word confidence is low (common in blur/low light).

· Combine Vision outputs into features (confidence, box size %, OCR success, object count).

· Train an Azure ML binary classification model (Accept vs Reject) using these features.

· Apply rule-based prechecks (hard failures) before ML scoring for faster rejection.

· Return a clear rejection reason (Blurry / Poor lighting / Wrong angle / Incomplete object) for feedback.

Please refer the below references for detailed extended reading.

https://azure.microsoft.com/en-us/products/ai-foundry/tools/vision?msockid=3cae6539807d64212ca57230811c6589

https://learn.microsoft.com/en-us/azure/ai-services/computer-vision/overview

https://learn.microsoft.com/en-us/azure/ai-services/computer-vision/

https://github.com/microsoftdocs/architecture-center/blob/main/docs/data-guide/ai-services/image-video-processing.md

https://github.com/microsoft/computervision-recipes

Please let me know if you have any questions.

Thank you!

Answer 2

For your PowerApps application involving barcode scanning, image quality validation, and real-time camera guidance, Azure provides several AI services that can effectively address your requirements:

Barcode Reading: You can utilize the Azure Computer Vision API, which includes capabilities for reading barcodes from images. This service can help you accurately scan the 25-digit barcodes you mentioned.
Image Quality Enhancement: While Azure does not have a specific service solely for image quality enhancement, you can leverage the Custom Vision Service to train a model that can identify and improve image quality based on your specific criteria. This service allows you to create custom classifiers that can detect various aspects of image quality.
Image Quality Check / Validation: For detecting blurry images, wrong angles, incomplete objects, and poor lighting, you can again utilize the Custom Vision Service. By training a model with labeled images that represent good and poor quality, you can create a solution that validates images before they are uploaded to Azure Blob Storage. Additionally, the Computer Vision API can assist in analyzing images for certain quality metrics, although it may not cover all your specific needs directly.

In summary, combining the Azure Computer Vision API for barcode reading and the Custom Vision Service for image quality checks and enhancements will provide a robust solution for your application.

References:

Share via

Guidance Needed: Best Azure AI Services for Barcode Extraction, Image Quality Validation & Real-time Camera Guidance in Canvas PowerApps

1 additional answer

Your answer