Store the Object detection Image frame result from IOT device to cloud

Vinay Solanki 1 Reputation point
2022-08-26T01:28:45.827+00:00

Hello,

I am looking the functionality where my object detection model detect the object from camera frame and that image with bounding box can be stored to cloud storage.

I can see routing for device telemetry but could not find anything to store image frame from IOT to cloud.

Could you help me if such functionality exist?

Thanks in advance.

Azure Percept
Azure Percept
A comprehensive Azure platform with added security for creating edge artificial intelligence solutions.
70 questions
{count} votes

1 answer

Sort by: Most helpful
  1. Sander van de Velde 28,161 Reputation points MVP
    2022-08-26T17:01:04.66+00:00

    Hello @Vinay Solanki ,

    you want to store both (a part of) the image and some related telemetry in the cloud.

    I'm neither sure how you connect to the cloud nor do I know what programming language you use but I expect you are able to use the Azure IoT Device SDK to connect to an IoT Hub.

    There are several approaches possible, it mostly depends on the size of the image and the platform you use.

    The maximum message size of a message sent to the IoT Hub is 256KB. If your image can be part of the message (eg. as byte array/base64 encoded), just send the message and do something with it in the cloud. Keep in mind that messages are devices in 4KB chunks so keep an eye on your daily message consumption.

    You can also use the IoT Hub file upload functionality, supported by most SDKs. Files are then stored in a related Storage Account container so you either trigger on file uploads or you only put a unique reference of that file in the message. The IoT Hub only counts the number of files stored without looking at the actual size.

    If you are using Azure IoT Edge, the file upload functionality is not offered. Instead, check out the Blob storage module that offers a file sync with the cloud.

    Extra, storing images from IoT devices in the cloud is not always recommended. It's slow compared to telemetry, takes a lot of bandwidth, and can have GDPR consequences if eg. people are shown on the images. Rethink your solution and decide if this is really what you need.

    ----------

    If the response helped, do "Accept Answer". If it doesn't work, please let us know the progress. By doing so, all community members who have a similar issue will benefit. Your contribution is highly appreciated.

    1 person found this answer helpful.
    0 comments No comments