---
title: "AI Vision IoT"
url: "https://maker.wiznet.io/josephsr/projects/ai-vision-iot/"
markdown_url: "https://maker.wiznet.io/josephsr/projects/ai-vision-iot/md"
type: "UCC: User Created Content"
author: "Peter Ma"
author_url: "https://www.youtube.com/watch?v=P4G2GHXkgEg"
editor: "WIZnet"
editor_url: "https://maker.wiznet.io/"
original_author: "Peter Ma"
original_url: "https://www.youtube.com/watch?v=P4G2GHXkgEg"
published: "2023-07-04"
language: "en"
tags: ["WIZ750SR"]
hardware: ["WIZnet WIZ750SR"]
likes: 0
views: 219
comments: 0
source: "WIZnet Makers (https://maker.wiznet.io/)"
---

# AI Vision IoT

> Object recognition with Helium IoT Kit and WIZ750SR

Original author: Peter Ma (source: https://www.youtube.com/watch?v=P4G2GHXkgEg)

## Components

- **WIZnet WIZ750SR** x 1 ([docs](https://docs.wiznet.io/Product/S2E-Module/WIZ750SR/))

## Article

[00:00](https://www.youtube.com/watch?v=undefined&t=0s)

so this is AI vision ilt what we've built here is using a simple camera and artificial intelligence to gather the object detection a method called single shot multi box detection to gather the information of the scene and report it back to the Google i/o t4o at the same time we also build a dashboard that reflects of the data so our first demo is you can see we have four bottles that's being tracked right now this is visited pull back through the stored on the Google Google datastore query and we right now you can see that you know we

[00:38](https://www.youtube.com/watch?v=undefined&t=38s)

have four bottles of milk and if we were to remove two we would only have two left right here as you can see from the scene here now we have you know - that's - being detected and we remove one more and now we will only have one at the same time we can actually change what the artificial intelligence what can detect through the whist 750s R so this is done in a socket and within the same Ethernet network over here and this allows more security as well as a much more immediate that's safe we want to track people

[01:15](https://www.youtube.com/watch?v=undefined&t=75s)

so that you know is for the customers and as well as employees during the hours as you can see the data is reflected right here and also right now is only one person that's being detected and added me and that's it we were to hide hide myself the - water would reflect it right away over here and similarly we can catch detect cars chairs as well as all the other things that's trained under the single shot detection that meteor Network how we built this is that we use the app scoreboard with the möbius in your

[01:54](https://www.youtube.com/watch?v=undefined&t=114s)

computing stick and this way we right now we're using a little PCIe over there this is allows us to do the inferencing in real time on the device itself we're uploading the entire IOT data through the helium atom and the helium hub the helium atom and how is basically connected directly to the Google IOT core and this allows the data to be transferred as seamlessly for the AI controlling part we're using the with 750 SR there's a serial to Ethernet port so this way we can kind of track you know people all sorts of things through

[02:32](https://www.youtube.com/watch?v=undefined&t=152s)

the devices right here and right now who basically tracking back the bottles instead the newer network built in can track off 220 things but uh we can also download other newer networker that can do a lot more um this entire projects between documented on haxor Daioh if you have any questions or um you know or any ideas on how to improve this feel free to comment on the project

---

Source: https://maker.wiznet.io/josephsr/projects/ai-vision-iot/