Let's build a Smart Fridge using IOTA and Azure IoT Edge



hey guys how's it going so in this video I want to show you a project I was working on the last couple of weeks it's basically a smart bridge which recognizes items like you know bananas strawberries for apples and once one of those items is actually move the the smart fridge actually asked the closest supermarkets for pricing information and actually order stopped all the communication is done buddy are you the protocol and we will be using Azure idh for any artificial intelligence service all right let me quickly explain what you see on the screen here so first of all we have the SG wise in our case this is a raspberry pi 3 I also connect the screen so we actually know which stage the device currently yes I know the video quality is not the best but you should really get a first impression here on this their operating system is raspbian us and if you want to follow this demo make sure to install OpenCV as well which is basically a great open source library for video processing I also installed the azure IOT at runtime which is basically our platform for the application and that's why we can take advantage of Microsoft address image recognized service which is a great way actually to create machining running modules next we have a USB camera this is actually nothing fancy here it's just a regular camera connected to respirate in my USB you can basically use any camera you configure and last but not least we have a fridge I know it's not the most beautiful fresh here but you can see we with an apple and we have over there a strawberry so if you look on the screen the raspberry the service is basically telling us that it still sees the apple and strawberry so let me actually remove the apple the background processing is being done by an external application hosted in the cloud this is quite a common parent for our IOT scenarios as heavy workloads cannot be processed by smaller devices so let me actually fast-forward here a bit to see some output from the application as you can see on the screen the device tells us that we remove the Apple and we are using edges image recognition service which are trying to do with you know apple and banana pictures especially a very easy way to get you started with AI services as the module is basically a drop container which can be deployed to any service in any platform the application asks the closest supermarkets for latest pricing information the uniform was pushed earlier by myself to our last hanger by a tool which will show you guys later in this video we're using a u.s. mass identification messaging here as it is a very cool way to actually stream information to any device the network was slow that they are recording this video so let me actually channel bit forward in a second you will see on the screen that the device figured out the best offer for the Apple since services on I would have wallet it starts the ordering process based on the best price it has their model token received from the MA and channel to the address of the supermarket let's quickly verify this using the web application I created for the proof-of-concept so build were a simple web application with render node if you're onto stuff locally I recommend you to check out the github link below this video so when we click on the word tab which is basically a wallet for the supermarkets it's talking for all transactions really need to and they will address and here we see a list sorted by the date we can actually sort by date as you can and the messages to each a transaction so I figure maximum size here is about 1 kilo body lessons click on the link here on the right side let's put the latest one and we can actually verify what's going on in the syringe excellence since we can see what's in the adjacent body there we see all the items which was entered by our restaurant or I guess let me also quickly show you how the devices actually selects the cheapest items from a supermarket so when we click here on this published magazine info tab you basically gets mass so that's quickly at free apples with different prices and push the stuff to the NBN channel after second the messages to be pushed to the mah we can verify this by copying the root key and browse to the M AMX pearl web application which is actually a very easy way of written by Alex Isola to access the ma M stream let me also select the current provider as the Stamos running on I would ask me Annette and choose the correct mode use public here since there's no password protection for a stable finally we can see what's going on inside of each I me a message my service is adding another poverty which basically tells the edge device where it needs to set focus the last step is to update the device to a property so the raspberry is actually aware of the latest changes to the a man channel just click on the bottom and right here move to the azure portal and you can see that the latest MA and channel 40 was updated to the Mei Mookie we created a minute ago let's finish the demo by putting back the Apple to the fridge so the whole process can start over I have to add some extra light as it was very dark that night I recorded Aminu off and my machine learning model is trained for a day and the images and there we go the device recognizes the Apple on says that all items are still in fridge so that's it for my son thank you guys for watching and I will see you next time

7 Comments

  1. stelzi290888 said:

    It´s always cool to see how people are able to transfer ideas into a working use cases. People like you are much more important to the world than we can imagine. Thanks.

    May 22, 2019
    Reply
  2. Newton Sheikh said:

    Brilliant

    May 22, 2019
    Reply
  3. Ted Mosbi said:

    You are f* awesome!

    May 22, 2019
    Reply
  4. Diogo Vitorino said:

    Excellent job! Congratulations

    May 22, 2019
    Reply
  5. virtual resistance said:

    BRAVO

    May 22, 2019
    Reply
  6. Nikhil Kumar Uttar said:

    Hey mate, this can be done on Xinfin as well, you should totally check it out, https://www.kramaa.com/, I would love to see how this will be done on Xinfin 🙂

    May 22, 2019
    Reply
  7. Handn Dkgkx said:

    Love you 🤖

    May 22, 2019
    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *