Before you read
If you want to go straight to the facts and theory you can skip this. If you’re reading this, thank you. This project is something I’m very passionate about. Growing up in Perú wasn’t hell but, there were a lot of things that happened that shouldn’t have, and a lot that should have but didn’t. That’s as specific as I’d like to be. However, I will say that I was a very sensitive child, I’m kinda still am, I remember being a kid and watching the news. Which if you know anything about Peruvian news outlets, is that they censor very little and they’re not careful of what they say. It’s easy to scare an 8 year old about murdering gangs, neglectful and even hateful parents. It’s a lot to process for a kid. I still to this day, at 23 years old, keep a healthy distance from watching the news. That doesn’t mean I’m apathetic now. I’m still a little angry and scared kid I was back then. It’s hard having a sense of safety when you feel that the world is a terrible place. I was growing up scared of leaving my home, scared that one of my loved ones would get hurt. Then one day, a little after I turned 12, my brother and I received a gift, it was the first Batman Arkham Game, Arkham Asylum. Playing this, at 12 years old and being a ball of anxiety, was quite the experience. No matter how scared or anxious I felt, my brother and I finished the game and something clicked in my brain. I know what I need to do. And yes being a young kid, my first thought was I want to be Batman and scare criminals as well as show people hope. As I grew up, the idea of being a vigilante seemed very silly, but the idea of inspiring people and giving people hope, that didn’t seem silly. Because of this game I got to see all kinds of crazy devices and gadgets, so “little me” made the decision to study electrical engineering. I thought that maybe one day I’d be able to build something that can help people. And that brings me to today, for my thesis project I wanted to build an ESP32 CAM and network powered with deep learning to identify different needs or medical emergencies that people may have.
The inspiration
Ever since I got into college, one question plagued my head for years, what am I going to do for my thesis project. I knew it had to be something that could potentially help save lives or at least be helpful for someone. My intentions, unlike some of my peers, were not to be the first or best at something, or become rich or famous. Some people have the option to even work with a big corporation to optimize something and present this as a thesis project. It’s necessarily wrong, it just wasn’t meaningful to me. I mean I wouldn’t feel ok just helping a company save money and make a living. I needed to do something else. Then, for once. the Youtube Algorithm was helpful, and presented a video about an engineer that developed a system named EEVEE
I was blown away, and the first thing that popped in my head was, this is it. So I began research, what else could I add? How can I make this cheaper? Can I make this reliable? My first Idea was to use LoRa technology, since I had been part of a project very recently.
It was a project in which we measured CO2 levels, sent it via LoRa to a gateway, which was an ESP32 connected to a LoRa module, and the data was uploaded to the cloud. Everything worked pretty ok, at least everything I worked on, which was the web platform and the ESP32 Gateway, unfortunately we classmates weren’t as productive. Anyway, the only problem is that this LoRa Module was very slow, and if I want to transmit images or a video stream, it’s going to be very difficult. If I wanted to use LoRa and do video streaming it was going to get very complicated and very expensive. So LoRa was now out of the question, however, I didn’t have to look too hard for a replacement, I just googled cheap MCUs with integrated cameras and the ESP32 CAM was one of the first things to appear.
And since it’s from the ESP32 family, I didn’t have to worry about integrating with bluetooth or extra Wifi modules, everything is already on the board. I even got a different camera since the OV2640 doesn’t have a good resolution. I found the OV5640 in aliexpress which has a higher resolution and when I tested it it even improved the frame rate.
Now at this point, I was procrastinating the fact that I had to sign up for another free AWS subscription because I have beef with Amazon and how they treat their employees, but for educational purposes, AWS will have to do. I designed the following architecture that would include a lot of moving parts. I’ll try to make blogs or maybe even videos for every part. For now I’ll end it here, next time I’ll share how to set up the AWS IoT core so that the ESP32 can publish images over MQTT.
