Hi there. Welcome to our IOT group project group 45 Security Camera. Our IOT project title is Object Detection Security Camera. And it belongs to the security domain to describe this IOT system. Um I would say that we would develop a security camera and we did take inspiration from surveillance home. Security uses personal uses. Government uses our companies and our camera. How it works is that when it sees an object, it tries to figure out if it's a person or not, this was previously a feature on our IOT project. But due to our security uh reasoning, we change this to only people. But regardless if it figures out if it's a person or not, then it uh if it is a person, it draws an outline around the object and also displays a percentage that it thinks it's right. So it has a threshold. We used Python with a computer vision library to achieve this. It achieves the camera output to a website. Oh, sorry. It streams the camera output output to a website for which we used flask. And our motive was to create a security camera that displays its output somewhere we use the web because it was, it would be an easy feature to implement and what it requires to design any physical equipment or change or do any drastic procedures on the whole. Our product fulfills the role of a security camera as it outputs itself to the website and informs the user if it detects a person. So our IOT project does use flask server and local host and it does use OPEN CV to achieve the detection feature. As I stated previously, our code used to be able to detect objects, but this wasn't the end result. We wanted, we wanted a more person slash human detection as it's mostly com it's most commonly used in home security. And the fact that we didn't continue with object detection is that it would sometimes overheat the Raspberry Pi due to its only due to its only having passive cooling. And our IOT does detect multiple people. It's not just one person but you. But because of this, uh the frame rate of the camera is lower than usual. We weren't, we weren't able to achieve the 30 fps that we wanted, but it is able to achieve human detection when it comes to similar markets or when it comes to similar products in the markets today that we can, that people can buy, users can buy. There are a couple of products that are similar to ours which provide the security feature. And uh prime example is NES CM IQ indoor, which is very similar to our pro product. The Nes CM also has an advanced, also uses advanced computer vision capabilities. The Nes CM uses A I algorithms in order to detect and identify people and then also has the ability to send an alert to the user's smartphone. When a person is detected in front of the camera, ours isn't able to do this as we use a website, it will outline the shape of the detective person and will then also provide a confidence score like ours. But we we call it a threshold and um through the help of computer vision, it determines the likelihood of the accuracy of the detection. There is the option for a continuous stream to the app which allows for remote remote viewing of a user's home by the user when they decide and more information on their website. Another notable example is the Arlo Ultra two, a wireless security camera system with advanced motion detection features and four KHDR video. It may not provide a clear outline of items it has identified, but it does provide customizable motion zones and notifications for certain activities such as human detection. It also sends a live video to the Arlo app so that they can be viewed and monitor monitored remotely, which has more information on their website. So to to to conclude our products are, our product is really similar to other products out there that people can buy. So when it comes to the functionality of this code. Uh It works by using a flash web application that performs object detection on video frames captured from the camera feed A K the module camera from the Raspberry Pi. It imports the code. The code imports necessary libraries such as flask for creating the web application open CV for computer vision tasks and NP for numerical calculate uh computations as I mentioned previously. So continuing on continuing on, it initializes a flask application sets up the object detection model. The model uses used here is SSD mobile net V three trained on the Cocoa data set which is here with all the the object detection names. It loads the class names model configuration file and weights. Now to the actual object detection function, the get objects function takes an image threshold, uh takes an image threshold values for confidence and non maximums suppression A KN MS and optionally a list of objects detect to detect it performs object detection using the loaded model on the provided image. It draws bounding boxes around detected objects and labels them if draw is set to true, then frame generation function which is down here. The generate frames, the generate frames function continuously captures frames from the camera feed. It performs object detection on each frame using the get objects function encodes the frames into JPEG format and yields them one by one with detection overlay onto our routes. The route renders a html template presumably that contains a video player for viewing the camera feed right here under their index HTML. It contains the object detection title a header and the actual URL for the video feed. The video feed route streams, the video feed with object detection overlay, it returns a response object with a generator function that yields frames. The main block for this starts the flask application and runs it on host 0.0 0.0 0.0 accessible from any IP locally and port 8000. So when you run this code and navigate to the appropriate URL, which is local hosts two dots 8000, you see a web page displaying the video feed from your camera with real time object detection overlay, specifically highlighting person objects. So that is how our IOT project works. And um there isn't much to it. That is all. Thank you so much for listening.