Parking Spot Detection with Raspberry Pi
The Parking Spot Detection project uses Raspberry Pi and the NanoNets API to detect open parking spaces given an aerial photo of a parking lot. The number of total parking spots and open spots are displayed.
Area of Interest
Cupertino High School
My final milestone is a model that can detect parking spaces and one that can detect cars. The number of spots open and number of total spots are displayed on the image of a parking lot using Open CV. The model’s were trained using images from Google Earth and the API used was NanoNets.
My second milestone for the Object Detection with Raspberry Pi project was to find a way to calculate the density of a parking lot – in essence – the amount of a parking lot that is taken. I ended up using a pixel method, where I calculated the pixels of the bounding boxes around the cars and the total pixel area and used the ratio to plot the density of a parking lot. This problem required a thought through top down approach which I then translated to code. The end result was a program that returned the percentage of a parking lot that was taken, accurate to about 5%.
My first milestone for the Object Detection with Raspberry Pi project was to first set up the Raspberry Pi and understand how it works. Additionally, I wanted to learn more about the NanoNets API and use one of their pre-trained models and get it to function on my Raspberry Pi. I chose a face detection model and used the camera to take pictures of my face and my family’s. Additionally, I wrote code to implement the API and drew boxes around each of the predictions and ran the program to see if it would accurately detect the faces.