-
-
The Smartrack home page
Inspiration
We were inspired from hearing about a problem a friend faced while working in an Amazon workhouse, where they had to count the number of boxes on the conveyor belt as they're being shipped out. We thought that we could leverage computer vision technologies to automate this process!
What it does
Our project smarTrack takes in a video of boxes on a conveyor belt as input. It will then count the number of unique boxes that have passed the camera that filmed the video, and then return that count!
How we built it
On the front end we used HTML for the UI and Flask to create the app. For box detection and trackng we trained a DeepSort model, and used OpenCV and Python to give the video input to the model to track the boxes.
Challenges we ran into
On the backend, challenges that we ran into were finding an appropriate model to train for not only detection but also tracking. We also had to handle dependencies that did not match our user environment, so had to spend time setting up the proper virtual environment. On the frontend, we ran into challenges integrating Flask with HTML due to unfamiliarity with the framework. We also had to figure out how to transfer data between the frontend and the tracker in the backend.
Accomplishments that we're proud of
At the end, what we're most proud of is that weve created a functional service fro tracking boxes, that can be easily modified to work in real life circumstances and aid in logistics.
What we learned
- DeepSort algorithm
- Flask
What's next for smartTrack
- tracking jiffys (white bags) alongside bags
- increasing model accuracy
- working with live camera footage
Log in or sign up for Devpost to join the conversation.