Project
# | Title | Team Members | TA | Documents | Sponsor |
---|---|---|---|---|---|
10 | Smart Squirrel Proof Bird Feeder |
Christine Li Linfei Jing Yitian Xue |
Shaoyu Meng | design_document3.pdf final_paper1.pdf presentation1.pdf proposal1.pdf |
|
Problem Squirrels are the “enemy” of backyard birders. Even though squirrels are cute, they are considered to be pests to the bird lovers. They don’t just want to share the bird food, but they can take all of them. Their amazing athletic ability, voracious appetite and ability to chew through almost everything makes them unstoppable in the backyard. The main goal of this project is to keep squirrels away from bird feeders and provide a peaceful environment for the birders. Solution Overview To solve this problem, we plan to build a smart bird feeder with a camera compatible as well as a smartphone application. The camera will first distinguish bird and squirrel by machine learning algorithm with pre-studied pictures. Other sensors such as pressure sensors can also be used to support the bird feeder to distinguish squirrels. After the identification, the feeder machine will either automatically load a reasonable amount of bird food based on its weight or prevent squirrels from stealing bird food. In addition, when the food is almost eaten up, the feeder will notify people to refill the food on the app. To please the backyard birders, we can also create an additional feature of taking birds’ pictures when birds are eating in front of the bird feeder. Our solution is an innovation to the existing products. The squirrel repellent bird feeders that sell currently passively prevent squirrels from taking the bird food, and squirrels are smart enough to beat the feeders. According to the youtuber Mark Rober, squirrels managed to overcome all the squirrel proof bird feeders he tested. Our smart squirrel proof bird feeders design to actively provide food for only birds. Solution Components Hardware [Bird image classification subsystem] The bird image classification system will have two outputs: either identified as bird or not. Components: camera, Arduino [Feeder subsystem] The feeder system will connect to the bird image classification system and take the signal transmitted back as input to either load the food or not. It is also implemented with a pressure sensor to notify the user of empty tanks. A mechanical part will be implemented that allows this system to load appropriate amounts of food to birds. Components: pressure sensor, mechanical component [Squirrels Repel System] The repeller system will generate ultrasonic waves that are above the audible frequency range, usually above 20,000 Hz, to repel squirrels. Components: square wave generator [Power system] The power system will support power to all other subsystems. Components: battery Software [User Application] Web App for users to interact with the smart bird feeder. It will notice the users when the food container is empty and send birds’ pictures captured by the camera. It also stores the data of the bird feeder. [Image processing program] Identify birds and squirrels with machine learning algorithms and pre-studied pictures. Criterion for Success Our solution will be successful if the camera can accurately identify birds within a short amount of time, the user will be notified when the tank is empty, the feeder system will load food with an accurate amount of food, and the squirrels can’t easily destroy the bird feeder and will be repelled. |