As a final project for DIG333: Physical Computing, I modified a garden owl which analyzes pi-camera data to figure out fashion trends, count how many people pass the camera, and the noise that is being created as each person passes. Birds have very advanced eyesight which is what motivates having an analyzed camera in the system. A microphone will be used to analyze overall volume of the environment. This will be used to understand how loud humans are being in the given area and the impact that humans are having on the audio-volume of a particular place. A pi LDR (light sensor) will help the owl decide when to sleep (or put only certain sensors to sleep). Tweets will be produced by the owl (lol) related to the data. These tweets will be composed of sentiment related to how busy the walkway is and how loud it's been in the past 2 hours along with the average color of pedestrians' shirts for the day. There's a possibility that data will be visualized using Dash, as well, on my website. See sentdex tutorial on Dash on YouTube.
Python Dependencies:
- numpy
- matplotlib
- cv2 (opencv)
To download opencv on linux:
sudo apt-get install libopencv-dev python-opencv
Sensors:
- pi-camera
- microphone
- pi light sensor
Design:
Potential location:
Dataflow:
Potential visualization:
- OpenCV - A computer vision module for Python.
- HaarCascades - Definition of objects to be found in video stream in conjunction with OpenCV.
- Dummy Owl
- Tucker Craig - Davidson College Class of 2020 - tucraig
See also the list of contributors who participated in this project.