Housewise was to be a window into the environment and consumption patterns of a house. We planned to set up an expandable network of sensors in the house, and collect all this information and report these numbers to a cloud server. At a high level, and as the first pass, we wanted to
- Use the bluey boards to control and read information from the sensors.
- They would then broadcast this information in BLE beacons, coded with their House-ID and Sensor ID. (Since the bluey does not have wifi)
- In addition, we had a webcam attached to a raspberry pi, that would be focused on a meter. We ran some scripts to capture an image at regualr intervals and then do some preprocessing of the image before passing it to an open source OCR library called Tesseract. Once a number was extracted from this process, ti would be directly posted to our DB in the cloud.
- An edison board, acting as a sort of a hub in this enviroment, would regularly scan for the BLE beacons, extract the values and post the tuples to the cloud DB.
- The frontend, would read the values from the DB and make fancy graphs and predictions or reports on quality and usage patterns.
After some hits and misses with sensors/sensor quality, we finally chose a noise sensor (REES52 Sound Detection sensor Module ), and an air quality sensor (Insert name here).
Most things worked as planned. Please see github (https://github.com/pnarasim/efh_housewise.git) for details and code on each module of this project. All the instructions to set up each peice of hardware and software (sensors, calibrations, bluey, firmware code, edison setup and scripts, cloud setup including the DB, frontend, etc) are included in this github repo.
However, we underestimated the image processing part of this project. We need to understand the steps that go in to preparing an image for OCR (cropping, edge detection etc) before we can make reasonably robust product. This is the part that we will be focusing on for the next steps.
This is a picture of the frontend showing the noise levels in the room on demo day. https://github.com/pnarasim/efh_housewise/blob/master/writeup/noise_levels_on_demo_day.png
Also a video showing the bluey connected to our noise and air quality sensors.