Online Data Collection Module for PyMouseTracks (Raspberry Pi recording module)
To clone the repository, run the following in terminal. The recursive tag is required for the RFIDTagReader submodule to work correctly.
git clone --recursive https://github.com/tf4ong/tracker_rpi.git
To install the setup scitpt by running in the cloned folder
sudo chmod 777 setup.sh
sudo ./setup.sh
To run the tracker cage, enter the command in the cloned folder
sudo python3 main2.py
For full video tutorial on setting up the software and editing recording settings, please refer to this link
To setup a Raspberry Pi, please refer to the following links
Official Raaspberry Tutorial
Tom's Hardware
Basic Concept and Essential Hardware
![](https://github.com/tf4ong/tracker_rpi/raw/main/concept.png)
- Any Pi compatible camera can be used as long as there is an over headview of all mice
- Timestamp of each frame is written to a csv file
- RFID readings and their corresponding timestamps are also written in a csv file
- Most of the hardware can readily purchased on distributing sites such as Amazon
RFID System
- To identify and validate mice
- ID-20LA RFID tag readers setup at the custome locations at the bottom of the cage
- Locations of the readers must be sparse enough to minimize electromagnetic interference (approx 12cm)
- RFID reader details
- RFID reader base details
- RFID tag reader module developed by Jamie Boyd
- Aprroximate cost for each reader is 86 CAD
- Tags cost about 6 CAD and can be reused after cleaning with solutions like ethanol
FPS at Different Resolutions on Pi3/4
![](https://github.com/tf4ong/tracker_rpi/raw/main/fps.png)
- can be record up to 40 fps at 960 x960 on a Raspiberry Pi 4
Recording Rodents in Various Settings
Traditional Open-Field/Three Chamber
| ![](https://github.com/tf4ong/tracker_rpi/raw/main/three_chamber.PNG)
Custom Home-cage Setups
![](https://github.com/tf4ong/tracker_rpi/raw/main/home_cage_example.PNG)
PRT Analysis Analysis
- Fully python based
- Yolov4 Detection and SORT tracking
- RFID matching methods based on Pandas framework
- More information can be found on the following github link
- To ensure efficient processing speeds in the offline analysis, it is recommended that each recording do not exceede 12 hours.
- If longer recordings are needed, a simple restart will be best
- Sample video can be found at Youtube
Simple Open Field Build tutorial