Logic and Core inspired by Full program details at: https://gregtinkers.wordpress.com/2016/03/25/car-speed-detector/
Light weight Python based traffic speed tracker using the Raspberry Pi. Tracking traffic in front of your street in a side view with traffic going from the left and right. It has been tested with the Pi Zero and RP4. It should work with any Ubuntu based distro, maybe other operating systems (NO not Windows) as long as dependencies are met. The speed tracker is not designed for highways or busy roads, it's targeted for rural and residential streets. It's great tool to get a sense of how traffic goes by in a day and how fast they are going.
Heres some of the features of the Web Control panels:
- Graphs by Hours, Times, and day of week, all searchable with custom plots points.
- Cam status give a quick view of the most recent 100 (Default but can be change) photos and let's you know if you speed cam is running. Note: all speed tracked are stored as photos, these can be viewed in the graph section section.
- Calibrating your camera. You can take quick snapshot of what you camera sees in the control panel. This allows you to focus and point the camera at the part of the road you want. You also have the ability to draw a box on the region of the road you want to monitor here.
- Editing configurations, is all in one file and should be pretty self explanatory, one you open the editor.
- Scheduling when the camera turns off and on can be done here as well.
- There is a debugger to help, to tell you what the AI see, in the as images. This can help determine if you have the camera physical location, need to configure the threshold, or redraw the a better section of road to track.
- There are real time logs to for the webserver, speed camera, and server (computer) load & resources.
- All configurations are managed in the web control panel.
- Data is stored in a SQLight3 database. No worries if you don't know SQLight, this is behind the scenes. The database is created and updates are handled by the app. The data is primarily used for reporting with graphs and such. If you have SQL skills and know SQLight3 you can pull you own report. The database is in
./db/speed-tracker.sqlite3.
- CSV files are also generated if you want to import to Excel. Each time the App start a new CSV file is created in the
./data/directory
data/
├── carspeed_20210208_0631.csv
├── ...
└── carspeed_20210213_0952.csv
- Images are stored in folders by
./media/images/YYYY/MM/DD. Each car that has a reported speed will have an image saved. Generally this image is used with the webserver, but is also documented in the SQLight database and CSV files.
media
└── images
└── 2021
└── 02
├── 08
├── 09
├── 10
├── 11
└── 12
- Query and cycle through each record
- Turn Cam on/off with the toggle switch
- See most recent entries
- Raspberry Pi 4 2gb RAM (Recommended, 4gb would be better)
- We do need the desktop environment to calibrate so don't go headless.
- Raspbian Buster - Should work with most Ubuntu cores.
- I've installed and it working fine on a PI Zero. A little sluggish but does the job.
- Camera module
- Opencv >=4.1.0.25
- python3
- Not a whole lot but a little terminal skills is nice