Find My Bike
Online monitoring of bicycles.
Tracking bike movement on the map
Statistical data collection
Video archive analysis
On the site you can get access to the the system, in configuration which was tested in the video surveillance system in Moscow 2019.
ДИТ
Test results*
- The probability of true-positive identification of the bike was from 94% to 99%
- In other words, even in poor conditions, cameras that are not designed to track the roadway, the share of passes was no more than 6%.
- The average level of reduction is from 200 to 400 times
- This means that if 5,000 bikes passes were recorded, the operator on average needs to view 15 to 20 images instead of 5,000.
The tracking of the bike on the city map
- by parametric description
- by photo
For demonstration purposes the system is connected to several public cameras in the Sokolniki park**.
**Public cameras do not guarantee a stable connection, and work with a small FPS, the results may differ from working with cameras in the city's video surveillance network
Statistical data collection
- number of passes per unit of time
- statistics on types of bikes
- ...
Demo Tell me more
Video archive analysis
- search and recognition
- detection in poor visibility conditions
- analysis at x10+ speed
We recorded video from public cameras located on the streets of cities in EU, Asia and the US. The link shows the results of the system operation on these records
*Data on the results of the first part of the pilot project with IT Department of Moscow (summer-autumn 2019).