
Bike Lane Violation Detection with Mask R-CNN
Software as a Service
NoCategory
Thesis DesignClients
Mapua University
Techstack
PythonDetectron2OpenCV
Purpose
The thesis aimed to detect motor vehicles encroaching on designated bike lanes from CCTV footage, providing an automated way to monitor bike lane compliance and improve cyclist safety without manual video review.
Description
A Python computer-vision system built on Detectron2's Mask R-CNN that processes CCTV video of road segments such as Shaw Boulevard. Operators draw bike lane regions directly on the frame; the system then segments cars, motorcycles, buses, and trucks with instance masks, measures their proximity to the marked lane, and flags violators in red while excluding bicycles.




