Vehicle Violation Detection Using Faster R-CNN and SPP-Net in Bike Lanes

Clients

Techstack

Purpose

Motorized vehicles frequently intrude into bike lanes, endangering cyclists. This study compares Faster R-CNN and SPP-Net to determine which algorithm detects bike lane violations more accurately, scalably, and stably for automated enforcement.

Description

Both convolutional neural network architectures were implemented in Python with TensorFlow and evaluated head-to-head on bike lane footage. The comparison measures detection accuracy, how each algorithm scales as the dataset grows, and processing speed and stability. The results identify which approach is better suited for monitoring bike lane usage, giving enforcement systems a grounded basis for detecting motorized vehicles that violate dedicated cycling infrastructure.