Reproduce “Recovering Origin Destination Flows from Bus CCTV: Early Results from Nairobi and Kigali” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Nthenya Kyatha, Jay Taneja
Researchers from the University of Massachusetts Amherst proposed a pipeline to recover bus Origin–Destination (OD) passenger flows by integrating existing CCTV and telematics data from buses in Nairobi and Kigali. The system, with domain-specific refinements, achieved an overall accuracy of 92.4%, demonstrating a feasible method for detailed transit planning despite operational challenges prevalent in sub-Saharan Africa.
Key insights
- Leveraging existing CCTV and telematics infrastructure is a technically feasible and cost-effective approach for recovering detailed Origin-Destination passenger flow data in challenging SSA bus environments.
- Domain-informed engineering refinements, such as explicit door-state gating for counting events and flexible stop classification, yield substantial improvements in accuracy, particularly for alighting passengers.
- The baseline system established a foundation while identifying specific failure modes like significant undercounting during extreme overcrowding, degradation from poor lighting and modality shifts, and tracklet fragmentation due to varied passenger postures.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.