Edge Computing Video Analytics Challenge

African cities desperately need intelligent video analytics for managing traffic chaos, ensuring public safety, and optimizing urban services, but cloud-based solutions are impossibly expensive and unreliable due to bandwidth limitations. Cities cannot afford $10,000+ per camera for commercial analytics, internet costs make cloud processing prohibitive, and privacy concerns prevent centralized surveillance. Meanwhile, traffic gridlock costs economies billions, crowd disasters occur at markets and stadiums, and security threats go undetected. Developers must build edge computing video analytics achieving 30+ FPS on Jetson Nano or Raspberry Pi for traffic, crowd, or security applications, bringing affordable, private, real-time intelligence to resource-constrained environments.

The Edge Computing Video Analytics Challenge seeks innovative solutions for real-time video processing that runs entirely on edge devices like Jetson Nano or Raspberry Pi, addressing critical needs in traffic management, crowd control, and security without expensive cloud infrastructure. This challenge recognizes that African cities need intelligent video analytics but cannot afford or rely on cloud-based solutions due to bandwidth costs and connectivity issues.

Participants must develop highly optimized video processing systems achieving 30+ FPS performance on edge hardware for applications including traffic flow optimization and violation detection, crowd density monitoring and anomaly detection, security threat identification and perimeter monitoring, or queue management and people counting. The system must implement efficient neural networks through techniques like quantization, pruning, and knowledge distillation to achieve real-time performance on resource-constrained devices.

Successful solutions will demonstrate robust performance in challenging African conditions including varying lighting (bright sun to dark nights), weather conditions (dust, rain), and crowded, chaotic scenes. The system should support multiple camera inputs, provide real-time alerts and analytics dashboards, maintain privacy through on-device processing, and operate reliably 24/7 with minimal power consumption.

We particularly value solutions addressing uniquely African challenges like informal transport management, market crowd control, wildlife monitoring, or informal settlement security. The platform should cost under $200 per deployment, work with existing CCTV infrastructure, provide actionable insights to city managers, and improve public safety while respecting privacy through edge processing that avoids central surveillance.