How Amplicon delivered a high-performance surveillance solution for the Security Forces for real-time analytics and facial recognition.
The Problem: Moving Beyond Traditional Surveillance Limitations in the Middle East
For the advanced requirements of the SIRA Oyoon program, traditional surveillance systems were inadequate. The client needed to overcome several critical challenges to build a truly intelligent and responsive city-wide network:
Low Image Accuracy
One of the most pressing challenges was low image accuracy in real-world conditions. Conventional surveillance cameras and back-end analytics struggled in environments with poor lighting, glare, shadows, and high-speed vehicle movement. Motion blur and low contrast frequently degraded image quality, leading to unreliable Automatic Number Plate Recognition and missed identification events. In a city where enforcement accuracy and legal reliability are paramount, even small error rates could undermine trust in the system and reduce its effectiveness.
Slow Data Processing
Equally limiting was the slow pace of data processing inherent in centralized surveillance models. Traditional systems depended on streaming continuous high-definition video from thousands of cameras to a central control center for analysis. The approach introduced unavoidable delays, as video needed to be transmitted, queued, processed, and interpreted before alerts could be generated. In real-world security scenarios, these delays translated directly into slower response times, reducing the ability of authorities to intervene quickly during incidents or violations.
Weak Network Connectivity
Network connectivity further compounded these issues. High-resolution, real-time video streaming places enormous demands on cellular and backhaul networks. In areas with weak signal strength, limited fiber access, or congested infrastructure, video feeds frequently dropped or degraded in quality. These connectivity gaps created blind spots in the surveillance network, precisely in locations where continuous monitoring was often most critical. For a city-scale deployment, even small pockets of unreliable connectivity represented unacceptable risk
No Onboard Behavior Detection
Traditional systems also lacked onboard intelligence. Cameras functioned as passive sensors, capturing footage but incapable of understanding what they were seeing. Specific behaviors such as seatbelt violations, mobile phone usage while driving, or other contextual traffic and security violations could not be detected at the source. Instead, all footage relevant or not was transmitted to central servers, consuming bandwidth and storage while offering little immediate value.
Centralized Bottlenecks
As the number of deployed cameras increased into the thousands, back-end servers became bottlenecks. Expanding capacity required massive investments in data centers, storage arrays, and processing clusters. This not only increased costs but also introduced single points of failure. Any disruption at the central level could impact large portions of the surveillance network, undermining reliability and resilience.
The Client: Government Security Forces
The client driving this transformation is a globally recognized government security authority and a pioneer in public safety innovation. With a reputation for adopting the latest technologies ahead of global trends, the organization has consistently positioned itself at the forefront of smart city development. Their leadership in deploying advanced surveillance, traffic management, and emergency response systems has made them a benchmark for cities worldwide.
The SIRA Oyoon initiative represents one of the most ambitious expressions of this vision. Conceived as a city-wide, AI-powered security ecosystem, the program aims to connect thousands of surveillance cameras across Dubai into a unified, intelligent network. Rather than relying on reactive monitoring, the goal is to enable proactive security, accelerated emergency response, and data-driven traffic management that enhances safety and quality of life for residents and visitors alike.
To realize this vision, the client sought a solution that could deliver real-time intelligence without overwhelming networks or central infrastructure. They required a partner capable of designing and deploying an architecture that balanced decentralization with centralized oversight, ensuring both autonomy at the edge and unified command at the city level.
The Technology: Edge AI Computers and 5G Routers in Dubai
A comprehensive hardware solution was engineered to provide intelligent processing and reliable connectivity directly at the edge, forming a powerful outdoor surveillance connectivity stack.
Edge AI Computers: A combination of Advantech MIC AI surveillance platforms (MIC-711, MIC-733-AO) and powerful, compact ASUS IoT edge computers (PE1100N, PE2100U) were deployed to perform on-site video analytics.
Industrial 5G/4G Routers: 5G routers for video surveillance from Robustel, including the R2010 model, provided secure, high-bandwidth uplinks for real-time data transmission.
High-Gain Antennas: Panorama 4G/5G surveillance antennas were used to ensure resilient, high-gain connectivity, providing a stable link even in challenging environments.
The Solution: Decentralized Intelligence and Real-Time Alerts
The integrated hardware stack delivered a powerful and scalable solution, creating one of the most advanced smart surveillance systems in the world.
The high-performance edge computing for CCTV platforms from Advantech and ASUS process HD video feeds directly at the camera site. It resulted in local AI analytics for high-accuracy ANPR, facial recognition, and real-time behavior detection generating instant alerts without the delay of back-end processing.
The rugged networking for security systems make sure that while intelligence is decentralized, critical alerts are reliably transmitted to the central command center via the Robustel routers and Panorama antennas.
The architecture is highly scalable and resilient, reducing server load and ensuring constant uptime. The solution also effectively supports both centralized monitoring of critical alerts and decentralized processing of raw data, making the entire network more efficient and intelligent.
