Description:
_Current Solutions
Current video analytics solutions are highly cloud-dependent. When relying solely on cloud infrastructure for computer vision applications, you need to send the video data to the cloud, process the data there and send the insight to where the decision is being implemented. Although analog or IP cameras can be equipped with AI chips to reduce such cloud dependencies, only a few simple algorithms can run in parallel on these devices, yet requiring cloud infrastructure for more complex applications such as data fusion.
_The Problem
Cloud dependent approaches are highly expensive (i.e. cloud cost and communication infrastructure cost), they endanger the security and privacy of the citizens' sensitive data and they introduce undesired latency for time-critical applications. These limitations prevent the wide-scale adoption of such solutions. _NATIX Edge Vision
NATIX Edge Vision turns a network of analog or IP cameras into an intelligent swarm that identifies relevant events and triggers corresponding actions. This minimizes the cloud dependencies of the system. NATIX Software Stack installed on cameras enables seamless deployment of AI models on these cameras. Additionally, the entire camera network can be utilized for data, computation power and application exchange to deliver better detection, prediction, and planning. This way, the user benefits from computational capacity and SaaS solutions comparable to that of the cloud while maintaining advantages of edge computing solutions.
_Benefits
__Drastically lower costs
No need to install additional sensor infrastructure, when you can utilize the already existing camera sensors. Also, with NATIX, you can decrease your communication infrastructure cost, save your network bandwidth and better utilize the computational capacity available within your fleet of edge devices.
__No need to worry about privacy
NATIX comes with a real-time anonymization AI module, removing sensitive data (e.g. faces and number plates) from the footage before storing or sharing it with other entities. Event detection through NATIX is highly privacy-preserving as no sensitive data will leave the cameras.
__Share the infrastructure, not the video stream
Different entities (e.g. Municipalities, Businesses, etc.) can cross-share the camera infrastructure for event recognition without providing access to the camera or the live video feed. Instead, they can share AI models that are executed locally with results (e.g. notification with the relevant footage and metadata) being shared with relevant stakeholders.
__React faster
Time is an important factor when it comes to certain applications such as infrastructure assisted driving. With NATIX, there is no need to send the data to the cloud or data centers and wait for a response. You can make sure everything is done locally, right there at the heart of the city.
__Choose your communication infrastructure more flexibly
NATIX eliminates the need for expensive wireline communication infrastructure (i.e. fiber optic cable) required to stream video data to the cloud or control room. Event detection is performed locally on the camera and only a notification with the relevant footage and metadata is shared with other entities. _Use-cases
Video information is a rich source of data empowering smart cities and smart campuses to better sense the environment at micro-levels. By the end of 2020, a sensor network of appx. one billion cameras will be deployed on government properties, infrastructure, and commercial buildings. This translates to 1.6 exabytes of data per day that, when combined with video analytics, can support various use-cases.
__Mass Mobility Behaviour Understanding
Understanding the mass mobility behavior of citizens and vehicles in public spaces helps cities to make real-time decisions and optimize traffic patterns. This includes tracking metrics such as origin, destination, modal split for transportation, hot-spots, and average travel duration.
__Parking Traffic Guidance
Finding a free parking spot at the desired destination is a challenging task, especially in urban areas. Detecting available parking spots, as well as the parking traffic, helps to minimize the expected time to find an available spot.
__Road Condition Monitoring
Road condition monitoring, including accident detection, obstacles on the road or damage of the road surface, helps to provide maintenance and preventive services more efficiently and in a timely manner.
__Crowd Detection
Analyzing crowd activities rather than individuals is valuable for cities and workplaces to detect and control abnormal behaviors that leads to dangerous situations. Such insight can also be used for applications such as crowd, disaster or energy management.
__Infrastructure-assisted Autonomous Driving
Infrastructure-assisted driving solutions are emerging in controlled environments (e.g. park houses, logistic centres or car manufacturing plants) for process optimization. The vehicle relies on on-board sensors, however, camera infrastructure assistance is crucial to cover vehicle’s blindspots and to increase the system reliability.
__Safety Enhancement of Workplaces
Proactive monitoring and on-time hazard alerts can increase workplace safety. Applications include people counting, zone counting, restricted/forbidden zone monitoring, anomaly detection (e.g. smoke and fire), machine tracking and workers’ compliance with safety regulations (e.g. helmet usage). _The Technology
NATIX provides a framework to create scalable infrastructures at the edge of the network leveraging commodity devices. Applications can be deployed and executed on selected devices using containerization technologies. Beyond deployment, NATIX assures high quality of service as well as the privacy and security of the system by monitoring the network health and dynamically managing resource usage.
NATIX Edge Agent is able to run on any commodity hardware with Linux OS making it compatible with most of the devices found at the edge (e.g. smart cameras, Raspberry Pis, LTE/5G stations, routers, IoT gateways, etc.). To manage an entire fleet of devices and to reach a self-governing network, decentralized orchestrators and validators are needed across the network. NATIX is designed to be capable of delivering endless applications through a fine-tuned framework. This framework aims to ease the process of building edge-compatible distributed applications for developers. The Fog Service layer aims to serve as a bridge between the edge network and the outside world and is crucial for achieving real-time analysis and flexible data storage. NATIX provides a full overview and control of data flow and policy-based contracting to the device owners (and other involved parties) ensuring transparency and integrity of the network.