The role of AI, IoT and edge computing in the future of physical security

‍

There was a time when technology like ML, AI and IoT devices were only seen in Hollywood movies. It is now quickly becoming common place, and technologies like AI and IoT are enabling access to rich data insights for a variety of industries faster than ever. IDC’s forecast that global spending on artificial intelligence (AI) to double over the next four years, growing from $50.1 billion in 2020 to more than $110 billion in 2024. Today, businesses don’t find investing in such technology a luxury but a necessity.

Like any other industry, such technology has impacted the security industry. The  Security Industry Association (SIA) announced artificial intelligence as #1 in its top 10 megatrends for 2021. AI claimed the first position for the first time, and it’s not difficult to understand why. Thanks to AI, machine learning, and the rise of IoT-based devices, providing complex intelligence at the touch of a button and allowing labour resources to be allocated toward more important tasks.

‍

Technology and its impact on different aspects of physical security

Machine learning (ML) and artificial intelligence (AI) to detect objects

Real-time object detection is becoming a significant area of interest among data scientists. From developing ‘smart’ cities, offices and retail centres, ML and AI now plays a crucial role in the detection of behaviour trends in environments, from foot traffic to vehicle flow. With new, more intuitive technologies on hand, we are now poised better to detect objects out of the background images than ever before, using a probability score to classify the type of object, defining the boundaries of the object with x-y origins and height and length values.

Leveraging AI, we can also analyse historical data, draw inferences, and make predictions. For example, with the help of deep learning AI, footage can be analysed to find anomalies in real-time much faster. A recent application of this is seen in the context of the recent COVID-19 pandemic, globally, AI and deep learning has been used in airports, offices and manufacturing facilities to enforce face mask wearing and social distancing. For many of us, this is the new world we live in, where AI can be used to make decisions accurately and efficiently.

Edge data processing on the camera to reduce data usage

With the arrival of high resolution cameras (HD and 4K), difficulties can arise when transmitting large volumes (gigabytes) of data across a network, including an associated increase in bandwidth cost and a latency in data transmission. Coupling this with the remote location of some cameras, it makes analysing camera footage in the cloud a slow and costly task.

Edge computing solves this problem. An edge device works at the source, processing the footage on the device, only sending relevant data back to the cloud, thus reducing the bandwidth needed. For real-time applications, a server on the cloud can send data back. Companies, including Vision Intelligence, are now utilising edge computing to solve the requirements of high volume data analysis. Solar powered edge processing adds another level of complexity due to the minimal power available.

How new technologies in CCTV cameras and the rise of IoT devices are impacting different sectors

Mining and heavy industry

Specialised and integrated video systems make operations easy under different weather conditions. Highly sophisticated devices that integrate video analysis systems allow automated in-depth analysis, otherwise impossible for man to do in that given timeframe. IoT devices can also help in process control by measuring weight, temperature and vibration,  monitoring the location and speed of vehicles and plant equipment, access control implementation, and many other control systems. In heavy industries, IoT has countless applications, from premises security, occupational health and safety to asset performance and accident prevention. IoT and integrated video surveillance systems help to improve visibility in day-to-day operations for managers to easily understand complex processes.

Agriculture

IoT and cameras are helping the agriculture industry too, from monitoring farm activity to the control of farm equipment. Livestock and equipment theft is a growing issue in Australia. With the use of IoT sensors and deep learning from cameras, crops, equipment and livestock can be monitored remotely, with triggers set based on certain conditions or behaviour, eliminating the need for a farm worker to be present 24/7. For example, if there are five cows not present in the view of a camera at hourly intervals, an alert can be sent to the relevant person notifying them.

Local government authorities

Anti-social behaviour is a costly and high-profile topic that councils deal with on a daily basis. From monitoring parking, to graffiti and rubbish dumping, it requires significant resource and investment. IoT devices and deep learning on cameras are transforming cities into ‘smart’ cities, reducing labour and maintenance, increasing productivity and aiding councils in tracking down serial offenders that damage council property. There are countless applications, from bin capacity levels to soil monitoring, however, the most popular application we find is for physical security. Utilising number plate recognition and movement sensors, councils can reduce response and clean up times to rubbish dumping and graffiti incidents. Biometric security technology is becoming more widespread in our communities, with automatic number plate recognition from 11,000 cameras now widely used across the United Kingdom, as an example.

What does the future look like for AI, IoT and the security market?

The future of IoT and security surveillance is continuing to develop at a rapid rate. According to a report prepared by Grand View Research, Inc., it is reported that "the global security market size is anticipated to reach USD 167.12 billion by 2025". New applications for IoT, AI and deep learning are being found every day, transforming how industries perform labour intensive tasks to improve safety and security. Nonetheless, it will take time as companies and local authorities develop the understanding on how to apply these technologies and outlay capital expenditure to upgrade their infrastructure.