Sensing, AI and imagination: How vision is shaping the Internet of Things

Vision is quickly becoming the leading sensing application in the development of the Internet of Things, which is profoundly changing our world.

Think about factories and manufacturing. Computer vision systems can transform modern factories by ensuring quality control, optimizing processes, reducing waste and driving continuous improvement. These systems help improve productivity, cost-effectiveness, and competitiveness of manufacturing operations.

In a recent Arm IoT survey, industrial respondents said the two main reasons they are adopting IoT technologies are to improve their use of data to change business decisions and improve customer experience. In commercial construction, a similar revolution is underway.

Building and IoT Vision Sensors

Building managers are leveraging IoT visual sensing technology to monitor and analyze activity inside buildings to improve space usage efficiency. By collecting and analyzing foot traffic data, office and work area occupancy, they are able to better plan office space layout and seating arrangements, as well as effectively allocate meeting room resources. This smart monitoring system gives them a more accurate picture of how different areas of the building are being used, allowing them to make more informed decisions and increase productivity and employee satisfaction.

Construction and factory managers have been thinking about outcomes like this since the dawn of digitalization, but what is happening now to help them realize their ambitions? What motivates developers to adopt visual sensing solutions so quickly and with such ingenious results?

Utilize efficient, low-power processing technology to process large amounts of data more effectively, and extend applications through artificial intelligence algorithms to achieve ultra-intelligent data processing.

CPUs and Neural Processors

The convergence of efficient CPUs and neural processors with artificial intelligence and machine learning software at the edge is opening up huge new business opportunities.

Surprisingly, it seems too early. I can't help but be reminded of the early days of the mobile phone industry: a rapidly forming ecosystem that enabled greater design flexibility and application development by abstracting software from hardware.

Anyone currently standing on the edge of visionary innovation risks being left behind. This isn't just about missed opportunities.

There is almost no reason not to take the initiative and get to work. Because the tools and processes needed to realize your personal vision are already in place and ready to go.

IoT Vision Sensing Considerations

Connectivity

Through Wi-Fi, Integrating connectivity into IoT devices through protocols such as Bluetooth Low Energy (BLE) has been a key development, similar to the integration of connectivity in smartphones.

Developers are free to choose the right communication protocol for their specific application. For example, smart vision systems in factories might take advantage of Wi-Fi's cost and scalability advantages, while developers building energy-hungry systems might choose BLE.

More far-reaching is the growing adoption of high-bandwidth 5G technology, which promises applications in smart cities. (Indeed, in a recent Arm survey of innovators, nearly half of respondents cited 5G as one of the factors that will have the biggest impact on IoT growth over the next five years).

Security

Security is a key issue in the Internet of Things - devices have been used in this field for many years - especially in imaging data aspect. IoT visual sensing continues to evolve, with challenges addressed through frameworks such as PSA Certified to ensure devices can be maintained and remain secure over the long term.

Machine Learning at the Edge

As more powerful and efficient processing is pushed from the cloud to the edge, machine learning applications are being deployed in new , a fascinating field. They are improving real-time performance and supporting the development of new solutions.

Standards

Common underlying APIs and frameworks (such as Trusted Firmware) enable developers to address core functionality consistently across multiple platforms, thereby promoting innovation and value addition. Thanks to the adoption of standards, fragmentation is becoming a thing of the past.

Seize the market

The journey of vision-based IoT systems from concept to reality has transformed in other ways. A generation of developers has grown up on open tools and platforms, like the Raspberry Pi.

Now, many developers (who first encountered technology like the Raspberry Pi as teenagers) are developing in the professional world. They demand the same easy-to-exploit experiences they had as teenagers.

All of these factors combine to spur innovation in vision-based applications, not only because the processing power and machine learning capabilities are already in place, but because the barriers to design and development are falling.

Imagine what could be achieved by installing an ML-enabled camera at the parking lot entrance (like we have at Arm’s Cambridge office). It can identify all vehicles entering and exiting throughout the day, eliminating the need to install sensors in every parking space within a building.

The capabilities of visual sensing in the Internet of Things have been significantly enhanced, and its diverse applications are truly fascinating. The sudden expansion of IoT capabilities enabled by vision technology is truly remarkable.

Early adopters win hearts and minds, but laggards (those waiting to see how early IoT adoption progresses) still have a huge opportunity to leverage vision technology to transform their businesses. You can see the possibilities. The only thing holding us back now is our imagination.