Skip to main content

Traffic management is increasingly image conscious

At the Vision show in Stuttgart, Germany, a wide variety of traffic-related solutions were on display. Adam Hill takes the temperature of the industry…
January 27, 2025 Read time: 8 mins
Cameras exhibitors in-person events Germany © Landesmesse Stuttgart GmbH & Co. KG
Vision technology is evolving (© Landesmesse Stuttgart GmbH & Co. KG)

All companies need to adapt if they are to survive in ultra-competitive markets. The vision technology sector is no different – especially when it comes to that subset which is concerned with providing products and solutions to enable smarter traffic management. 

“Smart sensors with edge processing and IoT [Internet of Things] connectivity are the ‘new normal’ for vision technology,” suggests Jan-Erik Schmitt, CEO of Carrida Technologies. “This makes perfect sense if you look at the additional benefits that you can obtain therewith.” Cameras with Carrida software, for example, can identify cars by their licence plate and also recognise make and model – providing high reliability, if registered and recognised vehicle type match. 

“In addition, it enables advanced traffic statistics that can be used to optimise services,” Schmitt continues. “This is even more so, if the traffic sensors are interconnected and if, for example, travel time and traffic routes can be analysed. Of course this requires GDPR-compliant and secure data handling, which is another ‘new normal’. AI, of course, also drives this development towards smarter solutions.”

 IDS NXT Malibu camera

 

‘New normal’

Jan-Friso Blacquière, vice president of sales at Hailo, sees the 'new normal' in the traffic sector characterised by the integration of advanced technologies and AI-driven solutions to improve safety as well as streamline traffic management. “With the rise of smart cities, we see an increasing reliance on cameras and sensors to monitor traffic patterns, detect incidents and accidents, and respond in real-time,” he says. “The analysis of data accumulated from countless sensors and cameras allows for better planning and optimisation of traffic and public transportation management, while real-time insights and alerts allow for immediate resolution of situations that could be dangerous or create congestion.”

For Andrea Van Landingham, Theia Technologies VP business development, larger sensors and larger pixels are becoming the norm. “That makes finding a variety of suitable high- performance, affordable lenses more challenging,” she adds. “When the sensor is larger the image circle of the lens grows, often resulting in lenses larger in diameter and longer, resulting in more material and cost.”

 

“We see larger sensors and larger pixels becoming the norm” Andrea Van Landingham, Theia

 

“Cameras and image processing have long contributed to enhanced safety and planning - not just recently,” says Neil-Steven Sandhu, area sales manager of IDS Imaging Development Systems. “Technologies like AI are now making them even more powerful. For example, real-time traffic analyses can prevent congestion and intelligently manage traffic flow.”

One indicator of the direction of a sector is the type of products and services companies are receiving requests for. Van Landingham is very specific about what Theia’s customers are currently demanding: “100mm telephoto lenses for the larger >1” sensors for ANPR.”

Meanwhile, Hailo provides a portfolio of processors designed to run AI applications and process data - specifically video data - at the edge. “Our customers are particularly interested in solutions that allow for the seamless integration of AI capabilities into smart transportation systems for security, access control, tolling, behaviour analysis and more,” Blacquière says. “Our comprehensive software suite allows applications developed for smart transportation to be easily deployed across various platforms, such as edge boxes, NVRs and security cameras, enhancing flexibility and scalability for our customers. We're seeing heightened interest in solutions that improve real-time analytics and enhance traffic management capabilities, as well as those that drive safety and efficiency in video surveillance and smart city initiatives.”

Cameras with Carrida software can identify cars by their licence plate and recognise make and model

 

Diverse market demands

For Schmitt at Carrida, the picture looks slightly different. “In the past, it was all about parking, where licence plate recognition simplifies booking and management, access control and payment, or smart city applications and law enforcement, where LPR is used to collect data in moving traffic,” he says.

“Recently, we also get many requests from companies in the field of EV charging and smart home automation. They integrate LPR into garage doors, install cameras to optimise the usage of small parking spaces and to manage e-mobility infrastructure. For EV charging applications, LPR also simplifies authentication. Our Plate-i ALPR camera series, with an extremely competitive price, enables such applications and we have received a very positive feedback from the market.”

IDS finds market demands are very diverse, notes Sandhu. “Notably, there is significant demand for embedded systems with light-sensitive, high-resolution image sensors and a high dynamic range,” he explains. “The aim is to consistently obtain high-quality images even under varying lighting conditions. Additionally, more and more application scenarios are emerging for AI-capable cameras - for instance, in traffic monitoring systems for detecting and classifying vehicles and pedestrians. However, there is no ‘one size fits all’ solution. Our team at IDS carefully examines each application and recommends the appropriate vision technology to our customers.”

 

“We're seeing heightened interest in solutions that improve real-time analytics and enhance traffic management capabilities” Jan-Friso Blacquière, Hailo

 

The ability to scan the horizon successfully is another way in which companies maintain relevance – and, crucially, profitability. Artificial intelligence is expected to play a big role in most industries going forward. 

“AI expands the scope and capabilities of industrial cameras, especially when it comes to real-time data analysis for more efficient traffic control,” says Sandhu. “They assist in predicting traffic flows and can ideally prevent incidents before they actually occur. More generally, AI is capable of processing even large volumes of data very efficiently and drawing conclusions, which typically leads to increased efficiency and effectiveness.”

So what are the next big developments in vision tech going to be? For Van Landingham, it is: “Combining sensing technologies and harnessing their data for AI in digital twins and other ways for solutions development.”

There is already a lot underway, says Sandhu, but advanced AI algorithms for predictive traffic analyses and the development of smart city infrastructures with networked vision systems are set to become increasingly important. “Nevertheless, classical image processing still holds a firm place in the traffic sector, with trends moving towards smaller, faster, and more cost-effective solutions,” he adds.

 

Flying the flag for the vision sector in Stuttgart (© Landesmesse Stuttgart GmbH and Co KG)

 

Next big development

AI boosts the development of smart devices, and vision tech provides the necessary data, says Schmitt: “We have been using AI to improve the reading results of our Carrida software and for additional features such as make and model recognition for a while already and the combination is very successful.” There will be further developments, he suggests: “On top, with cameras like our Plate-i series, that provide on-board licence plate recognition at a very competitive price, we’ll see even more appliances with cameras and LPR functionality in the future.”

The next big development in vision technology will be the rise of vision language models (VLMs), which integrate visual understanding with natural language processing, thinks Blacquière. “We believe this technology has the potential to transform how we interact with and interpret visual data,” he says. “As these models advance, we can expect systems to be able to analyse footage and generate summarisation and actionable insights based on both visual data and natural language queries. This will allow systems to not only identify objects but also understand their relationships and contexts within a given environment. Additionally, the integration of VLMs into smart transportation systems could lead to more intuitive interfaces for operators and improved decision-making processes. We see AI as the foundation to all these transformations.”

“Smart sensors with edge processing and IoT connectivity are the new normal for vision technology” Jan-Erik Schmitt, Carrida

 

All of this comes at a cost – and for customers in challenging economic times, price is always an issue. So are there tips and tricks to maximise the value of vision in the traffic sector? “Absolutely,” says Sandhu. “Depending on the application, utilising edge computing is advisable to reduce bandwidth and storage costs. To get the most out of an AI application, it is sensible to bring in the expertise of specialists - these could be research institutions or manufacturers like us. The more precisely a neural network is trained and the better a camera understands what is expected of it, the better the results.”

Van Landingham suggests: “Using varifocal instead of fixed focal length lenses can cover multiple focal lengths with minimal SKUs. Also, minimising set-up and downtime by using motorised lenses for remote focus and zoom. Using lenses with calibration data to combine with sensor level calibrations at the [camera] factory to facilitate quickly finding focus in the field.” 

Carrida LPR software is optimised to run on any embedded hardware and with minimal resources, says Schmitt. “This enables the development of extremely cost-effective solutions, be it for parking applications or LPR in fast-flowing traffic,” he adds. “We have also recently introduced the Plate-i Series of ALPR cameras. With its unmatched entry-level volume price below €500 it enables partners and OEMs to design completely new applications based on licence plate recognition.”

Applying AI analytics at the network edge is financially advantageous, as video data does not need to be streamed to the cloud, points out Blacquière. “Video analytics can take place on the edge, and only insights and metadata can be transmitted to the cloud for long term storage,” he explains. “Using edge AI processors such as the Hailo devices is particularly beneficial, as these processors are designed to run AI on the edge at high efficiency, resulting in cost saving both on compute power as well as power consumption.”

For more information on companies in this article

boombox1
boombox2