The automotive camera landscape has undergone a decade of transformation, driven by breakthroughs in image sensors, optics, system architectures, and AI-powered perception. To reflect on this journey, we spoke with Martin Punke, Head of Camera Product Technology at Aumovio, who shared his perspective on milestones achieved, challenges overcome, and what’s next for automotive vision. See the full Q&A below…
How has the automotive camera landscape evolved over the past decade, and what were the biggest technological milestones?
The automotive camera market has seen numerous changes over the past decade. This is true for the camera technology, the system architectures, and the market itself. Starting with the camera technology we have seen the rise of advanced image sensors that not only increased resolution from 1 to 8Mpix (and beyond) but also improved performance like low-light sensitivity and dynamic range.
In addition, the modern image sensors are full of additional features like ISPs and safety/security functions. Also, in the optics area the technology advanced thus giving us wider field of view lenses with much improved performance like better sharpness, lower F#, low ghosting and increased stability.
Besides the technological changes the system architectures in the car are undergoing significant changes. While 10 years ago most front cameras were so-called smart sensors with integrated processing, the trend is clearly going towards centralized architectures and the usage of more and more satellite cameras.
Furthermore, the market and the market players are constantly changing. A few years ago, the camera market was dominated by traditional Tier 1s doing the camera development while nowadays we have a mixture of traditional and new Tier1s, new SOC platform and CV stack providers and also car manufacturers that develop ADAS systems in-house.
From your perspective, what have been the biggest engineering challenges in developing automotive camera systems over the last decade, and what advancements were key in overcoming them?
There are at least three hardware pillars to this. First the advancement of better image sensors helped to enable sensing in low-light scenarios, with high-resolution and with better overall image quality. Secondly the constant improvement of the optical path was enabling the high-quality and high-resolution imaging we are used today. There was a long and tedious way both in component development but also camera module manufacturing.
The third pillar is the much more powerful processing power enabled by advanced SOCs in modern ADAS systems. These SOCs are also paving the way towards the advancements in computer vision the community has seen over the years. Besides the hardware development certainly the rise of machine learning algorithms has had tremendous effects on the performance of modern camera systems.
How has the role of cameras in ADAS changed with the rise of sensor fusion and AI-driven perception?
Actually, cameras were always one or even the sensor modality that all ADAS systems needed. The reason for this is easy: everything on the street today is made for human perception and a camera sensor is as close as it gets to the human eye. It can “see” high-resolution colour images with a high dynamic range that one needs to detect lane markers, traffic signs and lights, pedestrians, and cars.
Of course, a camera is not perfect so a fusion with other sensor modalities like radar or lidar makes perfect sense to increase the robustness of the system. AI-driven perception has enabled an even better usage of the provided camera images as it helped to “understand” the scene content and context.
As cameras are high-resolution sensors it is very important to “concentrate” on the scene objects that are relevant for the perception and driving task.
Can you share some insights on how regulatory changes have shaped camera technology development?
Regulations have always played a significant role in the advancement and roll out of ADAS camera systems. It started with mandates for backup cameras, followed by emergency brake systems that are often camera-based, and goes now further in ADAS functions for pedestrian detection also at night.
So regulations helped to develop the right camera systems to save human lives. Besides regulation, customer demand for advanced ADAS features and ratings by agencies like NCAP were the main drivers for better camera systems.
As we look to the future, what do you see as the next major technological leap for automotive vision systems? Where do you envision the industry in another 10 years’ time?
There will be still a lot of improvements of the current technology. Higher resolution sensors and more advanced manufacturing technologies will further drive the hardware performance.
In addition, the AI-driven scene understanding will make major steps forward in order to not only see the environment but also fully understand it and derive the right actions out of it. This should enable a “better and safer than human” driving experience.
Most likely there will also be new sensor modalities like event-based imaging, infrared imaging etc., that further enhance detection capabilities and robustness of ADAS systems. All cars in 10 years will be equipped with numerous cameras both for exterior and interior sensing.
AutoSens is celebrating its 10-year anniversary; looking back, what role has this event played in the evolution of camera technology for ADAS?
AutoSens is the place to be for the past 10 years if you want to learn about the latest advancements in sensor technology and connect with engineers ranging from car manufacturers to component suppliers. It helped tremendously to build up a great community that is constantly pushing technology forward.
One great example are the activities around standardization of image quality assessment in the P2020 working group. Here AutoSens played a pivotal role in bringing the right people together and enabling the exchange of knowledge and opinions from all sides of the industry. I am looking forward to a fantastic journey with all peers in the next ten years!
From backup cameras to AI-powered vision systems, the past decade has proven that cameras remain at the core of ADAS. With advances in machine learning, optics, and sensor integration, the next 10 years promise even greater strides toward safer, more intelligent mobility.
AutoSens has been at the heart of this evolution, bringing the ecosystem together for its first decade — and shaping the road ahead for the next one. To read the full interview series and explore the complete ADAS landscape, download your free copy of The ADAS Guide.
Don’t miss Martin’s Opening Keynote at AutoSens Europe 2025 – See the details below…
Wednesday 8th October | 09:00 | Auditorium
This presentation charts the evolution of automotive camera systems, from early 1.3MP sensors to today’s 8.3MP designs with innovations like LED flicker mitigation and smaller pixel sizes. It explores advances in optics, the growing role of sensor modeling, and the challenges of manufacturing high-performance modules. Finally, it reviews market shifts—from ADAS to robotaxis and back to a renewed focus on ADAS and automated driving—offering a clear view of the future of automotive vision technology.
