News

Samsung Takes On LED Flickering with its First Automotive ISOCELL Image Sensor

July 13, 2021 by Jake Hertz

Today, Samsung announced its first ISOCELL image sensor designed for the automotive space.

In the developing world of autonomous vehicles, one of the hottest debates is what, if any, combination of sensors will provide maximal performance. 

When Tesla chose to use exclusively cameras moving forward, it riled up the industry by foregoing popular options like LiDAR and radar. However, almost every other company feels that cameras alone are insufficient and instead opts to use a combination camera, LiDAR, and radar. 

There is undoubtedly a lot of development that must go into cameras for autonomous vehicle purposes, and Samsung is amongst the many companies working towards this end. Today, it announced its first ISOCELL imaging sensor designed specifically for automotive applications. 

 

Samsung’s new ISOCELL Auto 4AC hopes to improve automotive imaging.

Samsung’s new ISOCELL Auto 4AC hopes to improve automotive imaging. Image used courtesy of Samsung

 

This article will discuss some places where cameras still need improvements for automotive applications and dive deeper into Samsung’s newest release.

 

Places Where Cameras Fall Short 

One of the most significant shortcomings of cameras is how severely they’re affected by environmental changes. 

Notably, a camera’s vision is highly impaired by unavoidable environmental factors such as rain, snow, and dirt. Similar to when a human driver cannot see the road clearly during a heavy rainstorm, cameras lose their vision in these conditions. Thus, when planning on making a fully self-driving car, losing vision is not an option. This setback is one of the reasons engineers chose to incorporate radar in their systems, which has the benefit of immunity to weather. 


Cameras are not perfect for all imaging needs.

Cameras are not perfect for all imaging needs. Image used courtesy of Analog Devices

 

Another unavoidable environmental factor that impacts a camera’s vision is changes in ambient lighting. For example, when a vehicle exits a tunnel, the camera will pick up a sudden change in ambient lighting from total darkness to daylight. This could cause image blur and exposure issues that could confound camera data. Another example would be the reflections of headlights off of street signs which could have a similar effect on the camera. 

With these setbacks in mind, how does Samsung hope to circumvent some of them with its newest sensor?

 

Samsung’s ISOCELL for Automotive

To try to improve some of these shortcomings, Samsung released its first ISOCELL imaging sensor for automotive applications. 

Called the ISOCELL Auto 4AC, the new sensor comes in a 1/3.7-inch optical format offering 1.2 million 2.0 µm pixels. Impressively, the sensor claims to offer a high dynamic range (HDR) or 120 dB along with an embedded image signal processor and integrated LED flicker mitigation to help alleviate issues related to changes in lighting. 

 

Samsung’s new ISOCELL Auto 4AC.

Samsung’s new ISOCELL Auto 4AC. Image used courtesy of Samsung

 

To achieve these feats, Samsung developed a CornerPixel, a specialized pixel structure where there are two embedded photodiodes in a single-pixel area, one 3.0 µm pixel for low light imaging, and one 1.0 µm pixel placed at the corner of the big pixel for viewing brighter environments. The two pixels capture images simultaneously in different exposures, allowing for HDR images, minimal motion blur, and smoother imaging in lighting transitions. 

Samsung states that extending the smaller photodiode's exposure time could help minimize LED flickering. This minimization of flickering prevents pulsing LED light from being displayed as flickering on the camera screen, which results in more accurate image data on LED-embedded objects for the automotive system to recognize. 

 

Moving Foreward

Samsung envisions their new ISOCELL image sensor to be very versatile in automotive applications, including front, surround, and rearview cameras. The company says the sensor has already met AEC-Q100 Grade 2 qualifications and is currently in mass production.

As autonomous driving continues to ramp up, more image sensors are sure to be heading our way. It will be interesting to see the developments and where the industry will be focusing.

 

Featured image [modified] used courtesy of Samsung