Optimizing Image Sensor Platforms to Address Automotive Requirements in the Most Challenging Scenes

by  Bahman Hadji  - 10-25-2017 

With a rise in demand for driver assistance requirements driven by both safety and convenience features, the number of systems on a vehicle with imaging capabilities is increasing at a rapid pace. These advanced driver assistance systems enable features such as adaptive cruise control and automatic emergency braking. This is in addition to parking assistance features such as rear view cameras, now mandatory in the United States on every new vehicle, 360-degree surround view systems, and emerging ones like camera monitor systems for mirror replacement.

Automotive environments often present very challenging operating conditions and parameters for image sensors. Scenes can have dynamic range exceeding 140dB, with extreme contrast between the brightest parts coming from, for example, the sun directly in the field of view of the camera and dark areas inside a tunnel. Capturing the detail in the scene with a high dynamic range output is critical in order to give a good representation of the scene for both the ADAS algorithms as well as the driver. If the imaging technology is not able to clearly show objects and hazards because of these real-world factors, then there may be potential safety consequences.

The imaging problem is further compounded by the recent trend of carmakers to move to pulse-width modulated LED lighting on vehicles and the increasing presence of the same technology for electronic traffic signs and vehicle messaging systems. The gained popularity of these pulsed lighting systems is due to the ability to save power and to implement advantageous lighting effects and enhance design aesthetics. But while human eyes cannot see this pulsing, when captured by a traditional camera, especially in bright environments, these pulsed lights will appear to flicker since the camera will often be capturing the scene when the lights are actually off. This can be problematic for ADAS algorithms which may miss critical info on an electronic traffic sign and pose a distraction to drivers – and may even cause them to confuse a vehicle’s flickering headlights with those of an emergency vehicle.

It is therefore essential that image sensors used in automotive applications address the issue of flicker caused by pulsed LED sources. However, the longer exposure times needed to ensure the capture of the pulsed LED during its brief ‘on’ state would be in conflict with the achievement of good high dynamic range performance required for mixed lighting scenes, and introduce oversaturated images in bright areas. To tackle this, and provide a solution that can provide high dynamic range output with LED flicker mitigation, innovative high-charge capacity CMOS image sensors which can expose long enough to capture the pulsed light source without saturating bright areas of a scene are needed. This is exactly what ON Semiconductor is introducing with the new Hayabusa™ platform of automotive image sensors, which can deliver a ‘Super-Exposure’ to enable exposure times more than five times as long as that of a traditional automotive image sensor.

The Hayabusa platform is an image sensor family ranging from 1Mp to 5Mp  to serve a variety of applications, with a common pixel, design, and real-time functional safety architecture which addresses the imaging problems outlined above. An additional benefit of the platform-based approach is that this can simplify design, speed time-to-market, and optimize scalability across different vehicle models offered by a manufacturer. This allows engineers to leverage development work and data collection done on one sensor and port them over to another of a different resolution without requiring an entirely new development, and avoid repeating time- and cost-consuming design work.

 

Tags:Automotive, CMOS, Image Sensor, CMOS Image Sensor
What's your reaction?
0 0 0 0 0

Post a Comment

Your email address will not be published. Required fields are marked *

Name*

Email*

Comment

Rules for Commenting

Comment

Craig 2017-10-29 00:00 Comment rating: Vote down 0 Vote up
all on one sensor - excellent
Your request has been submitted for approval.
Please allow 2-5 business days for a response.
You will receive an email when your request is approved.
Request for this document already exists and is waiting for approval.