Sensor improvements usher cars to safety

By Inayat Khajasha
OmniVision Technologies Inc.

A revolution is taking place in the global automotive market. It eyes small cameras and imaging technology playing an ever-increasing role in vehicle safety, security and driver-assistance features. Such is the impact of imaging technology on the automotive market. It features lane-departure detection and warning systems, blind spot displays, night vision, intelligent airbag deployment, panoramic rear view cameras, adaptive cruise control, rain detection systems and car security, which are being touted as major selling points for security and driver assistance.

Underpinning this video and imaging "revolution" are several important drivers that have been critical in opening up the lucrative automotive market for imaging device and camera module manufacturers. Improvements in component quality and reliability, and lower production costs mean that the use of cameras in automotive applications is no longer limited to high-end luxury cars and SUVs. CMOS image sensors are at the heart of many of the camera modules used in the automotive market. With such a market opportunity, innovation, differentiation and value-add have become important factors in the battle for market share.

This article examines how two technologies are helping the automotive market improve the car's security and assistance and provide benefit for OEMs. It also discusses how they enable new possibilities in the automotive market and further integrate image sensors in cars.

Enabling night vision
OmniVision's Near IR (NIR) capability enables image sensors to see in the dark. CMOS sensors have an inherent degree of IR capability. What the company has done is develop and use proprietary process level enhancements that extend the sensor's spectral light sensitivity up to 1,050nm. As a result, the CMOS sensor becomes more sensitive to low light conditions. When used with low-power LEDs, it allows automotive cameras to see both beyond and outside the range of a vehicle's headlights, and to perform object detection in complete darkness.

Essentially, a color sensor with no IR cut filter can provide images during daylight hours that are almost natural. Keep in mind that no matter how good the process technology is, reproducing natural green, for example, is a challenge when no IR cut filter is used. As soon as natural light levels fall below a pre-determined lux level, the sensor automatically switches to black and white night vision mode (with active illumination). This dual-mode night vision capability is especially useful in driver assistance and safety features such as pedestrian, object and sign detection, and rear view or backup camera applications.

Growing number
A growing number of automotive security applications are also using image sensors (e.g. being "black box" anti-theft camera systems that record video when activated by motion detection around or inside the vehicle). These sensors provide night vision using just a single low-power LED, which has a negligible effect on vehicle battery life. Thus, due to very low current draw by the sensor during power down mode, the system will remain active even when the vehicle is not operated for a lengthy period.

The NIR capability improves functionality for automotive safety and security applications while allowing OEMs and automotive manufacturers to simplify system designs and reduce the overall system BOM. In addition, there are financial benefits to the vehicle owner. In the United States, the development of many automotive security applications that effectively use night vision capabilities is being driven in part by the automotive insurance industry. Vehicles with these systems qualify for lower insurance premiums because they reduce the risk of theft, vandalism and other vehicle-related crimes.

Color recognition
High dynamic range (or HDR) technology is another area of development. Dynamic range refers to the capacity of an imaging system to capture widely differing light levels in a single image and to adjust rapidly to changes in light levels. HDR enables CMOS sensors to function much like the human eye, adapting to rapidly changing light conditions to capture and process image data simultaneously from bright sunlight to dark shadowed regions.

While HDR technology has a very important role to play in the vision systems found on modern vehicles, OmniVision has developed color HDR, which brings accurate color recognition. Using color HDR, a vehicle can now distinguish, for example, between yellow and white lines on the road under extreme contrast conditions. As people start to depend on the reliability and performance of their automotive vision systems, functionality such as accurate color recognition assumes greater importance. Without it, a vision system could end up giving misleading information that might cause extremely dangerous situations.

OmniVision's color HDR sensor, the OV10620, allows cameras to recognize clearly color on traffic signs, lights and road lines and markings. This capability is critical to the proper functioning of vision- and sensing-based automotive driver assistance applications. It works by using an on-chip HDR algorithm processing pipeline. This allows the sensor to switch rapidly to HDR mode (to handle extreme variations of bright and dark conditions within the same scene) and automatically switches back to non-HDR mode when conditions return to normal. Incorporating all of this functionality onto a single-chip removes the need for additional external components, thus lowering BOM, shrinking the camera module size and reducing the time-to-market.

Typical applications for color HDR in the automotive field include backup/rearview cameras, blind spot detection systems, driver/passenger sensing for smart airbag deployment systems. It can also be used in forward looking sensing or visual/display based applications such as headlight dimming/bending, night vision, rain sensing, traffic light monitoring, sign recognition, and lane departure warning and guidance systems.

NIR and HDR technology are just two of the innovations that improve the functionality of the image sensors used in automotive applications. As process level enhancements continue to be implemented and more functionality integrated on-chip, we can expect to see image sensors in more automotive apps in the future.

About the author
Inayat Khajasha is senior marketing manager of worldwide automotive at OmniVision Technologies Inc.