Technical Analysis of Image Sensing in Automotive Domain
Increasing the safety of the driver and the car is the main focus of the automotive sector. Manufacturers are planning to incorporate image sensing or digital imaging in various advanced driver assistance systems (ADAS). Two type of image sensing are currently available in the market i.e. CCD image sensor and CMOS image sensor. CMOS image sensors is widely accepted in the market and many automotive OEM’s has progressive R&D strategy for CMOS image sensing. Applications powered by image sensing and advanced driver assistance systems are capable of pedestrian detection, adaptive cruise control, park assist, blind spot detection, and lane departure warning.
The key forces of transformation for the image sensors in the automotive industry are as follows:
Time-of-flight image sensors may fail or become inefficient in distinguishing the correlated light (illumination) and non-correlated light (sunlight and other sources), leading to an error in calculating the distance of an object or recording the correct depth information of an object. 3D imaging sensors are needed that can provide in-depth information about the surroundings.
The sensing accuracy of conventional complementary metal-oxide-semiconductor (CMOS) image sensors installed in the vehicle is hindered when recording LED traffic lights and signs because of LED flickering. Also, such sensors may be unable to record high-precision images in bright and low-light conditions, creating demand for high-resolution image sensors with efficient electric shutter mechanism in the automotive sector.
Key automakers offer thermal infrared cameras to enhance drivers’ night vision; however, system-on-chip image sensing systems can enable improved processing so the sensors can work better alongside self-driving capabilities and allow for improved obstacle avoidance during compromised lighting conditions.
High Dynamic Range
Cameras installed to monitor traffic conditions need to overcome various challenges such as varying lighting conditions that range from dark shadows to bright scenes. Surroundings containing bright and dark areas are common in outdoor environments; therefore, the image sensor needs to capture both the dark and bright areas without underexposing or saturating any pixels.
North America: Research in North America is focused on developing a system to track occupants in a fixed environment and under variable illumination; a mirror-less vehicles with image sensors mounted on windows; intensity variation in headlamps; forward-looking imaging systems; driver drowsiness detection; steering control apparatus; and a back-gate-modulated image sensor (which can provide benefits such as high fill factor). Some companies with a North American presence and with patents include Schott AG, Innova Electronics Inc., Rockwell International Corporation, Aptina Imaging Corporation, and Bluetechnix Group GmbH.
Europe: Research in Europe is focused on developing a system to detect white lines on roads, display and 3-D measurement of obstacles, ADAS, lane recognition, fingerprint recognition, and multi-beam occupant detection. Some companies with European patents include Ford Global Tech LLC, DENSO Corporation, Nissan Motor, Ford Motor Corporation, Delphi Technology Inc., General Motors, and Texas Instruments.
Asia-Pacific: Research in Asia-Pacific is focused on developing a tear line for the interior trim, adaptive cruise system, object detection, and a dynamic range display for rear-view images and parking. Some companies with Asia-Pacific patents include Robert Bosch GmbH, Tekville.com Inc., SCHOTT AG, Fuji Photo Film Co., Ltd., Sumitomo Precision Products Co., Ltd., and Great Wall Motor Company Limited.
Europe is leading the patent filing and innovation process. As the home for key automakers, Europe has witnessed strong traction for CMOS. Companies are more focused on developing a level 5 self-driving vehicle where full automation is achieved; therefore, the European automotive sector is experiencing intense R&D. Image sensors are integral to automated driver assistance systems (ADAS).
Key Innovations in Image sensing:
FLIR Systems, Inc., USA, has collaborated with US-based Movidius to develop the next-generation thermal imaging device, the Boson camera core. This device implements a Myriad 2 vision processing unit, along with its expandable infrared video processing architecture, XIR. The device embodies a cost-efficient and simple thermal imaging system.
New Imaging Technologies, France, specializes in image sensors with a high-dynamic range. The company developed the NSC1105 sensor, which is suitable for the difficult traffic conditions. The sensor has a wide dynamic range of over 140 decibels (dB), allowing it to capture extreme lighting differences in the same scene.
BMW, Germany, has developed a camera system to replace the rear-view mirror and side mirrors. The company provides 4 different cameras which includes 2 on the side windows and 2 on the rear window. The camera’s output is shown on the LCD, providing a wide field panoramic view.
Toshiba America Electronic Components Inc., USA, has developed a 2-megapixel CMOS image sensor, the CSA02M00PB, for automotive cameras. The sensor minimizes image flickering caused by LED lights and provides clear and accurate images at a faster rate.
Mobileye, Israel, has developed an advanced collision avoidance system, the Shield+™ for trucks, buses, and commercial vehicles. The system utilizes up to 4 multi-vision smart cameras with high dynamic range CMOS technology to avoid and mitigate imminent collisions. Images are processed in the system with the help of the image recognition software, and unnecessary warnings that can desensitize drivers over time are reduced.
As cameras became a key component for increasing car safety, driving assistance, and driving comfort, the market share for image sensors in the automotive industry is increasing. Blind-spot viewing, lane-departure warning, park assist, and driver drowsiness monitoring are some of the applications for imaging in automobiles. The research community is continuously exploring new ways to overcome technical shortcomings and enhance the speed, resolution, and image quality of image sensing systems. Additionally, existing materials have been tuned to improve the performance of sensors.