Introduction
Growing customer preferences for in-vehicle comfort and ease of driving is leading to digitalization of passenger cars. Features such as touch screen infotainment and reverse parking cameras, which were once installed only in premium cars are being widely adopted in mass produced cars. The deployment of vision based ADAS systems is also showing us that passenger cars are heading towards an evolutionary path to autonomous vehicles. LiDAR, RADAR and camera sensors are three important sensors currently being used for object detection in ADAS equipped vehicles. This blog will discuss types of automotive camera sensors, their applications, growth drivers, competitive scenario in the automotive camera sensor market.
Types
In general, based on the technology of camera sensing, camera systems either can be classified as having a single source of vision, i.e., mono-vision or a combination of at least two mono-vision systems to form a stereo vision. Monovision system uses a single sensor to capture length and breath of the image, and is popular and inexpensive way to capture two-dimensional images like traffic signals. Stereo vision system uses two sensors, one for capturing the image (like mono-vision) and the other for capturing the depth information. Stereo vision camera systems are used to capture 3D images and distance information (like RADAR and LIDAR systems).
Traditional camera sensors are suitable for working in visible light, but provides a challenge – to see an object at night. To address this issue, automakers are integrating far-infrared sensors in camera systems to provide heat maps of images, by detecting the temperature differences between the object (for example an individual) and its ambient surroundings. These camera systems are generally employed in night vision systems.
Applications
ADAS features | Premium segment | Mid segment | Economy segment | Sensors used |
ACC | x | x | x | Camera/LiDAR/RADAR |
AEB | x | x | x | Camera/LiDAR/RADAR |
Blind spot | x | x | Camera/RADAR | |
Drowsiness | x | x | Camera | |
Park assist | x | x | x | Ultrasonic/Camera |
Lane control assist | x | x | x | Camera |
Pedestrian | x | x | Camera | |
Traffic sign recognition | x | x | x | Camera |
High auto beam | x | x | Camera | |
Night vision | x | x | Camera/Bolometer |
Park assist applications account for a majority of automotive camera sensors due to low requirements of camera specifications for this function. However, the share of parking assist in total camera sensor installations is expected to decline over the next decade with increasing penetration of other ADAS features (such as lane control assist) using camera sensors.
Drivers
The demand for camera sensors will also be driven by their unique ability to detect colours (to see traffic lights, read stop signs, etc.) which is an indispensable feature of an autonomous vehicle. Lower cost of camera sensors when compared to RADAR and LiDAR makes them ideal for many ADAS applications. Due to their lower cost, camera sensors are increasingly becoming the “base sensor” for many OEMs building their ADAS strategy, and is most likely be the first ADAS sensor being fitted in affordable cars. Automotive camera sensors would remain in high demand for the next decade as level 1 and level 2 ADAS features are increasingly incorporated in affordable cars. Lower price of camera sensors (when compared to LiDAR and RADAR) would lead to their increased penetration in economy cars.
With growing passenger and pedestrian safety concerns, governments are making basic safety features such as rear view camera mandatory in all vehicles. For instance, rear view cameras are mandatory for all new cars sold in the U.S. and Canada. Even developing countries like India are planning to make their installation mandatory in new cars sold. Stringent safety regulations would continue to benefit the sales of camera sensors over the next decade.
The growth of camera sensors will depend upon the advancements in image processing speed. Currently, processing of the captured image takes considerable time which limits the use of cameras in many ADAS applications. Moreover, the advancements in software algorithms will affect the growth of automotive camera sensors. Introduction of potentially “free” algorithms (Tensorflow) from companies like Google are likely to speed up the adoption of camera sensors in vehicles.
Competition
The automotive camera sensor market is consolidated with ON Semiconductors accounting for around 50% of market share, followed by Omnivision, Sony and Toshiba, each accounting for 12-18% of market share. Other important players in the market includes Gentex and Melexis. The competition in the market is increasing with players introducing technologically advanced sensors. For instance, in 2017 Sony launched IMX324 image sensor with 7.42 megapixel resolution, capable of capturing long distance (160m) objects. The low lighting sensitivity of 2666 mV allows the sensor to capture obstacles even in the dark.
Conclusion
In short to mid-term, camera sensors are expected to be widely used as parking sensors due to their increasing penetration in economy cars, especially in emerging economies. Regulatory mandates for installing back cameras will further spur the growth of camera sensors during this period. Declining prices of LiDAR and RADAR, camera sensors would affect the demand of camera sensors in applications such as AEB and ACC. Other applications such as drowsiness detection and night vision are expected to remain niche during this period. With the advancement in sensor technology, cameras will be extensively used in machine vision applications for autonomous cars in the long term. Ability of cameras to accurately capture colours and shapes would make them preferable compared to other sensor types. Also the lower prices will make it the preferred solution for economy cars. Adding to that, there are several countries that are planning to make rear view cameras mandatory and this will further spur the demand of camera sensors in the automotive market.
* Definitions
Premium: Includes passenger cars with MSRP more than $50,000. Examples in this category include the BMW 5 series and the Audi A8
Mid-segment: Includes passenger cars with MSRP $20,000-$50,000. Examples in this category include the Volvo S60 and Mercedes-Benz C Class
Economy: Includes passenger cars with MSRP less than $20,000. Examples in this category include the Nissan Micra and the Honda Fit