Intelligent vehicle revs up demand of automotive cameras, single cameras dominated

2018-03-19  |  By Irene  |  In knowledge

In autonomous driving areas, ADAS help vehicle sense the surroundings and guide them to take corresponding action. ADAS collectively refers to multiple subsystems, whose operation process is sensing, analyzing and decision-making, action-making.

Camera configuration of ADAS and subsystems

(Source: TRI, TrendForce)

In terms of sensing, different subsystem utilizes different sensor solution. Vehicle-mounted camera, MMW radar and ultrasonic radar are usually equipped for sensing external environment. MMW radar is more resistant to fog, rain, snow and other natural conditions, and it is the mainstreamed sensing solution at the beginning of ADAS’ entering into before-market. But MMW radar is not capable of organizing pedestrian, sensing vehicle-side and detecting traffic lights.

Meanwhile, because of the improvement of image recognition algorithm and image processing ability, sensor fusion that integrates camera and MMW radar is gradually used by automakers.

A mass-produced model is usually equipped with four to eight cameras, and the number will increase with the demand of more intelligent functions to be achieved. Several countries including USA, Europe and Japan have listed subsystems such as FCWLDWAEBBSD into new vehicles evaluation index, or issued related regulations to encourage automakers to install above subsystems, which revs up the penetration of automotive cameras as well.


Front view camera is the main part to sense the surroundings; single camera is the mainstream

Depending on the installed location, automotive camera can be divided into front view camera, rear view camera, surround view camera and inside view camera. Currently, front view camera is utilized most frequently in ADAS. For example, functions of FCW, AEB, LDW, PCW, ACC, TSR and LKS are achieved based on analyzing information that front view camera senses.

Front view camera adopts single camera system, dual camera system and multiple camera system. Mobileye, a camera-based ADAS solution supplier, is committed to single camera system before 2016. Single camera detects objects and calculates the distance between vehicle and barrier by the sizes of objects, but it has limitations in detecting three-dimensional and incomplete object. Therefore, since 2016, more and more models adopt dual camera system, like BMW 7 Series, 5 Series, and Mercedes Benz S Class, and test models from Volkswagen, Toyota and Nissan are tested based on dual camera system as well.

Dual camera system imitates human visual system with two cameras. Therefore, the measurements of two cameras must be the same. Otherwise, the measurement accuracy will  be reduced. Currently, if measurement error of two cameras is no more than 1%, camera makers and automakers will challenge high cost and high product defective rate.

Multiple camera system was proposed to solve the problem between sensing view angle and the distance that when the camera’s view angle is larger, the sensing distance will be shorter. But multiple camera system is more difficult in solving image; its cost is higher than single camera; and there is more requirements in both chip and hardware reliability. Therefore, single camera is still the mainstream of the automotive camera solution in the short term.


High technical threshold of aspheric lenses

Compared with cell phone camera, on-board cameras should feature much higher reliability and stability, including heat dissipation of the fuselage, electromagnetic interference resistance, waterproof and dustproof, seismic resistance, etc. Also, rigid requirement that car cameras should keep working even when electricity is out is put forward as cameras are an essential part of the safety system.

Aspheric lens are more preferable for car cameras thanks to better transmittance, tolerance resistance and wider operating temperature.


There are three processing technologies of aspheric lens.

1. High precision processing, including Single Point Diamond Turning (SPDT), diamond grinding, electropolishing, etc. The technology is mainly used for the manufacture of ultra-precision lenses.

2. Optical machining fabrication, including traditional lapping and polishing, CNC machining, etc.

Both of the technologies above require manual grinding, therefore, their production time is longer and the cost tends to be higher as well.

3. Molding technology is the main technology to produce aspheric lens. Though this processing technology also requires manual grinding, it’s more economic efficient with simpler procedure, however, the threshold for molding is high.


Currently, not many camera makers are able to master the stated processing technologies.

Assahi, Hoya, Opti-asia, Calin, even leading automotive camera maker Sunny Optical purchase aspheric lenses from abroad.

Foxconn Interconncect Technology, a subsidiary of Hon Hai / Foxconn Technology Group, recently announced partnership with Sharp to establish a joint venture with 3.02 billion yen for the development, design, manufacture and marketing of automotive cameras. Kantatsu, a subsidiary of Sharp, provides camera technologies, while the Hon Hai-backed optical molding company Zhongyang Guangdian and optical lens provider Hongli Guangdian will strengthen Hon Hai Group’s deployment in the automotive camera field. However, whether Zhongyang Guangdian will master the processing techniques of aspheric lens, and whether Hongli Guangdian could overcome the technical threshold of automotive grade camera and enter into the before market remain questions.