Auto-driving era quietly come, car camera will be used as the vehicle network information processing an important entrance. According to the IHS survey, the global shipment of in-vehicle cameras will grow from 28 million in 2014 to 82.7 million in 2020, a six-year CAGR of 19.8%. With the advent of fully autonomous driving after 2020, in-vehicle cameras The market will grow geometrically.
Importance of car camera for autopilot
ADAS system solutions include camera solutions, radar / lidar solutions, and sensor fusion. Early stage of market development Because radar technology is mature and weather-free, radar / Lidar solutions are the mainstream of the market. However, with the development of ASICs (ASICs) and the improvement of image processing algorithms and the high accuracy of radar technology in identifying metallic obstacles, it is powerless to identify non-metallic obstacles such as pedestrians and can not be accurately identified Vehicles coming in from the side and can not distinguish lanes, debris or road potholes.
The car camera's visual processing technology can better identify the information such as road signs and pedestrians, and can also calculate the trajectory of pedestrians and vehicles through algorithms. Compared with the radar technology, the technology has the advantages of lower cost, more comprehensive functions and higher accuracy. Camera-based imaging technology is increasingly being accepted by mainstream vendors. Taking into account the limitations of the car camera's pixels for image recognition technology and the reduced functionality in extreme conditions such as foggy weather and rainy days, camera-based sensor fusion will become the mainstream.
Car networking architecture from bottom to top are the perception layer, network layer and application layer, respectively, as the information collection, transmission and processing functions. Video capture storage (perception layer) as the underlying structure of the car networking, the main technology of car DVR and car IP Camera. Car Mobile DVR commonly known as car recorders, is based on digital video compression storage and 3G wireless transmission technology, built-in GPS, car black box, CANbus, G-SENSOR and other technology applications.
The car IP Camera based on digital signal processing (DSP) and network technology, CMOS image sensor to the scene of optical signals into electrical signals, these electrical signals into digital signals transmitted through the data interface to the DSP memory to complete the image compression, encoding At the same time the data stream sent to the hard drive or other storage devices to save. In distance, scalability and cost compared with traditional analog systems and DVR is different.
Car camera has a wide range of applications, according to the application can be divided into driving assistance (driving recorder, ADAS and active safety systems), parking assistance (car around the ring) and the car staff monitoring (face recognition technology) Driving to the parking the entire process, so the camera working time and temperature have higher requirements. According to the installation location can be divided into front view, rear view, side view and the vehicle monitoring 4 parts.
Why choose car camera CMOS technology?
Since the car camera so important, its technology and technology will have what requirements? For automotive applications, automotive camera and mobile phone camera, the main use of CMOS instead of CCD as the optical sensor, the main reason for three points:
First, the primary features that sensors used in active driver assistance systems should have are: fast speed. Especially at high speeds, the system must be able to record critical driving conditions, evaluate the condition and initiate actions in real time. In essence, CMOS is a faster image acquisition technology - cells in a CMOS sensor are typically actively controlled and read by three transistors, which significantly accelerates the image acquisition process. Currently, high-performance CMOS-based cameras can achieve levels of about 5,000 frames per second.
Second, the CMOS sensor also has the advantages of digital image processing. CCD sensors typically provide analog TSC / PAL signals and may have to be converted with additional AD converters or CCD sensors to work with the progressive scan method with digital image output. Either way, providing CCD-enabled cameras with digital video signals significantly increases system complexity; CMOS sensors, on the other hand, provide LVDS or digital output directly, and components within the active driver assistance system are direct and without delay Deal with these signals.
Moreover, in order to achieve such a goal, the car camera manufacturers must consider the use of lower cost CMOS sensor. And, in the presence of strong light, the CMOS sensor does not produce Smear noise when using the CCD. This will reduce the adjustment time due to operational errors.
The advantages and disadvantages of COMS and CCD sensors
What are the characteristics of car camera module?
In addition to using COMS technology for automotive applications, automotive camera modules have other requirements in terms of process and packaging. Compared to mobile phone camera, vehicle camera technology more difficult, mainly due to its high reliability requirements. Different from the average camera, the car camera continuously working longer hours, where the environment is often more vibration and once the failure will be a deadly threat to the safety of users, and therefore strict requirements on the module and package. Car camera tests need to be immersed in water for days, as well as more than 1000 hours of temperature testing, but also from minus 40 degrees to 80 degrees on the rapid jump. And car camera need to have night vision function to ensure normal use at night.
The unique specifications of the vehicle camera module has four main points
(1) Noise suppression during low-light shooting can be suppressed. In particular, it is required to capture the image of the rear side and the side of the vehicle, and the image must be captured easily even at night.
(2) Another feature of the car camera module is the horizontal viewing angle is expanded to 25 ° ~ 135 °. The camera's camera in the phone's horizontal viewing angle is mostly about 55 °. To achieve wide-angle and high resolution around the image, use at least 5 or so lenses.
(3) Car camera module body is made of aluminum alloy die-casting, high material costs. Car head-mounted module does not use the resin and the use of aluminum die-casting products, is to ensure reliability, including the following three reasons: good heat dissipation; the fuselage as a ground layer can inhibit electromagnetic interference; shape thermal stability.
(4) automotive camera module mechanical strength and high temperature resistance is one of the decisive criteria. These modules will be packaged in a special package that will provide the camera with the toughness and penetration resistance required. Since the cameras used for active driver assistance systems are traffic-safe components, they must also work reliably when the power supply system is temporarily powered down.
Due to the vehicle camera for the stability and specifications of the special requirements of the module and packaging requirements higher, in addition to the higher threshold of technology and technology, the car camera into the pre-installed market cycle longer than many other types of cameras, from the design -win to generate revenue for at least a year or more period of time.
What is a car panoramic imaging system?
Car camera for ADAS and automatic driving has an important role, and the application of automotive camera panoramic car imaging system can greatly improve the driving safety and convenience. Panoramic imaging system Chinese can also be called the 360 ° panoramic imaging system, or simply MVCS (MulTI-View Camera System). Panoramic Surround View System provides car drivers with more intuitive driving image information, can quickly and accurately find the vehicle is difficult to be observed near the situation, to achieve a precise driving control, especially for novice drivers, can improve driving safety and Reduce unnecessary scratch.
Panoramic view of the system through the car around the erection of 4 to 8 wide-angle high-sensitivity camera cover the vehicle around all the field of view, through the same time before and after the acquisition of the car around the image by the acquisition unit converted to digital information sent to the video synthesis, processing Components, after the image processing unit distortion reduction → perspective conversion → image splicing → image enhancement into analog signal output to generate a 360-degree body top view, and finally displayed on the center console screen to allow drivers to clearly view the vehicle around the existence of Obstacles and obstacles to understand the relative position and distance to help drivers parked vehicles easily.
While displaying the panorama, you can also display a single view of either side, along with the scale line to accurately locate the location and distance of obstacles. ADAS controls the vehicle body around the vehicle by controlling the body to assist in safe driving, while the panoramic camera system by controlling the vehicle body around the camera to capture the impact of the safe parking. The two systems operate independently and consistently throughout the journey.
Panoramic view of the system will be based on the dynamic trajectory and movement, providing 360-degree vehicle around the screen. Often cost-effective links such as LVDS or Fast Ethernet deploy four to five high-dynamic-range (HDR) 1-megapixel cameras. Video compression is generally used to reduce the required communication bandwidth and reduce wiring requirements (eg, unshielded twisted pair or coaxial cable can be used). Other system requirements include a multiport LVDS or Ethernet switch, a power supply, an integrated DRAM for fast access to external memory, and an embedded flash memory to reduce system cost.
Car camera core technical barriers
Both the panoramic imaging system and ADAS undoubtedly will give drivers a better experience and enhance the safety of vehicles. The panoramic imaging system still faces challenges in image mosaic and video processing, and finally we are here to return to the basic car Camera to the core of the current technical barriers.
Night vision function will become one of the core barriers of the car camera. According to the National Highway Traffic Safety Administration (NHTS), half of all accidents occur while driving at night takes only one quarter of the total road traffic. 70% of the accidents caused by poor night vision. Therefore, automotive cameras must be required to have a strong photosensitivity, making the whole day work, that is, the near-infrared wide spectral range (from 400nm ~ 1100nm), the future of night vision function will become the standard car camera.
Night-vision technology has been put into use in three categories: low-light night vision technology, passive infrared night vision technology, active infrared night vision technology. The glimmer of light uses low-intensity natural light reflected by a nighttime target to magnify it up to hundreds of thousands of times to achieve an image that is suitable for overnight observation by the naked eye. Passive infrared night vision technology is based on receiving the difference between the detected heat source and the background infrared radiation for imaging. Compared with the low light and active infrared technology, the passive infrared night vision technology does not require extra light sources, and the detection distance is the farthest, and the accuracy is high but the imaging is also the most obscure. low. Active infrared technology, also known as near-infrared night vision technology, emits invisible light through an infrared searchlight to illuminate the target and uses reflected light to image with a clear visual distance and clear imaging.
Compared with the passive night vision technology, Active Night Vision technology imaging clearer, you can directly use the image recognition of night road signs, pedestrians to detect, so the active night vision technology more in line with the automotive application scenarios. At the same time, due to the passive infra-red night vision system core infrared focal plane imaging materials, technology was embargoed, so passive infrared night vision technology costs much higher than the active infrared night vision technology.
The core laser night vision technology needs to have a full range of near-infrared, mid-close range laser night vision imaging and processing technology to solve technical problems such as all-weather imaging, bidirectional high-speed moving high-speed focusing and laser speckle elimination and the like, And flashlight effect, technical difficulty, therefore, night vision function will become one of the core barriers of the automotive camera.