Optimize the image sensor platform to meet the requirements of challenging shooting scenes in cars

     Driven by both safety and convenience features, the demand for driver assistance systems is growing, and the number of systems in the car with imaging capabilities is also rising rapidly. These advanced driver assistance systems enable functions such as adaptive cruise control and automatic emergency braking. In addition, this includes parking assistance features such as rear-view cameras (now mandatory in the United States for every new car), 360-degree round-view systems, and emerging applications such as camera monitoring systems that replace traditional rear-view mirrors.

     The operating conditions and parameters in the automotive environment are often very challenging for image sensors. The dynamic range of the scene can exceed 140dB, with a sharp contrast between the brightest areas (such as sunlight directly on the camera's field of view) and the dark areas (inside the tunnel). For this type of scene with a high dynamic range output, it is important to photograph every detail possible in order to present a clear scene for the Advanced Driver Assistance system (ADAS) algorithms and the driver. If imaging technology does not clearly present objects and hazards, it is likely to compromise safety

     The imaging problem is complicated by the recent trend by automobile manufacturers to adopt pulse-width modulated LED lighting in vehicles and the increasing use of the same technology in electronic traffic signs and vehicle communication systems. The reason why this type of pulse lighting system is becoming more and more popular is that it saves electricity, the lighting effect is superior, and the design can be improved. Although this pulse is invisible to the human eye, when shooting with a traditional camera, especially in bright light, because the camera is usually shooting the scene with the light off, the captured pulse light appears to flicker. This can cause problems for the ADAS algorithm, causing it to miss key information on electronic traffic signs and distracting drivers, and it can even cause the ADAS algorithm to confuse the flashing headlights of ordinary vehicles with those of emergency vehicles.

     Therefore, image sensors used in automobiles must solve the flicker problem caused by pulsed LED light sources. However, capturing in the short "on" state of pulsed leds requires longer exposure times, contradicts the high dynamic range performance required to achieve hybrid lighting scenes, and introduces susaturated images in bright areas. To solve this problem and provide a high dynamic range output including a solution to reduce LED flicker, an innovative high charge-capacity CMOS image sensor is required that can be exposed long enough to capture the pulsed light source without saturating the bright areas of the scene. That's what On Semiconductor is doing with its new HayabusaM automotive image sensor platform, which features a "super exposure" feature that allows exposure times more than five times that of conventional automotive image sensors.

     The Hayabusa platform is a family of image sensors with an image cable range of 1Mp to 5Mp for a wide range of applications, using a common image cable, design and real-time functional safety architecture to solve these imaging problems. This platform-based approach has the added advantage of simplifying design, reducing time to market, and optimizing scalability for different models from the same manufacturer. In this way, the development work and data collected by engineers on one sensor can be transplanted to another sensor with a different resolution, without the need for new development, avoiding time-consuming and costly repetitive design work