1. The Status Quo and Defects of Autonomous Driving Technology

In recent years, autonomous driving technology has gradually matured in the field of automobile manufacturing. The primary purpose of auto autonomous driving is to improve safety, followed by intelligence. Allegedly, vehicle traffic accidents in the United States is nine times out of ten is due to driver error, which makes driving safety has always been a place that cannot be ignored, the car manufacturers, in particular, if improve the safety of vehicle driving and intelligent, reduce the number of traffic accidents, is need to continue to focus on and solve problems. Self-driving cars are said to save nearly 40,000 lives a year in the US by replacing human drivers.

At present, all self-driving cars adopt the integrated mode of visible light camera, radar technology and laser radar sensor, but still cannot avoid many defects, such as safety problems in extremely complex weather, dark night, dazzling light and haze weather. Under these complex conditions, the visible light camera's ability to detect the road ahead is greatly compromised. Millimeter-wave radar's ability to identify obstacles is still very weak, while ultrasonic radar is beyond its reach. Lidar does have this ability, but the weather has too great an impact on lidar. This shows that even if these three technologies merge, the defects of autonomous driving will still exist. For example, on January 20, 2016, a Tesla car crashed directly into a road sweeper in the Handan section of the Beijing-Hong Kong-Macau Expressway, hebei province. The driver died after the car was damaged on the spot. In the face of overwhelming evidence, Tesla has finally been forced to admit that the car was on autopilot at the time of the crime. On March 22, 2018, an Uber self-driving test car hit and killed a pedestrian while driving at night in Tempe, Arizona. An investigation report released after the accident said that in one case, the Uber driving system had detected the passerby six seconds before the accident, but identified it as an unidentified object, a vehicle and a bicycle.

2. Infrared Thermal Imaging Technology Improves Autopilot Safety

Auto-making systems that "see" pedestrians and respond to them by various standards still face enormous challenges, whether on quiet country roads or in cities where roads are tangled, especially in extreme weather such as fog or sun glare. In such an unusual but real situation, thermal camera can provide the most efficient and rapid classification of potential risks in the near and far distance to assist the vehicle in relative response. For visible light surveillance cameras, it is challenging to classify light sources as insufficient, safe driving at night, sunny glare and extreme weather. Because of the length of the visible light surveillance camera in the electromagnetic wave wavelength wall detected by the thermal camera, this technology does not present the problem of being unable to distinguish the potential road risks at night or during the day. Even vehicles, non-motor vehicles, cyclists, animals and other objects directly in front of the road are able to distinguish clearly.

Whether it's a human eye or a high-definition camera, when night falls, seeing things becomes something you can't quite reach. This is due to the principle of human eye imaging. Then there is an urgent need for autonomous driving to detect pedestrians and animals on roads at night to enhance the safety of mobile travel. The thermal imaging sensor equipped with an identification system can efficiently and stably identify pedestrians and vehicles on the road, give early warning in time, and greatly increase driving safety. Since pedestrian detection is still one of the main concerns of self-driving cars, thermal imaging technology will be the key to promote the development of self-driving car technology. The large-scale application of thermal imaging sensor technology in the automotive field will lay a solid foundation for the future of autonomous vehicles. Until auto makers began installing thermal cameras in self-driving cars in recent years, infrared thermal imaging night vision technology was considered unique for dealing with complex situations, and other technologies such as lidar, radar and vision cameras may have limitations. Such sensors detect heat emitted by objects, so the technology is more effective at detecting obstacles at different depths in the road. Compared with radar and lidar systems, thermal cameras capture higher resolution images and are particularly good at detecting pedestrians. "Pedestrians give off heat. Therefore, in the thermal imaging environment, pedestrians can be easily identified without being affected by complex environmental conditions such as day and night, bright light and shadow." Thermal imaging technology has proved so reliable that it can classify pedestrians, bicycles and vehicles in complete darkness and detect them at nearly four times the range of a standard vehicle headlight. It also classifies targets under other challenging lighting conditions such as fog, smoke, shadow, bad weather and sun glare.

3. Infrared Thermal Imaging Technology to Improve the Intelligence of Automatic Driving

In addition, thermal cameras present redundant but separate data information for visible light, lidar, or radar detection system software. For example, radar detection or lidar data signals from non-motor vehicles can be obscured by vehicle data signals and other objects in the chaotic natural environment. If a pedestrian walks between two vehicles or is partially blocked by a tree branch, it is almost impossible to detect the reflected data signal of the approaching person, and the feedback information will only confuse the intelligent control system operating the vehicle. Thermal cameras, by contrast, can detect the heat of a person or animal through a shaded area, based on the contrast with the surrounding environment. Such unique advantages, combined with deep learning categorization techniques, allow people and animals to be separated from the background, thus ensuring the normal operation of the vehicle. Low light under the standard of visible light cameras also is all about "near infrared cameras, must also be some visible light can operate, and the thermal camera, km and thermal imaging system even in the distance under the condition of poor visibility, also can perceive, detect and identify obstacles, can be applied under the standard of completely not light.

Yoseen infrared thermal imaging camera through fog to find vehicles and pedestrians

Obviously, the thermal imaging device's distinctive working ability can improve the safety coefficient of self-driving cars and the intelligence of autonomous driving, which not only makes up for the defects of the three existing technologies of autonomous driving, but also makes the icing on the cake for intelligent autonomous driving. The scale application of thermal imaging devices in the autonomous driving industry has been a set of thinking, so we are looking forward to the arrival of this day.