ResearchInChina: Automotive Thermal Vision is of Little Value

ResearchInChina publishes a report "Global and China Automotive Night Vision System Industry Report, 2019-2025." Few quotes:

"For the automotive sector, night vision system is of little value and seems like “chicken ribs” – tasteless when eaten but a pity to throw away.

In function, night vision system is a special solution for automobiles now that it enables a vehicle to see an object more than 300m ahead at night (compared with a mere 80m offered by headlamps) and gives driver more time to react, ensuring safer driving. ADAS and other technologies (like LiDAR and ordinary optical camera), however, play a part in night driving safety as well. And the stubbornly high price justifies the sluggish demand for night vision systems such as infrared night vision system.

According to the statistics, night vision system was a standard configuration for 58 of vehicle models available on the Chinese market in March 2019, just less than in 2015, of which 18 were Savana (caravans). Audi, Mercedes-Benz and BMW are less enthusiastic about the technology, and just equip it to their luxury models each priced above RMB1 million (a combined 67% of models carrying the system).

In the meantime, the insiders hold such different views on night vision system as follows:

Negative:

“It’s not something that’s really necessary because optical cameras actually do pretty well at night and you have a radar system as backup that is not affected by light,” said Dan Galves, a senior vice president at Intel Corporation’s Mobileye.

Bosch argues that technical advances bring about the decreasing demand for night vision system. One reason is that ordinary camera alone can work outstandingly at night with the maturity of image sensing technology. Also, the progress in technologies for automotive lighting, like LED headlamp, offers a horizon as long as 100-200m. So Bosch has shifted its attention away from night vision solution.

Positive:

Tim LeBeau, the vice president of Seek Thermal, thinks that the current optical radar for autonomous cars cannot detect the heat of an object to ensure whether it is a creature or not, and that the cost of thermal sensors is slashed by about 20 percent a year as they get widely used.

People who detest high beam agree that headlamps delivering 200m beam will interfere with other drivers’ sight, and the solution combining low beam and passive night vision (infrared thermal image) system is the best choice.

Still, some vendors are sparing no efforts in making the technology more feasible for automotive application. Examples include Veoneer whose third-generation night vision system capable of detecting both pedestrians and animals is integrated with rotary LED headlamps which will automatically turn to the front object detected by the system; and Adasky’s Viper system that can classify the obstacles through convolutional neural network-based unique algorithms and display them on the cockpit screen to remind the driver.

Vendors will also work on laser-based night vision, low-light-level night vision, bionic night vision and head-up display (HUD) as well as headlamp fusion.

In brief, as long as price comes down to an affordable level, “the chicken ribs” will become “a delicious homely dish.”"