Thermal vision with thermal vision cameras could improve autonomous vehicle driving at night.
Self-driving cars can have trouble distinguishing between a person walking and a person’s cardboard cutout when it’s dark, and especially at night Rain. A system that uses artificial intelligence to identify objects based on their thermal patterns could help autonomous vehicles operate more safely in all outdoor conditions.
researchers of Purdue University in Indiana developed a thermal vision and detection (HADAR) system by training an artificial intelligence to determine the temperature, energy signature and physical texture of such objects for each pixel in the thermal images.
To train the artificial intelligence, the researchers collected data outdoors at night using sophisticated thermal imaging cameras and imaging sensors that can show emissions across the electromagnetic spectrum. They also created a computer simulation of outdoor environments to provide additional training for the AI.
HADAR learned to recognize objects and estimate the distance to those objects with 10 times more accuracy than traditional ones night vision technologies, so the research. Nighttime performance is also equivalent to daytime performance for traditional object detection systems.
This HADAR proof of concept is still years away from being operational in self-driving vehicles. The bulky and expensive camera and imaging equipment has yet to be manufactured in a smaller form and at a much lower price – the HADAR demonstration tested both a $10,000 thermal imager and a military-grade hyperspectral imager costing more than $1 million costs dollars.
Another challenge is that the process of data collection and processing still takes about a minute, while this period should ideally be within milliseconds for a self-driving car to use such a system.
The accuracy and reliability of such a system has yet to be demonstrated in a wide variety of environments. The HADAR concept is a potentially promising new capability to complement the existing cameras and sensors in self-driving cars.
The technology could immediately prove useful to the wildlife in the Night to monitor or in future biomedical applications.