AI could help driverless cars spot objects based on their heat emissionsShutterstock/Ivan Smuk
When it is dark or raining, driverless automobiles may have difficulty distinguishing between a pedestrian and a cardboard cutout of a human. A method that employs artificial intelligence to recognise things based on their heat emission patterns might enable autonomous cars function more securely in all outdoor circumstances.
Zubin Jacob and his colleagues at Purdue University in Indiana created a heat-assisted detection and ranging (HADAR) system by teaching an AI to determine the temperature, signature of energy radiating from objects, and physical texture of such objects for each pixel in thermal images.
The researchers used advanced thermal-imaging cameras and image sensors capable of detecting energy emissions across the electromagnetic spectrum to train the AI. They also developed a computer simulation of outside situations to aid with AI training.
According to Jacob, HADAR learns to recognise objects and estimate their distance with ten times higher precision than previous night-vision technology. For typical object-detection systems, nighttime performance is also comparable to daylight performance.
This HADAR proof-of-concept demonstration is still years away from becoming practical on self-driving vehicles. The large and expensive camera and imaging equipment must still be created in a smaller and less expensive form – the HADAR demonstration tested both a $10,000 thermal-imaging camera and a military-grade hyperspectral imager costing more than $1 million.
Another issue is that the process of gathering and processing data still takes around a minute, although it would ideally be within milliseconds so that a driverless car could use such a system while on the road.
According to Miroslav Pajic of Duke University in North Carolina, the precision and dependability of such a system still need to be verified in a wide range of various contexts. However, he hailed the HADAR concept as a potentially interesting new capacity to supplement self-driving cars’ existing cameras and sensors.
“Having a new way of thinking about the environment, especially in situations where cameras don’t perform well in low light,” explains Pajic.
The device may be more immediately beneficial in aiding nighttime animal monitoring or in future biomedical uses. “I believe we’ll see a lot of breakthroughs on the thermal front in the next five to seven years,” Jacob predicts.
FAQs
1. How does AI-enabled heat vision enhance nighttime driving for driverless cars?
AI-powered HADAR system recognizes objects based on their heat emission patterns, allowing autonomous vehicles to function more safely in dark or adverse weather conditions.
2. What technology does Purdue University’s HADAR system use for object detection?
The HADAR system uses advanced thermal-imaging cameras and image sensors capable of detecting energy emissions across the electromagnetic spectrum.
3. How does HADAR compare to traditional night-vision technology in object detection?
HADAR demonstrates ten times higher precision in object recognition and distance estimation compared to previous night-vision systems, with comparable performance in both daylight and nighttime.
4. Is the HADAR system currently practical for self-driving vehicles?
While the HADAR proof-of-concept is promising, it is still in the developmental stage and requires smaller, more affordable equipment for practical implementation.
5. What are the potential applications of AI-enabled heat vision beyond driverless cars?
HADAR could be beneficial for nighttime animal monitoring, future biomedical uses, and is expected to see significant advancements in thermal imaging technology in the coming years.
