Original Research Article, B. Miethig and E. Huangfu, June 2021
For autonomous driving, thermal imaging sensors combined with convolutional neural networks can ably detect and classify objects, even when available data is limited.
The field of autonomous driving has made considerable strides in the past decade with the advent of cost-effective sensors and improved computing capabilities. For these systems to be adopted by the masses, they must consistently perform better than human drivers in all driving conditions. While existing vehicle sensors offer important information about a vehicle’s surroundings, there have been several documented cases in which these sensors did not analyze their environment properly, leading to collisions.