The current de-facto Automotive Driver Assist System (ADAS) sensor suite typically comprises mutually dependent visible-light cameras and radar, but when one of these sensors becomes ineffective, so too does the entire sensor suite. This scenario happens often especially when it comes to pedestrians, cyclists, and animals at night or in inclement weather. We will discuss a novel modality known as monocular 3D Thermal Ranging™ that dramatically improves pedestrian safety to reduce accidents and save lives. The solution is based on custom HD thermal imaging and innovative AI based computer vision algorithms. Operating in the thermal spectrum these algorithms exploit angular, temporal and intensity data to produce ultra-dense point clouds (up to 150x that of LIDAR) along with highly refined classification for object detection and identification. We will discuss how to derive ultra-high-density range maps from a monocular thermal camera running a purpose-built AI CNN and bespoke embedded optics. The resulting new sensor modality provides all the benefits of a thermal camera, including all weather and day/night operation and instant detection of animals and vehicles, while simultaneously delivering a geospatially registered 3D range map of such density that perception stacks may enjoy unprecedented awareness. This new sensor maybe an ideal complement to the ADAS & AV sensor suite where thermal perception is sorely needed and the redundancy of real time imaging and ranging channels will be most welcome to improve the utility, comfort, and safety of autonomous and semi-autonomous vehicles.
|