FLIR advances vehicle safety with regional dataset program
Autonomous Vehicle Technology announces 2020 ACES Award Winner in Autonomy | Sensors category
FLIR Systems, Inc. has taken a leadership role in advancing ADAS and AV safety by created the Thermal Imaging Regional Dataset program for machine-learning development. Specific to major cities, the company first released the San Francisco dataset followed by an upgraded Santa Barbara dataset. The datasets enable developers to evolve convolutional neural networks with FLIR’s ADK (Autonomous Developer Kit), which includes a cost-effective thermal camera developed for ADAS and AV testing. San Francisco, in particular, is a significant autonomous vehicle development hub, with the data enabling developers and researchers to create safer vehicle systems. Additional regional datasets will be released as they are developed to provide engineers with expanded object classes, weather, and seasonal driving conditions. The program builds on a free dataset program FLIR launched in 2018 of more than 14,000 annotated thermal images of day and night scenes. The San Francisco thermal dataset features nearly 10,000 annotated thermal images with 181,000 annotations in thermal and the corresponding visible camera images. It includes new variations in weather, including fog and rain, plus additional driving scenes at different hours of the day. With the introduction of the regional dataset program, FLIR also increased the number of annotations to include car, sign, light, people, truck, bus, hydrant, bike, rider, motorcycle, and train classes. Having regional specific datasets enables developers and carmakers to more quickly ramp up and execute thermal imaging testing as part of the ADAS/AV sensor suite. In the near future, FLIR says that thermal imaging data will fused with visible, LiDAR, and radar data, and through machine learning, this will create a more comprehensive, redundant, and safer system for identifying and classifying roadway objects, especially pedestrians and other living things in an autonomous driving mode.
Click to view AVT’s 2020 ACES Award Winners best in class of