Real-world challenges do occur from accidental situations that are not expected at the training phase. In this paper, we address a practical multispectral fusion issue as unexpected image contamination in day and night conditions.
Over all days, we successfully captured 50km sequences of synchronized multiple sensors at 25Hz using a fully aligned visible and thermal device, high resolution stereo visible cameras, and high accuracy GPS/IMU inertial navigation system.
Geometrically aligned RGB + Thermal images for pedestrian detection
Multi-sensor dataset for various computer vision tasks
We introduce a novel calibration pattern board for visible and thermal camera calibration.
We introduce a low-cost multicamera synchronization approach.
We provide the first aligned visible/thermal all-day dataset, including various illumination conditions: day, night, sunset, and sunrise. With this dataset, we introduce multi-spectral loop-detector as a baseline.
We propose a multispectral pedestrian dataset which provides well aligned color-thermal image pairs, captured by beam splitter-based special hardware. With this dataset, we introduce multispectral ACF, which is an extension of aggregated channel features (ACF) to simultaneously handle color-thermal image pairs.