Evaluating faster-RCNN and YOLOv3 for target detection in multi-sensor data

Intelligent and autonomous systems like driverless cars are seeking the capability to navigatearound at any time of the day and night. Therefore, it is vital to have the capability to reliablydetect objects to predict any situation. One way to capture such imagery is through multisensordata like FLIR (Forward Looking Infrared) and visible cameras. Contemporary deep objectdetectors like YOLOv3 (You Look Once Only) [1] and Faster-RCNN (Faster Region basedConvolu- tional Neural Networks) [2] are well-trained for daytime images. However, noperformance evaluation is available against multi-sensor data. In this paper, we argue thatdiverse contextual multi-sensor data and transform learning can optimise the performance ofdeep object detectors to detect objects around the clock. We explore how contextual multisensor data can play a pivotal role in modelling and recognizing objects especially at night. Forthis purpose, we have proposed the use of contextual data fusion on available training databefore training these deep detectors. We show that such enhancement significantly increasesthe performance of deep learning based object detectors