Unmanned aerial vehicle target detection method based on event camera

• Main Authors: HU TIANYANG, LI HAOLUN, ZHANG ZHIDONG, YANG GONGLIU, MA BOYI, WU FALIN
• Format: Patent
• Language: Chinese; English
• Published: 07.12.2021
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention discloses an unmanned aerial vehicle target detection method based on event camera, and relates to the field of target detection. The method is mainly applied to rapid target detection of the unmanned aerial vehicle in extreme environments such as poor illumination conditions and fast target movement. Firstly, the event camera is used for carrying out data collection on an unmanned aerial vehicle target, event streams generated by the event camera are preprocessed and converted into image data with the same size as frame images, then event images and the frame images are input into a neural network at the same time for feature extraction, and obtained feature maps are fused and predicted. Loss calculation is carried out according to the prediction result and the real label, network structure parameters are updated through back propagation of the neural network, and a to-be-detected target image is sent into the trained network to obtain a detection result. According to the method, the detection