Developing the GoogleNet neural network for the detection and recognition of unmanned aerial vehicles in the data Fusion System

This work reports a study into the possibility of using GoogleNet neural network in optoelectronic channel Data Fusion system. The search for most accurate algorithms detecting and recognizing unmanned aerial vehicles (UAVs) systems has been carried out. data processing scheme was selected (merging SVF state vectors merging MF measurements), as well sensors recognition models on each model based Kalman Filter chosen, integrating radar channels. Mini-radars LPI-FMCW were used channel. Evaluation effectiveness UAV detection is accuracy. main aimed at determining under conditions different range classes. drones developed transfer training technology. For training, validation, testing network, database built, special application MATLAB environment. capabilities studied 5 variants distance to object. objects Inspire 2, DJI Phantom 4 Pro, F450, DU 1911 UAVs, not included database. accuracy by 98.13 % up m, 94.65 20 92.47 90.28 100 88.76 200 m. average speed this method 0.81 s.