TWI783748B

This invention discloses a method and a system of image recognition for an obstacle avoidance flight control for an unmanned aerial vehicle (UAV) by applying the deep learning and applications thereof. The system includes a UAV, a wireless communication unit and an information processing unit. The UAV includes a flight control module and an image capturing device. The flight control module controls the flight of the UAV according to a flight path. The image capturing device continuously captures flight status of the UAV and images it as a flight image. The information processing unit receives the flight image through the wireless communication unit. The information processing unit contains a deep learning algorithm module and a feature database with an object feature sample built in. Each object feature sample is defined with an object name. The deep learning algorithm module is configured to extract the object feature for each flight image and input it into the feature database to predict a probability of ma