Target Recognition Method Utilizing Super-Resolution Processing for Drone Automatic Flights

Research on autonomous driving for mobile objects and development of applications are widely conducted. GPS can be used to move drones to specified positions, however the accuracy of GPS will cause an error of up to 15m. Furthermore, in order to operate them, it is expensive in terms of camera devices.In this study, we use the total variation regularization method to generate super-resolution images. The aim is to increase the recognition rate even if a marker is far away from the camera device.