Drones provide excellent chances to acquire images from points of views that are otherwise impossible or very difficult and costly to achieve. This advantage can successfully be exploited in numerous civil applications within industrial inspection precision agriculture disaster prevention and relieve safety and security and environment protection. This vast amount of images needs to be automatically and quickly processed and interpreted for its optimal exploitation so that here is where new DL technologies play an essential role. This talk presents several examples on ML technologies applied to aerial image recognition in industrial inspection and precision agriculture as examples for its increasing potential usage. Finally reinforcement technologies are presented for autonomous drone navigation in dynamics environments in tasks as landing on a moving platform and dynamic obstacle avoidance.