Path Planning For Cinematic Drones And Its Applications For Real Scenes

As quadrotor technology rapidly advances and the prices of drones with high-quality cameras go down,drone filming becomes quite popular.While more and more stunning aerial videos appearing online,this creative task remains challenge,as it requires the precise control of the drone’s position and the camera’s orientation simultaneously with good perception of filming aesthetics.And the operator should make sure the movement as smooth as possible during the whole flight.Things become even hard in the case of making a good uncut aerial video with only one-long-shot for the large-scale outdoor scenes.Virtual camera control is one of the classic topic for computer graphics scientists in a very long time.The increasing use of drones in the cinema industry calls for the creation of dedicated tools.Though there is a range of techniques to automatically control drones for a variety of applications,none have considered the problem of producing cinematographic by designing reasonable path.We adapt virtual camera control techniques to drone path planning for creating cinematic content.This work presented a camera path planning algorithm based on cinematic sensation and introduced some application of drone aerial filming.At the meantime,it proposes a higher level tool designed to enable even novice users to easily capture a compelling aerial video of a collection of static landmarks.Some specific application scenarios also have been discussed in detail.In this paper we present our approach to UAV path planning for autonomous cinematography.Using a coarse 2.5D model of a scene,the user is only expected to specify starting and ending viewpoints and designate a set oflandmarks,with or without a particular order.Our system automatically generates a diverse set of candidate local camera moves for observing each landmark,which are collision-free,smooth,and adapted to the shape of the landmark.These moves are guided by a landmark-centric view quality field,which combines visual interest and frame composition.An optimal global camera trajectory is then constructed that chains together a sequence of local camera moves,by choosing one move for each landmark and connecting them with suitable transition trajectories.This task is formulated and solved as an instance of the Set Traveling Salesman Problem.Our trajectory design tool has been tested and performed well in several large-scale outdoor scenes.A detailed comparison was made between the proposed method and other existing path planning products on the market.