Research On Earth Observation Coverage For Multiple Unmanned Rotorcraft

For both military and civilian, micro unmanned rotorcraft has a wide range of applications and earth observation is one of these important applications. As the traditional earth observation works, it is inefficient and not a real-time method to scan in lines to cover a target area and rebuild the panorama offline which is impossible to cover in a long period and gives rise to research observing target area with multiple UAVs. There are few applications for this problem which need further research.We solve three problems in this paper:how to build onboard system; how to find the position for the minimal UAVs to accomplish the observation task; how to stitch the captured pictures. It is meaningful to solve these problems for the research on earth observation for UAVs.According to the demand of earth observation problem, we build an onboard computer based on flight control system and flight management system. Flight control system is based on DSP+FPGA and flight management system is based on OMAP3530Beagle Board. We also put forward a strategy for multi-UAVs synchronization mechanism.To solve the problem of UAV placement in earth observation, we abstract the original problem to a problem of covering the polygonal region with rectangles in this paper. We simplify this NP-complete problem by making some assumption and build a basic model to solve where to place UAV in the observation problem. By introducing the overlap percentage of images to the basic model as constrains, we put forward a non-linear model take image stitching into account. With some simulations and experiments, we figure out the advantages and disadvantages of these two models and find some possible reason. We also verify the two models by some simulations and experiments.Furthermore, in order to stitch the captured images of different UAVs, we research the theory and algorithm of image stitching based on ORB. After that, we research the relationship between image stitching and the percentage of overlap. We also give examples in real projects and combine the image stitching algorithm with micro rotorcraft to verify the model for earth observation coverage problem.In a word, to solve the problem of earth observation by rotorcrafts in this paper, we design the platform for rotorcraft; we build models to solve the placement problem for rotorcrafts to meet the demand of monitoring completely; we research the image stitching algorithm based on ORB and verify the performance with the images captured by rotorcraft and solve the image stitching problem.