State Estimation of Micro Autonomous Helicopter Based on Computer Vision

In order to achieve autonomous flight, micro helicopter should have the ability to estimate its state information. In this paper, a special target used as a landmark is designed and the corresponding recognition algorithm is developed, which mainly relies on the color feature. Besides, three algorithms of state estimation, linear and nonlinear optimization algorithms based on four planar feature points and data fusion technology are compared and analyzed respectively. Finally data fusion method is chosen for state estimation. Actual flight experiments show our object recognition algorithm is robust, efficient and has higher correct recognition rate (about 99%). Moreover the state estimation algorithm is accurate and simple in actual applications by the fusion of other sensors´ information