Avian eye-inspired visual attention approach to UAV target detection

The purpose of this paper is to propose an avian eye-inspired visual attention approach for unmanned air vehicles (UAV), with the objective of improving the capability of UAV target detection. A visual attention model inspired by the functional architecture of the avian visual system is proposed in this paper. The neuro memetic model employs multiple contrast features as primary channels inspired by the contrast sensitivity of avian retina. The global context strategy is also adopted for focus attention based on the visual mechanisms of tectofugal pathway of avian cortex. Our approach is evaluated through quantitative simulations in various environments. Experimental results demonstrated the feasibility and effectiveness of the proposed method: The proposed method is capable to guarantee efficiency of target localization in the complex environment.