Avian contrast sensitivity inspired contour detector for unmanned aerial vehicle landing

作     者：DENG YiMin DUAN HaiBin 

作者机构：State Key Laboratory of Virtual Reality Technology and Systems School of Automation Science and Electrical Engineering Beihang University Beijing 100191 China 

出 版 物：《Science China(Technological Sciences)》 (中国科学（技术科学英文版）)

年 卷 期：2017年第60卷第12期

页      面：1958-1965页

核心收录：

学科分类：08[工学] 081105[工学-导航、制导与控制] 082503[工学-航空宇航制造工程] 0825[工学-航空宇航科学与技术] 0811[工学-控制科学与工程] 

基　　金：supported by the National Natural Science Foundation of China(Grant Nos.61333004 61425008&91648205) 

主　　题：contrast sensitivity function surround suppression contour perception unmanned aerial vehicle landing 

摘      要：Runway detection is a demanding task for autonomous landing of unmanned aerial vehicles. Inspired by the attenuation effect and surround suppression mechanism, a novel biologically computational method based on the avian contrast sensitivity is proposed for runway contour detection. For the noisy stimuli, deniosed responses of the biologically inspired Gabor energy operator are generalized followed by the denoising layer and the multiresolution fusion layer. Moreover, two factors such as contour effect and texture suppression are considered in the contrast sensitivity based surround inhibition. Different from traditional detectors, which do not distinguish between contours and texture edges, the proposed method can respond strongly to contours and suppress the texture information. Applying the contrast sensitivity inspired detector to noisy runway scenes yields effective contours, while the non-meaningful texture elements are removed dramatically at the same time. Besides the superior performance over traditional detectors, the proposed method is capable to provide insight into the attenuation effect of the avian contrast sensitivity function and has potential applications in computer vision and pattern recognition.