Detection model of target motion direction simulating the optic lobe neural network of compound eyes
How to detect and perceive the motion direction of objects(targets)in a chaotic background accurately and reliably is an important research area in the field of computer vision.In nature,flying insects(such as flies and dragon-flies)can sense the movements of their target with high adaptability and reliability.Based on the latest research results of the physiology of the optic lobe neuropile layer of the flying insect drosophila,a multilevel model system for detect-ing target motion direction based on visual perception of drosophila is proposed.By shooting the video sequence of samples under different scenarios for experiment and test and comparing with the 2-Q motion detector model based on ON and OFF signal channel processing motion information,the effectiveness and robustness of the proposed system for horizontal and vertical motion detection of video targets under a chaotic background are confirmed.
万方数据
维普
