Author:
Yang Yang,Li Zhijun,Wu Yunqiang,Li Ke,Lu Jie
Abstract
Abstract
Regarding the relative detection of the moving direction in vertical between the unmanned aerial vehicle (UAV) and the ground target, we have been inspired by the functional advantages of flying insects in nature, such as lightweight, low computational complexity, low power, and the natural characteristics of high adaptability and high reliability in the detection and tracking or escape process of targets. Based on this, with the help of LGMD (lobula giant movement detector) neuron modeling, this paper proposes a detection model of the moving direction in vertical for UAV ground infrared targets based on LGMD neurons modeling (referred to as LGMD-UAVGIT model). In this paper, experiments and tests are mainly carried out on infrared imaging videos taken by UAVs, thus verifying the effectiveness of the new model proposed in the paper for detecting the two typical vertical movement directions of UAVs that are relatively far away from and close to the ground targets.
Subject
General Physics and Astronomy
Reference9 articles.
1. How fly neurons compute the direction of visual motion[J];Borst;Journal of Comparative Physiology A,2020
2. From the eye to the brain: Development of the drosophila visual system[J];Neriec;Current Topics in Developmental Biology,2016
3. A robust collision perception visual neural network with specific selectivity to darker objects[J];Fu,2019
4. Compound-Eye Imaging Imitation-Based Whole-Field Flow Measurement;Xu;Computers and Electrical Engineering,2021
5. Collision avoidance using a model of the locust LGMD neuron[J];Blanchard;Robotics & Autonomous Systems,2000