Insect-inspired holistic visual navigation methods for navigation of wheeled and flying robots in outdoor environments

受昆虫启发的视觉机器人整体导航方法是普遍存在的基于特征的方法的一种新兴的替代方法。整体方法不是提取和匹配局部特征描述符，而是将整个图像相互关联。整体方法保证了较低的计算量和在不同环境下的良好导航性能。该项目的目标是大幅推进整体方法，克服当前的局限性，特别是通过增加可处理的自由度，通过改善光照耐受性，以及通过开发能够在复杂深度结构的环境中操作的方法。由于这三个主要要求在室外环境中特别强烈，因此整体方法的适用性将在室外机器人实验中得到演示，无论是在执行检查或监视任务的飞行机器人(多翼机器人)上，还是在轮式室外机器人上，在割草机场景中。

Insect-inspired holistic methods for visual robot navigation are anemerging alternative to the ubiquitous feature-based methods. Rather than extracting and matching local feature descriptors, holistic methods interrelate entire images with each other. The holistic approach promises low computational effort and good navigation performance over different environments. The goal of this project isto substantially advance holistic methods and overcome currentlimitations, especially by increasing the degrees of freedom which can be handled, by improving illumination tolerance, and by developing methods capable of operating in environments with complex depthstructure. Since these three main requirements are particularly strong in outdoor environments, the applicability of holistic methods will be demonstrated in outdoor robot experiments, both on a flying robot (multicopter) in an inspection or surveillance task and on a wheeledoutdoor robot in a lawn mower scenario.