基于监督的室内机器人导航系统研究

Indoor Robot Navigation System Based on Supervision

针对室内光照多变对机器视觉带来的不良影响所造成的室内机器人定位难的问题,提出了一种基于监督的室内导航模型。将鱼眼相机固定在室内房顶的中心位置,并且进行多次拍照,拍照后的图片经过图像处理,即可以完成获取整个地图,以及可行驶区域的人工标定。以屋顶固定参照物灯管为中心,结合位于机器人上的移动设备在不同位置向房顶参照物拍摄的图像数据和实际坐标位置,组成了投影轮廓数据库。根据在移动过程中机器人上的移动设备对房顶参照物进行拍摄,并与数据库匹配,实现了当前机器人的位置信息的获取。经测试,本方法在中低精度下具有一定的可行性,并且具有速度快的优点。本次研究实现了在精度要求不高情况下室内的监督导航。

In order to solve the problem that indoor robot is difficult to locate due to the bad influence of indoor illumination change on machine vision,a kind of indoor navigation model based on supervision is proposed.The fish-eye camera is fixed in the center of the indoor roof,and is used to take photos many times.After image processing,the whole map can be obtained,and manual calibration of driving area is realized.With the fixed reference light tube on the roof as the center,the projection profile database is composed of the image data taken from the mobile device on the robot to the reference object on the roof at different positions and the actual coordinate positions.According to the mobile devices on the robot,the roof reference is photographed and matched with the database to achieve the acquisition of the current robot position information.Finally,the method is tested to be feasible in low and medium accuracy and has the advantage of high speed.This research realizes the indoor supervision navigation under low precision requirement.