【作者】 张华；

【导师】 姚万业；

【作者基本信息】 华北电力大学 ， 工程硕士（专业学位）， 2018， 硕士

【摘要】 随着科技的进步以及人们生活水平的不断提高,越来越多的服务型室内机器人出现在各行各业,基于机器视觉的室内机器人定位技术成为一大研究热点。基于视觉的室内定位技术是通过摄像机获取目标图像,将机器坐标转换为世界坐标实现目标定位的技术。由于机器坐标在转换为世界坐标时存在误差,从而使定位不准确。针对此问题,研究了一种新型基于视觉的室内定位系统。首先,设计了一种便于机器识别的红外标签,从标签编码设计与电路设计两方面来增加机器人可活动范围。该标签编码简单,易于机器识别可满足较大空间的室内定位。然后,设计了一种基于机器视觉的室内定位方法,此定位方法包括目标识别、目标追踪,定位计算三部分。通过摄像头采集标签图像,然后对目标图像进行图像预处理及目标识别,当摄像头采集到的目标图像偏移图像预设值时,需要对目标追踪,当追踪到目标时通过角度传感器记录下电机转动过的角度,通过AAE法定位出机器人的当前所在位置从而实现机器人定位。本定位系统采用改进随机Hough变换算法识别红外目标;在目标追踪前需要建立追踪系统模型,采用POS法对模型辨识,确定模型参数,从而达到更好的控制效果。最后,对视觉定位系统进行设计与实现,主要包括定位平台的搭建,硬件选型、硬件电路的设计与软件设计。然后在此平台上对定位系统进行了静态测试及动态测试,分析定位系统定位误差与追踪速度,验证定位系统的可行性。更多还原

【Abstract】 With the progress of social science and technology and the continuous improvement of people’s living standards,more and more service-oriented indoor robots appear in all walks of life.Indoor robot positioning technology based on machine vision has become a hot research topic.Based on the vision of indoor positioning technology is through the camera coordinates and world coordinates to determine the location of the label to achieve positioning purposes.Because there is an error when converting to the world coordinate with the machine coordinate,it is easy to cause huge error to the positioning result,so that the positioning is inaccurate.In response to this problem,a new vision-based indoor positioning system is studied.First of all,an infrared tag is designed to increase the range of motion of the robot from two aspects of tag coding design and circuit design.The label coding is simple,easy to identify the machine to meet the larger space of indoor positioning.Then,a indoor positioning method based on machine vision is designed.The positioning method is to capture the label image through the camera and then pre-process and identify the target image.When the target image offset image preset value collected by the camera is offset,It is necessary to track the target.When the target is tracked,the angle sensor records the rotation angle of the motor,and the AAE method locates the robot’s current position so as to realize the robot positioning.This positioning method involves target recognition,target tracking,positioning calculation of three parts.The positioning system uses an improved random hough transform algorithm to identify the infrared target.The tracking system model needs to be established before the target tracking.The POS method is used to identify the model and determine the model parameters so as to achieve better control effect.Finally,the visual positioning system is designed and implemented,including positioning platform,hardware selection,hardware circuit design and software design.Then on this platform,static and dynamic testing of the positioning system was conducted to analyze the positioning error and system feasibility.更多还原