摘要
因地面含有丰富信息,常用来为室内移动机器人提供地图创建与导航的环境信息。考虑到光线反射对地面造成的干扰较强,在相似颜色环境下地面区分的难度较大,因此将高强光反射区定义为"缺陷"进行检测。利用其周边信息填充缺陷,有效增强了地面颜色的统一性。结合HSV联合密度进行彩色分割,利用地面位置区域特性,可准确获得地面与障碍物间的分割。试验表明,提出的方法具有运算简单、范围广、准确度高、便于机器人实时避障等优点。
The ground is usually used to provide environmental information of map creation and navigation for indoor mobile robots because it contains rich information. Considering the strong interference caused by light reflection, it is difficult to distinguish the ground surface under similar color environment, so the high intensity light reflection areas are defined as"defect"to be detected. By filling defect with its periphery information,the ground color uniformity can be effectively enhanced. Combining with the HSV joint density,color segmentation is conducted,and using regional characteristics of ground position, the segmentation of obstacle with ground is obtained precisely. Experiments show that the proposed approach features simple operation,wide range,high precision,and ease to implement obstacle avoidance for robot in real time.
出处
《自动化仪表》
CAS
2016年第10期13-15,共3页
Process Automation Instrumentation
基金
四川省教育厅重点基金资助项目(编号:14ZA0096)
四川省科技支撑计划基金资助项目(编号:2015GZ0035)
四川省科技创新苗子工程基金资助项目(编号:2015024)
四川省重点实验室开放基金资助项目(编号:13zxtk05)
西南科技大学研究生创新基金资助项目(编号:15ycx119)
西南科技大学创新团队基金资助项目(编号:14tdtk01)
关键词
图像处理
图像分割
图像识别
地面分割
颜色模型
缺陷检测
缺陷填充
彩色增强
阈值分割
二值化
联合概率
Image processing Image segmentation Image recognition Ground segmentation Color model Defect detection Defect fillColor enhancement Threshold segmentation Binarization Joint probability