园林自动喷药机器人杂草识别与导航方法探究

Investigation of Weed Identification and Navigation Methods for Garden Automatic Spray Robots

园林杂草与景观植物伴随生长,在防控不力的情况下能够很快发展为优势种群,引起自然景观早衰和退化。喷洒除草剂是清除园林杂草的有效方法,但会威胁操作人员的健康。喷药机器人若要准确、高效地完成喷药作业,则必须具备精准喷药和自主导航的功能。为此,将计算机视觉应用在园林喷药机器人上,根据颜色和形状特征识别杂草,根据颜色特征识别路径并规划获得行走路线。试验中,喷药机器人对园林杂草具有较高的识别精度,实际行走路线与规划的路线基本吻合,能够实现精准喷药和自主导航;机器人处理单幅图像仅耗时0. 1 s,可以满足实际作业的需求。

Garden weeds and landscape plants are accompanied by growth,and they can quickly develop into dominant populations in the absence of prevention and control,causing premature aging and degradation of natural landscapes.Spraying herbicides is an effective way to remove garden weeds,but it can threaten the health of operators.Spraying robots must have accurate spray and autonomous navigation functions to complete the spraying operations accurately and efficiently.In this paper,computer vision is applied to garden spray robots to identify weeds based on their color and shape features.At the same time,we identify paths based on color features and plan to obtain walking routes.In the experiment,the spray robot has high recognition accuracy for garden weeds,and the actual walking route is basically consistent with the planned route,enabling accurate spray and autonomous navigation.It takes only 0.1 s for the robot to process a single image,which can meet the needs of actual operations.The level of intelligence and automation has been improved.