基于视觉算法的割草机器人作业系统优化

Optimization of Mowing Robot Operation System Based on Vision Algorithm

以割草机器人作业系统为研究对象,通过对割草机器人作业过程进行分析,利用高清摄像头进行目标区域草地图像采集,采用视觉算法进行草地图像转换,获取草地图像的颜色和边界特征集合,实现草地识别和草地形状识别,并根据结果进行割草作业自动化控制。试验表明:基于视觉算法建立的割草机器人作业控制系统,草地面积识别误差不大于1.5%,草地图像像素识别误差不大于3%,系统在进行草地状态识别时具有较高的可靠性。

This paper takes the mowing robot operation system as the research object,through the analysis of the operation process of the mowing robot,uses the high-definition camera to collect the grassland image in the target area,uses the visual algorithm to convert the grassland image,obtains the color and boundary feature set of the grassland image,realizes the grassland recognition and the grassland shape recognition,and the mowing robot control system according to the visual algorithm recognition results for automatic control of mowing operation.The experimental results show that the error of grassland area recognition is less than 1.5%,and the error of grassland image pixel recognition is less than 3%,which indicates that the system has high reliability in grassland state recognition.