摘要
为了准确引导机械臂接近苹果目标,避免机器视觉研究中复杂的立体匹配过程,在识别苹果的基础上,提出了一种基于激光视觉系统的苹果空间定位方法。根据该系统的构成、扫描及运动控制特点,建立了激光视觉系统坐标系,用于表示目标的三维空间信息;进行了图像坐标系、视觉系统坐标系及机械臂坐标系间的转换,用于为机械臂的运动控制提供指向性数据。通过对果园环境下苹果的抓取实验可知:当特征参数准确提取率RRC大于90%时,机械臂的准确抓取率为100%。
In order to guide robot arm to be close to apple target accurately,and avoid complex stereo matching of machine vision,a fruit locating method based on laser vision system was presented.According to the features of the system structure,scanning and motion controlling,a coordinate system of laser vision system was established,which was used to represent 3-D spatial information of target.Three different kinds of coordinate systems were transformed for offering directional data to robot arm,which separately belonged to image,laser vision system and robot arm.The grabbing experiment in orchard showed that the accurate grabbing ratio of robot arm was 100% while recognition index RRC was more than 90%.
出处
《农业机械学报》
EI
CAS
CSCD
北大核心
2013年第S1期236-239,共4页
Transactions of the Chinese Society for Agricultural Machinery
基金
国家自然科学基金资助项目(31071333)
保定市科学技术研究与发展计划资助项目(12ZG011)
中国农业大学博士创新基金资助项目(2013YJ008)
关键词
苹果
激光视觉系统
机械臂
坐标转换
空间定位
Apple Laser vision system Robot arm Coordinate transformation Spatial location
