摘要
在农田非结构化自然环境下,通过机器视觉提取导航线的处理算法,利用透视投影变换得到摄像机相对于导航路径的偏移距离和偏移角,成功地在TMS320DM642系统上得到实现,并对程序进行了基于C语言的代码优化。试验结果表明,偏移角的最大误差为2°,横向偏移距离最大误差为5 cm。
In the farmland non-structurization natural environment, navigation route＇s extraction algorithm was studied based on machine vision. The distance and angle between camera and navigation route was gained by the perspective projection. The algorithm was successfully transplanted to TMS320DM642 hardware platform, and optimized on a C programming language level. In the examination, it shows that the biggest error of deviation angle between camera and navigation route is 2° and the biggest deviation distance between camera and navigation route is 5 cm.
出处
《农业机械学报》
EI
CAS
CSCD
北大核心
2009年第7期171-175,共5页
Transactions of the Chinese Society for Agricultural Machinery
基金
国家"863"高技术研究发展计划资助项目(2006AA10A304)
关键词
机器人
自定位
视觉导航
农业环境
数字信号处理
Robot, Self-orientation, Visual navigation, Agriculture environment, DSP
