摘要
传统的淡水鱼去头尾作业是通过人工操作实现的,效率低,不能满足现代生产的要求。该研究提出一种视觉引导淡水鱼自动去头尾系统方案,通过计算机视觉技术实现对淡水鱼头、尾的准确定位,并应用可编程逻辑控制器(PLC)实现准确对刀及自动完成去鱼头尾作业;基于图像不变矩理论,设计鱼头鱼尾匹配定位算法。结果表明,采用的匹配判据能有效地区分鱼头(鱼尾)与鱼的其他部分,且检测算法用于鳙鱼鱼头鱼尾定位检测有效率达86%以上。
The traditional operation of removing the head and tail of freshwater fish is achieved by manual operation with a low efficiency, and cannot satisfy the requirement of modern production. Suggest a plan for head and tail auto-removing system to freshwater fish by vis- ual guiding. The position of the head and tail of freshwater fish is lo- cated accurately through the computer vision technology, and accu- rate tool setting and auto completing the operation of removing the fish head and tail is achieved by applying programmable logic control- ler （PLC）; the algorithm of matching localization of the fish head and tail is designed based on the moment invariants of pictures. The study shows that the matching criterion adopted can effectively dis- tinguish between fish head （tail） and the other parts of fish. The ef fective rate of detection algorithm for the detection of the head and tail of the bighead reaches above 86 %.
出处
《食品与机械》
CSCD
北大核心
2014年第5期141-143,共3页
Food and Machinery
基金
湖南省科技厅应用研究项目(编号:2013FJ3101)
关键词
计算机视觉
淡水鱼
不变矩
computer vision
freshwater fish
invariant moment
