基于机器视觉的兰花几何参数测量方法
Method for Measuring Geometric Parameters of Orchid Based on Machine Vision
谢忠坚 1郭佳欣 1刁胤洲 1廖珩宇 1李元航 1文春明 1廖义奎 1辜松2
作者信息
- 1. 广西民族大学电子信息学院,南宁 530006;广西民族大学广西混杂计算与集成电路设计分析重点实验室,南宁 530006
- 2. 华南农业大学工程学院,广州 510642;华南农业大学南方农业机械与装备关键技术省部共建教育部重点实验室,广州 510642
- 折叠
摘要
为实现兰花几何参数快速测量,以"金凤蝶"兰花为研究对象,搭建图像采集平台并设计了一套兰花花朵几何参数测量算法.采用Canny-Harris算法实现兰花角点检测,采用K-means算法、RGB阈值分割算法实现兰花分割,并采用MER算法及遍历算法实现兰花具体几何参数测量.进行兰花算法识别测量试验,结果表明:单株兰花测量仅需10s左右,相较人工测量速率提升了近 60 倍,测量数据可为兰花选育种工作提供判断依据.
Abstract
In order to rapidly measure the geometric parameters of orchid,an image acquisition platform was built and a set of algorithm for measuring geometric parameters of orchid flower was designed with the Golden Phalaenopsis as the re-search object.The algorithm includes Canny-Harris algorithm to detect orchid corner point,K-means algorithm and RGB threshold segmentation algorithm to segment orchid in picture,and MER algorithm and traversal algorithm to meas-ure specific geometric parameter of orchid.Through recognizing and testing orchid with the algorithm,the test results show that it only takes about 10 s to measure a single orchid plant,the algorithm measure rate is nearly 60 times higher than the manual measurement rate,and the measurement data can provide a judgment basis for orchid selection and breeding.
关键词
兰花/视觉测量/几何参数/Canny-Harris/K-meansKey words
orchid/visual measurement/geometric parameter/Canny-Harris/K-means引用本文复制引用
出版年
2025
国家科技期刊平台
NSTL
万方数据