摘要
针对鸽子种蛋人工挑选效率低、主观性大的问题,研究了基于机器视觉的鸽子种蛋识别技术并设计出鸽子种蛋识别装置。采用工业相机完成鸽子种蛋图像的采集,运用灰度化、二值化等算法提取了所采集的鸽子种蛋图像中鸽子蛋区域以获得鸽子蛋的蛋重和蛋形指数信息。分析了鸽子蛋受精特征的特点,对鸽子种蛋区域图像进行灰度转换、二值化和形态学膨胀处理以获得包含受精特征的区域,提出了利用绿色通道图像增强受精特征的方法,获取红、绿、蓝三通道的受精特征区域图像,将红色通道受精特征区域图像与其余两通道图像对比获得受精特征,再结合受精特征的区域大小进行受精判断。实验结果表明该方法能够识别鸽子蛋的受精特征,综合蛋重、蛋形指数和受精情况3项数据进行种蛋挑选。
Aiming at the problems of low efficiency and large subjectivity in manual selection of pigeon eggs,the technology of pigeon egg identification based on machine vision is studied and the device of pigeon egg identification is designed.The images of pigeon eggs are collected by industrial camera.In order to obtain the information of egg weight and egg shape index of pigeon eggs,grayscale and binarization algorithms are used to extract the region of pigeon egg images collected.By analyzing the characteristics of pigeon eggs fertilization,grayscale conversion,binarization and morphological expansion are performed to obtain the region containing fertilization characteristics.A method of enhancing fertilization characteristics by green channel image is proposed to obtain the fertilization feature region images of red,green and blue channels.Fertilization characteristics are obtained by comparing the fertilization feature region image of red channel with the other two channel images.And fertilization judgment is made based on the area size of fertilization characteristics.The experimental results show that this method can identify the fertilization characteristics of pigeon eggs.The hatching eggs are chosen by comparing egg weight,egg shape index and fertilization condition.
作者
冯燕柱
蔡肯
宋安亮
林钦永
铁风莲
Feng Yanzhu;Cai Ken;Song Anliang;Lin Qinyong;Tie Fenglian(School of Automation,Zhongkai University of Agriculture and Engineering,Guangzhou 510550,China)
出处
《机电工程技术》
2023年第11期183-185,212,共4页
Mechanical & Electrical Engineering Technology
关键词
机器视觉
鸽子种蛋
自动识别
machine vision
pigeon hatching egg
automatic identification
