基于图像处理技术的受精蛋挑选方法研究  

The Research on Selection Method of Fertilized Eggs Based on Image Processing Technology

作　　者：何海燕 邓国浩 胡鹏飞 李红飞 He Haiyan;Deng Guohao;Hu Pengfei;Li Hongfei(Shanxi Agricultural University,College of Information Science and Engineering,Shanxi Jinzhong 030801)

机构地区：[1]山西农业大学信息科学与工程学院,山西晋中030801

出　　处：《南方农机》2022年第10期79-82,共4页

基　　金：山西农业大学2021年大学生创新创业训练项目(S202110113162)。

摘　　要：随着我国经济发展水平的提高,禽蛋业迅猛发展,但早期种蛋的孵化率不高,孵化率成了影响家禽饲养行业经济效益的一个重要因素,因此对受精蛋的挑选至关重要。目前工厂仍然采用人工的方法进行挑选,但成本高、效率低。本研究利用机器视觉,通过图像采集装置获取鸡蛋孵化第3天的图像,选择深度学习的YOLO V5算法自动获取图像特征并进行分析训练,可完成对受精蛋的自动挑选和定位,避免了人工挑选的缺点,适合应用在工厂自动化生产过程中。With the improvement of Chinaʼs economic development level,the poultry egg industry develops rapidly,but the hatching rate of early breeding eggs is not high,the hatching rate has become an important factor affecting the economic benefits of poultry raising industry,so it is very important to select fertilized eggs.Therefore,the selection of fertilized eggs is very important.At present,the factory still adopts the manual method for selection,but the cost is high and the efficiency is low.This study uses machine vision to obtain the image of the third day of egg incubation through the image acquisition device,selects the deep learning YOLO V5 algorithm to automatically obtain the image features and carry out analysis and training,which can complete the automatic selection and positioning of fertilized eggs,avoid the shortcomings of manual selection,and is suitable for application in the factory automation process.

关 键 词：深度学习 受精蛋 图像识别 YOLO 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]