基于鸡冠图像特征的蛋鸡产蛋早期产蛋量预测研究  

Study on Egg Production Prediction of Laying Hens Based on Comb Image Features

作　　者：郭蓓佳 张晓雨 籍颖[1,2] 锡建中 周荣艳 陈辉[4,5] 王德贺 GUO Beijia;ZHANG Xiaoyu;JI Ying;XI Jianzhong;ZHOU Rongyan;CHEN Hui;WANG Dehe(College of Information Science and Technology,Hebei Agricultural University,Baoding,Hebei 071001;Hebei Key Laboratory of Agricultural Big Data,Baoding,Hebei 071001;Graduate School,Hebei Agricultural University,Baoding,Hebei 071001;College of Animal Science and Technology,Hebei Agricultural University,Baoding,Hebei 071001;Key Laboratory of Broiler and Layer Breeding Facilities Engineering,Ministry of Agriculture and Rural Affairs,Baoding,Hebei 071001)

机构地区：[1]河北农业大学信息科学与技术学院,河北保定071001 [2]河北省农业大数据重点实验室,河北保定071001 [3]河北农业大学研究生学院,河北保定071001 [4]河北农业大学动物科技学院,河北保定071001 [5]农业农村部肉蛋鸡养殖设施工程重点实验室,河北保定071001

出　　处：《中国家禽》2023年第5期31-38,共8页China Poultry

基　　金：河北省鸡现代种业科技创新团队(21326303D);财政部和农业农村部:国家现代农业产业技术体系(CARS-40)。

摘　　要：为探究产蛋前期的蛋鸡鸡冠形态与产蛋量的关系,试验采集20周龄蛋鸡鸡冠侧视图像,利用最大类间方差法对L^(*)、a^(*)、b^(*)颜色空间中的a^(*)分量进行阈值分割,基于鸡冠形态的全局特性,采用凸包分析方法定位鸡冠的最高冠齿,提取鸡冠形态的主要特征参数并与产蛋性状进行相关性分析,构建灰狼算法优化支持向量机(GWO-SVM)的产蛋量分级模型,进而将差分进化算法与灰狼算法的搜索机制相融合建立产蛋量预测模型(DEGWO-SVR)。结果显示:开产日龄与冠高、冠长呈极显著负相关(P<0.01),产蛋量与冠长、冠高呈极显著正相关(P<0.01);GWO-SVM模型测试集分类准确率可达94.28%,较传统的SVM模型准确率提高了5.71%;DEGWO-SVR模型测试集的决定系数(R^(2))为0.861。表明DEGWO-SVR模型预测蛋鸡产蛋早期产蛋量精度较高,可以为产蛋早期对鸡冠形态的选育和产蛋量的预测提供理论依据。In order to explore the relationship between comb morphology and egg production in pre-laying period of laying hens,the side view images of laying hens aged 20 weeks were collected,and the a*component in the L^(*),a^(*),b^(*)color space was threshold segmented by using the maximum interclass variance method.Based on the global characteristics of comb′s morpholo⁃gy,the convex hull analysis method was used to locate the highest comb′s teeth.Then extracted the main characteristic parame⁃ters of comb′s morphology and analyzed the correlation between comb shape and egg laying traits.An egg production classifica⁃tion model of support vector machine optimized by gray wolf algorithm(GWO-SVM)was constructed.After that,an egg produc⁃tion prediction model(DEGWO-SVR)of SVM optimized by improved gray wolf algorithm was proposed.The results showed that the age at first egg significantly correlated with the crown height and crown length(P<0.01),and the egg production significantly correlated with the crown length and crown height(P<0.01);The classification accuracy of GWO-SVM model test set could reach 94.28%,which was 5.71%higher than that of the traditional SVM model;the coefficient of determination(R^(2))of the DEGWO-SVR model test set was 0.861.It showed that the prediction accuracy of DEGWO-SVR model was high for egg production of laying hens,which could provide theoretical basis for the selection of chicken comb morphology and the prediction of egg production in pre-laying period of laying hens.

关 键 词：图像处理 鸡冠 SVM 灰狼算法 差分进化算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] S831[自动化与计算机技术—计算机科学与技术]