当前位置>主页 > 期刊在线 > 计算机技术 >

计算机技术2020年23期

基于LeNet 深度学习模型的雏鸡性别智能识别
杨陆野
(上海大学,上海 200444)

摘  要:不同性别雏鸡在生产生活中的价值及培育方向均有较大差别,尽早无损伤高效地分辨出雏鸡性别从而进行定向培育,实现产业利益最大化。传统的肛门鉴别法、伴性性状鉴别法、蛋内尿囊液检测法和核磁共振测定法均有其不足之处。受肛门鉴别法中对肛部图片两大特性传统人眼检测方案的启发,在对原始图像数据进行关键区域特征增强和图像增广后,依托于LeNet深度学习模型的强大图像处理能力实现极高效率的雏鸡性别识别。


关键词:雏鸡性别鉴定;深度学习;图像识别;图像增广;LeNet



中图分类号:TP391.41         文献标识码:A         文章编号:2096-4706(2020)23-0064-03


The Gender Intelligence Recognition of Chicks Based on the LeNet Deep Learning Model

YANG Luye

(Shanghai University,Shanghai 200444,China)

Abstract:Different gender chicks have a large difference in the value and cultivation direction in production and life,and distinguish the gender of chicks as early as possible and efficiently without damage,so as to carry out directional cultivation and realize the maximization of industrial benefits. The traditional methods of anal identification,sex-linked charateristics identification, egg allantoic fluid detection and nuclear magnetic resonance detection have their shortcomings. Inspired by the traditional human eye detection scheme of anal image two characteristics in anal identification method,after the key area feature enhancement and image expansion of the original image data,relying on the powerful image processing ability of LeNet deep learning model to realize highly efficient chicks gender recognition.

Keywords:gender identification of chicks;deep learning;image recognition;image augmentation;LeNet


参考文献:

[1] 王健,漠纵泉,肖小珺,等. 鸡胚性别鉴定——家禽科学研究的新视点 [J]. 中国家禽,2003(14):26-36.

[2] 姜桂苗,何剑斌,马振乾,等. 家禽性别控制的研究进展 [J],黑龙江动物繁殖,2010,18(5):14-16.

[3] 燕海峰,肖兵南,TREFIL P,等. 家禽的性别鉴定方法 [J]. 动物学杂志,2001(6):58-61.

[4] ZHANG A,LI M,Lipton Z C,等. 动手学深度学习 [M].人民邮电出版社,2019:15-188.

[5] 郑睿,余童,程龙阅. 基于改进LeNet-5 的形状类似物体识别方法 [J]. 信息技术与网络安全,2020,39(6):31-37+43.


作者简介:杨陆野(1993—),男,汉族,上海人,初级工程师,硕士,研究方向:计算机视觉。