深度学习在海关图像识别技术研究中的应用  

The Application of Deep Learning in Customs Image Recognition Technology

作　　者：吕玉峰 张晓龙 高文娜 杨宇 韩晶 韩立亮 乔彩霞 LYU Yu-Feng;ZHANG Xiao-Long;GAO Wen-Na;YANG Yu;HAN Jing;HAN Li-Liang;QIAO Cai-Xia(Science and Technology Research Center of China Customs,Beijing 100026;Qingdao Qingshu Technology Co.,LTD.,Qingdao 266114)

机构地区：[1]中国海关科学技术研究中心,北京100026 [2]青岛清数科技有限公司,青岛266114

出　　处：《中国口岸科学技术》2024年第S2期4-12,共9页China Port Science and Technology

基　　金：国家重点研发计划课题(2022YFC2302704)

摘　　要：在全球化背景下,各种潜在的疫病疫情、生物入侵以及野生动植物非法贸易等,对国家生态安全和经济发展构成了严重威胁,因此应用深度学习技术实现对目标物的快速准确识别成为维护国门生物安全的重要手段。本文以卫生检疫、动植物检疫和国门生物安全监测三个领域中深度学习图像识别技术的应用为例,根据项目研究数据显示,采用深度学习技术的图像识别算法模型在对各类有害生物的识别中具有较高的准确率和较强的鲁棒性,为提高海关工作效率和智能化水平、推进智慧海关建设提供了有力的技术支持。In the context of globalization,various potential epidemics,biological invasions and illegal trade activities of animals and plants pose serious threats to national ecological security and economic development.The application of deep learning technology to achieve rapid and accurate identification of target objects has become an important means of maintaining national biosafety.This article takes the application of deep learning image recognition technology in the fields of health quarantine,animal and plant quarantine,and national biosafety monitoring as examples.Based on project research data,image recognition algorithm models using deep learning technology have high accuracy and strong robustness in identifying various harmful organisms,providing strong technical support for improving customs work efficiency and intelligence level and promoting the construction of smart customs.

关 键 词：深度学习 图像识别 卫生检疫 动植物检疫 国门生物安全监测 智慧海关 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP18[自动化与计算机技术—计算机科学与技术]