摘要
目的构建鼠情智能远程监测系统,实现长期、高效、准确地监测区域内鼠情及鼠密度分布。方法基于人工智能算法、大数据、物联网等技术,通过终端识别拍摄,生物智能识别技术,识别分析与记录,数据远程传输,数据整理分析,研究鼠情智能远程自动监测技术。结果研制了鼠情智能远程监测系统,该系统通过3万幅图模拟运算,鼠识别准确率达到96.6%,同一只鼠分辨准确率在72.8%以上。在选定的5个区域内进行试运行,均监测到鼠类活动,并自动转换为图表和分析报表等形式,向用户展示鼠情及鼠密度分布等情况。结论本系统可长期持续识别、分析、记录监测目标区域鼠类活动情况及鼠密度情况,监测数据客观、准确度高,并可大量节省人力。系统对大数据进行自动存储、分析预警,为科学防控提供数字化的决策支持。
Objective To develop an IoT-based rodent tele-monitoring system, so as to realize long-term, efficient and accurate monitoring of rodent population in the region.Methods Based on the AI, big data and IoT technologies, a system that can automatically tele-monitor, recognize, analyze and record the rodent and the rodent density in the target areas was designed and applied. Results After being trained and tested, the model achieved 96.6% accuracy of rodent detection,and 72.8% accuracy of rodent recognition. This system was piloted in five areas,and it captured rodent activities in all these areas, also the data of distribution was automatically converted into charts and analytical reports for users. Conclusion This system detects the rodent with cutting-edge technologies, improves data accuracy and saves manual works. All the objectivity data are processed and stored with big data manners, and easily downloaded for detailed analytics in support of intelligent and informative decision-making,and to provide digital scientific prevention and control of the rodent through automatic data analysis.
作者
黄清臻
贾瑞忠
田志博
朱庆伟
梅燕莉
HUANG Qing-zhen;JIA Rui-zhong;TIAN Zhi-bo;ZHU Qing-wei;MEI Yan-li(Center for Disease Control and Prevention of PLA,Beijing 100071,China;Spider IoT Technology Co.,Ltd,Beijing 100025,China;Beijing Haidian District Environmental Protection Bureau,Beijing 100039,China)
出处
《中华卫生杀虫药械》
CAS
2019年第6期519-522,共4页
Chinese Journal of Hygienic Insecticides and Equipments
基金
军队疾控专项(编号:16SYF005802)