摘要
传统红外测温技术因未考虑三维因素,导致测温准确度受限,难以适应当前生产生活的精确要求。因此,研究利用大尺度红外成像技术,结合改进的图像密集融合网络(Dense Fuse,DF)提出一种空间温度分布实时监测模型。结果显示,研究方法的信息熵、差异相关性总和、多尺度结构相似性度量以及相关系数指标值最优分别达到了7.9、1.52、0.99、078,均优于对比方法。系统集成测试结果显示在夏季温室纵剖面以及横剖面温度分布监测中,每日最大误差小于0.7℃,平均误差小于0.4℃,误差较小。结果表明方法能够实现对空间温度分布的监测,减小温度测量误差,提高空间温度分布监测的准确性和可靠性,为各行业提供有效的温度监测依据。
Traditional infrared temperature measurement technology is limited in accuracy due to the lack of consideration of three-dimensional factors,making it difficult to meet the precise requirements of current production and life.Therefore,a real-time monitoring method for spatial temperature distribution is proposed by using infrared imaging technology and an improved image dense fusion network(Dense Fuse,DF).The results showed that the information entropy,total correlation of differences,multi-scale structural similarity measurement,and correlation coefficient index values of the research method were optimal at 7.9,1.52,0.99,and 078,respectively,all of which were better than the comparative methods.The system integration test results show that in the monitoring of temperature distribution in the longitudinal and transverse profiles of the greenhouse during summer,the maximum daily error is less than 0.7℃,and the average error is less than 0.4℃,indicating a relatively small error.The results show that the method can achieve monitoring of spatial temperature distribution,reduce temperature measurement errors,improve the accuracy and reliability of spatial temperature distribution monitoring,and provide effective temperature monitoring basis for various industries.
作者
田茂城
叶露
田维文
陆秀波
王伟
TIAN Maocheng;YE Lu;TIAN Weiwen;LU Xiubo;WANG Wei(China Southern Power Grid Co.,Ltd.,Ultra High Voltage Transmission Company Guiyang Bureau,Guizhou Guiyang 550000,China)
出处
《自动化与仪器仪表》
2025年第8期109-112,117,共5页
Automation & Instrumentation
基金
中国南方电网有限责任公司超高压输电公司贵阳局科技项目资助(0102002024030301SJ00101)。
关键词
红外成像
空间温度
深度学习
检测
视场
infrared imaging
space temperature
deep learning
testing
field
