二维傅里叶变换在露点仪镜面图像偏移修正中的应用  

Application of 2D Fourier transform in mirror image displacement of dew point hygrometer

作　　者：代中华[1] 高太长[1] 赵世军[1] 马宁[1] 苏伟 

机构地区：[1]解放军理工大学气象海洋学院,南京211101 [2]南京英恩特环境技术有限公司,南京211153

出　　处：《气象科学》2016年第4期531-536,共6页Journal of the Meteorological Sciences

基　　金：国家自然科学基金资助项目(41475022);公益性行业(气象)科研专项(GYHY201206035)

摘　　要：图像式露点仪在采集镜面图像时由于温度波动和振动冲击会使采集图像发生偏移,本文针对偏移造成露点图像识别率低的问题,基于二维傅里叶变换的空间位移性,建立相对关联函数(RCF)对偏移量进行计算和修正,通过寻找其能量分布在低维空间的峰值响应,从理论上推导了偏移量的计算方法,并结合实际设计了算法。将算法应用到露点图像识别中,通过实验测试和对比分析表明该图像偏移检测方法精度较高,可行性好,对熵值法的识别准确率提高有明显的作用。Because of the temperature fluctuation and vibration shocking, the dew point hygrometer would make the detector surface image of metal mirror move and would lower dew recognition efficiency. To solve this problem, a new image displacement measuring method was presented, and based on spatial displacement of two-dimensional Fourier transform, a Relative Correlation Function （RCF） is set up and the image displacement is measured by searching the peak output of RCF＇ s energy distribution in lower- dimension space. An algorithm is proposed according to actual situation and applied it to dew image rec- ognition. The experiment and contrast analysis show that the method has high precision, good feasibility and increasing image entropy recognition efficiency obviously.

关 键 词：露点仪 镜面图像 偏移修正 二维傅里叶变换 凝结状态识别 

分 类 号：P412.13[天文地球—大气科学及气象学]