一种基于图像纹理分析的冷镜式露点仪研制

A research of the chilled-mirror dew-point hygrometer based on the technology of the image texture analysis

针对目前冷镜式露点仪不能实现对镜面露和霜的检测和识别的问题,研制基于图像纹理分析技术的冷镜式露点仪。首先,利用高倍CCD快速采集镜面露霜图像,然后将图像进行灰度压缩和正规化的预处理,通过灰度差分比对和纹理特征提取分析等技术,训练构建识别函数和阈值,实现对镜面生成露霜的检测,以及对露点和霜点的自动判别,最后结合双压法标准湿度发生器和人工观测进行试验验证。结果表明,该冷镜式露点仪能够快速准确实现对露霜的检测和识别,可作为实验室湿度标准器使用。

According to the problem of the Chilled-mirror dew-point hygrometer, which can not realize the detection and identification to the dew and the frost on the mirror, a new Chilled-mirror dew-point hygrometer based on the technology of the image texture analysis and the pattern recognition was designed. First,used the high power CCD to get the dew and frost image, and pretreated the image by the technology of the gray compression and the regularization. The gray differential technology and the texture analysis technology were used to train and build the recognition function and threshold value of the system, and to realize the detection and identification to the dew and the frost on the mirror. Finally,the test verification has been done by the double standard humidity pressure generator and the manual observation. The results showed that the chilled-mirror dew-point hygrometer can detect and identify the dew and the frost rapidly and accurately,and it can apply in the laboratory as the humidity standard instrument.