基于可见光——红外图像信息融合的云状识别方法

A cloud classification method based on information fusion of visible and infrared images

云状的正确观测对降水测报具有指示性意义,云状自动识别技术是气象要素自动化观测领域的难题之一。本文基于全天空可见光成像仪采集的云图与中红外热像仪获取的云图结合,对天空云状进行分类和测量。结果表明:通过在北京、杭州和丽江气象台站采集的大量云图,从云图特征和降水指示性方面将云状划分为Clear、C_H、C_L、C_B及C_M共5类。选取14个色彩和纹理特征值作为云状计算参数,采用552张云图作为训练样本,信息分类利用特征值加权最小距离算法,对于5类500个被测样本进行云状的判别。对应自拟的标准云状分类,平均准确率为82%。基于可见光—红外图像信息融合的云状识别方法结合了可见光图像色彩信息丰富的特点及红外图像可以降低雾霾干扰的优势,对比单一可见光传感器云测量,准确性有所提高。本文在可见光与红外图像传感器等多种云观测设备的信息融合方面进行了有益的尝试。

Correct observation for cloud has an indicative significance for rainfall forecasting and cloud classification technology is one of the problems in the field of automatic meteorological observation. In this paper, the cloud images collected with the whole sky imager instrument and the infrared imaging system were integrated for classification and measurement of cloud. Results show that on the basis of a great deal of cloud images collected from Beijing,Hangzhou and Lijiang weather stations,considering the features of cloud images and their indications for precipitation, the cloud shapes are divided into five categories including Clear,C_H,C_L,C_B and C_M. In addition,based on the 14 eigenvalues on colors and textures selected as the parameters for calculating cloud form and 552 cloud images used as the training samples, integrating a minimum distance algorithm with eigenvalue weighted method, the observed 500 samples are distinguished and divided into the above-mentioned 5 categories according to cloud form. Taking the standard cloud form classification as reference, the average accurate rate reaches 82%. This method combining the characteristics of visible light image with rich color and the advantage of infrared image in reducing the fog and haze interference shows a higher accuracy relative to the cloud measurement method with the single visible light sensor. This research makes a meaningful attempt in the field of information fusion of visible and infrared cloud image sensors.