多普勒雷达图像逆风区的自动监测识别

在多普勒雷达降水回波径向速度场中及时准确地发现逆风区,对灾害天气预报预警具有重要意义.根据逆风区在雷达径向速度图中的物理图像特征,采用数字图像处理和分析方法实现了逆风区自动监测识别.首先,以雷达图像色标为依据,采用阈值法分别获取正、负速度区域二值图像,再对2幅图像分别进行形态学运算,然后将上述4幅图像做交叉逻辑运算,得到逆风区监测识别结果和相关参数.通过在2005—2011年长沙雷达站47幅根据实况进行人工标注后的多普勒雷达径向速度图像上进行实验,表明该方法对逆风区可以进行快速准确识别,与人工标注结果比较准确率可达89%,满足实际应用需要.

Automatic monitoring and recognition of adverse wind area in Doppler radar velocity images

Accurate monitoring on adverse wind areafrom Doppler radar images is of great significance in the fast and automatic warning system of severe weather.In this paper,an in-depth study on the structural features of adverse wind area of Doppler radar radial velocity image iscarried out,then an accurate and automatic monitoring algorithm is designed and implemented which runs as follows.Firstly,the Doppler radar radial velocity color image is processed into two binary images,namely the positive speed binary image and the negative speed binary image,based on the color bar.Secondly,the two binary images are operated by morphology for de-noising.After that,the above four images are operated together by cross logic and the monitoring result with some related parameters are obtained accurately.In order to show the effectiveness of this algorithm,forty-seven Doppler radar radial velocity images include one or more adverse wind areas,from Changsha Radar Station of Hunan province,are tested.All experiments present an exciting result that the proposed algorithm can be 89% in accuracy,which is extremely effective for practical application.