基于数学形态学的阵风锋识别算法

Automatic Recognition of Gust Fronts Based on Mathematical Morphology

提出基于数字图像处理技术中的数学形态学的阵风锋自动识别算法。该算法首先基于阵风锋的反射率因子窄带形态、速度辐合和快速移动特征,计算相邻时次体扫反射率因子差、速度切变和速度梯度这3个参量场,综合这3个参量场,得到一个二维特征量;然后采用数学形态学算法对这个二维特征量进行二值化、膨胀、腐蚀、细化、剪枝及线段联通等处理,得到一条反映阵风锋骨架特征的曲线。利用武汉雷达观测的2个阵风锋过程的共53个体扫数据进行阵风锋识别算法检验,结果表明:该算法正确识别阵风锋39个,错误识别3个,漏识别11个;识别曲线与阵风锋位置和走向较为一致。

Based on mathematical morphology of digital image processing technology,an automatic recognition algorithm of gust fronts is devised.Based on the narrow-band shape,velocity convergence,and moving rapidly of gust front,three parameters（difference between reflectivity factors of two adjacent times,velocity shear,and velocity gradient）can be gained.Taking into account of the three parameters,a2 Dcharacteristic quantity field can be obtained.Applying binaryzation,dilation,corrosion,thinning,cutting,and conjunction,a curve corresponds to the skeleton of a gust front is shown as results.To evaluate the algorithm,the dataset of gust fronts（53volumes）are applied.The results show that the algorithm recognized 39 cases correctly,and meanwhile,3cases are recognized falsely,and 11 are missing recognition.