基于不变矩理论的雷达回波图中超级单体识别

Identification of super cells in radar echo based on moment invariants theory

根据超级单体在反射率因子图中的强度信息、水平尺度、典型的钩状回波等特征,利用图像处理技术对多普勒天气雷达回波图中超级单体进行识别,从而更好地为天气预报服务。首先通过图像分割获取灰度图中超级单体的结构,然后结合轮廓跟踪法和贴标签的方法对目标进行区域分割,获取目标存在区域,最后基于不变矩具有对位移、尺度和旋转变换的不变性对目标进行识别。仿真结果表明该方案可行且具有良好的识别性能。

A method to identify the super cells in the Doppler weather radar echo is proposed using the image process- ing technology primarily based on the intensity information of the radar reflectivity factor, the scale level and the typ- ical characteristics of the clasp echo of the super cells so as to serve the weather forecast better. The super cell struc- ture in the gray-scale image is obtained through the image segmentation. The targets are segmented and the existed goal regions are obtained by the contour tracking and labeling methods. The targets are recognized based on the prop- erty of the moment invariants which is of the .same displacement, scale and rotational transformation. The result of the simulation shows that this scheme is feasible and has a perfect recognition performance.