Cell Nerve Network (CNN) has been used to process the cloudy radar image. And then we use mathematic to diagnose the cloudy is hail or not. Veins is useful in diagnose the hail cloudy in weather forecast. The veins of the radar image have been picking up according the CNN. And then find the regular polynomial to processed radar image .Then analysis the image with polynomial fitting. We enlarge eight times of the key part of the radar image, and then detect the edge of the image. Dug some data as information .At last we find some regular to distinguish the cloud with hail or not. We find those are useful way for forecasting of hail.

In this paper, a kind of transform has been given for an image that switches to a matrix. A Radar image is an RGB image. An RGB image is usually represented to three layers: Red, Green, and Blue. Each layer can be changed to a matrix. We give a way to transform the three layers to one matrix. All image information will be remained. According to this way, the hail cloud radar image had been transformed to a matrix. The information of hail will been remain in this matrix.