基于距离判别法的雹云识别模型

【目的】提高雹云识别准确率,降低因冰雹造成的经济损失。【方法】依据气象雷达反射率图像,利用Kmeans聚类提取云层的内外轮廓,计算其距离方差;利用雷达软件提取云高数据并计算云高的一阶统计测度;将云层内外层轮廓的距离方差与云高一阶统计测度相结合,构造雹云判别模型。【结果】利用此模型对已有样本检测,可知该模型识别率判别准确率为88.75%,准确率较高。【结论】通过内外轮廓方差与云高一阶统计测度构造的距离判别模型有较好的判别效果。

Hail Cloud Recognition Model based on Distance Discriminant Analysis

Objective] Improving the accuracy of hail cloud identification can reduce the economic losses caused by hail. Methods] According to the meteorological radar reflectivity image, the inner and outer contours of the clouds were extracted using K-means clustering, and the distance variance was calculated. The radar software was used to extract the cloud height data and calculate the first-order statistical measure of the cloud height. The distance variance of the inner and outer layers of the cloud was combined with the first-order statistical measure of the cloud height to construct a hail cloud recognition model. Results] Using this model to detect existing samples, we could see that the recognition accuracy of the model was 88.75%, and the accuracy was higher. Conclusion] The results showed that the distance recognition model constructed by the internal and external contour variance and the first-order statistical measure of cloud height had a better recognition effect.