Detection And Identification Of Gust Front

Gust front is a strong convective weather, with characteristics of sudden strong destructive, irresistible and so on, and the whole process of its formation and development is very rapid. In its sphere of influence, there will be thunderstorms, heavy rain, strong winds, hail, and other disasters, so people's livelihood and agricultural production are impacted greatly. Therefore the early detection and identification of gust front is essential.Based on the echo image constructed from the Doppler weather radar data, this paper analysis the geometry characteristics of the bow echo triggered by the gust front and find that the shapes of non-bow echoes are similar to the round, while the shape of the bow-shaped echo is similar to the shape of the crescent moon. According to this characteristic, we propose to find out the unique features of bow echo region in frequency domain. Firstly, we convert RGB image into gray image, then detect the edge of echo region using sobel edge detection operator. In order to get consistency edge, the closure computing operation is applied on the image. Secondly, we use the scan and label chain code tracking algorithm to extract the edge of the closed areas of echo image, and transform the 2-D edge data into 1-D wave data. Lastly, we convert the 1-D expression of edge to frequency domain using Fourier transform, calculate their power density spectrum and made it normalization. We find that the bow echo the region's power density focus more on the density near the peak, so the peak of the spectrum can distinguish whether the region is a gust front. Apply to the actual radar images, the result shows that the algorithm can effectively detect the gust front.