MAP-MRF Estimation Based Weather Radar Visualization

Suk-Hwan LEE  

IEICE TRANSACTIONS on Information and Systems   Vol.E101-D   No.7   pp.1924-1932
Publication Date: 2018/07/01
Online ISSN: 1745-1361
DOI: 10.1587/transinf.2017EDP7212
Type of Manuscript: PAPER
Category: Image Processing and Video Processing
weather radar image,  reflectivity image,  contour line retrieval,  MRF-MAP estimation,  PPI image,  

Full Text: PDF(2.1MB)
>>Buy this Article

Real-time weather radar imaging technology is required for generating short-time weather forecasts. Moreover, such technology plays an important role in critical-weather warning systems that are based on vast Doppler weather radar data. In this study, we propose a weather radar imaging method that uses multi-layer contour detection and segmentation based on MAP-MRF estimation. The proposed method consists of three major steps. The first step involves generating reflectivity and velocity data using the Doppler radar in the form of raw data images of sweep unit in the polar coordinate system. Then, contour lines are detected on multi-layers using the adaptive median filter and modified Canny's detector based on curvature consistency. The second step interpolates contours on the Cartesian coordinate system using 3D scattered data interpolation and then segments the contours based on MAP-MRF prediction and the metropolis algorithm for each layer. The final step involves integrating the segmented contour layers and generating PPI images in sweep units. Experimental results show that the proposed method produces a visually improved PPI image in 45% of the time as compared to that for conventional methods.