Monitoring of weather radar digital products based on pattern recognition method

Abstract

After the construction and development of weather radar detection network in recent decades, hundreds of new generation weather radar detection network have been applied in practice. From the perspective of R&D and application of weather radar digital products, under the influence of radar hardware quality, electromagnetic interference and other factors, there will be many abnormal images in weather radar image products, which will directly follow the weather forecast results, so meteorological business must combine intelligent algorithms to solve these problems. In this paper, an application algorithm for automatic detection of weather radar abnormal images is proposed, which includes four parts: image preprocessing, edge detection, feature extraction and classifier based on artificial neural network. The final experimental results show that this method can effectively solve the problem of abnormal image observation and inspection, and can ensure the efficiency and quality of practical application.