Automatic classification method and device for precipitation cloud type based on diversified 3D radar echo characteristics

Abstract

The invention discloses an automatic classification method and device for the precipitation cloud type based on diversified 3D radar echo characteristics. The method comprises the following steps: analyzing networking radar data to obtain a plurality of radar echo characteristics, randomly selecting two echo characteristics from the plurality of echo characteristics, obtaining an optimal solutionof a cost function by using a neural network model, and drawing a learning curve graph; calculating a high variance and a high deviation based on the drawn learning curve graph, increasing the numberof training samples of the neural network when the high variance condition is met, inputting the increased training samples of the neural network into the neural network model, and drawing the learning curve graph again; when the high deviation condition is met, increasing the number of echo characteristics; inputting the added echo characteristics into the neural network model, and drawing a learning curve graph again; when the high variance condition is not met and the high deviation condition is not met, an optimization model is obtained; obtaining the optimized characteristics and the optimized characteristic quantity for training; and inputting data to be classified into the trained model, and carrying out classification to obtain a classification result. Thus, automatic cloud body classification of high recognition accuracy and speed is realized.

Description

technical field [0001] The invention belongs to radar echo processing technology, in particular to automatic classification of precipitation cloud types based on diversified 3D radar echo features. Background technique [0002] In recent years, the use of short-wavelength radars to form radar networks has become the mainstream way to overcome the inherent shortcomings of radars. For example, problems such as beam occlusion, low low-altitude coverage, and energy attenuation in the radar detection process can all be connected to ground radar networks. In the radar network scanning process, different scanning strategies will be adopted for convective clouds and stratiform clouds, so improving the accuracy and speed of cloud classification is a top priority. [0003] Existing algorithms can be roughly divided into two categories, one is based on the feature threshold algorithm, the algorithm for the selection of the threshold is fixed, and this setting is not applicable to all w...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to aim at the disadvantages of slow recognition speed and low accuracy in the prior art, and use the diversified radar echo products that can reflect the cloud structure as the input of the neural network for training, and finally obtain a high-precision, high-speed recognition Automatic methods for cloud types

Images