Dual-polarization meteorological radar precipitation particle classification method based on discrete attribute BNT

Abstract

The invention discloses a dual-polarization meteorological radar precipitation particle classification method based on discrete attribute BNT. The method comprises the following steps: firstly, discretizing an input polarization parameter, then constructing a Bayesian network with discrete attributes by using discretized data, determining a class prior probability by fully utilizing prior information, and finally classifying precipitation particles according to a Bayesian principle. Compared with a traditional FLA classification algorithm, the recognition rate of precipitation particles is increased, errors caused by selection of a probability model or a membership function in a traditional method are effectively avoided, the recognition capacity of precipitation particles in non-unimodaldistribution such as ice crystals is obviously improved, and the algorithm operability and generalization are higher. Simulation experiments verify the effectiveness of the method.

Description

technical field [0001] The invention belongs to the technical field of weather radar signal processing, in particular to a method for classifying precipitation particles of dual-polarization weather radar based on a discrete attribute Bayesian network (BNT). Background technique [0002] The reasonable identification of the phase state of precipitation particles in clouds has very important scientific significance in the fields of cloud precipitation physics and artificial weather modification. It is not only important for understanding the generation and transformation of hydrometeors in clouds, but also for improving the measurement accuracy of precipitation value, and can provide an important reference for the decision-making and evaluation of artificial weather modification. [0003] Since traditional single-polarization weather radars only transmit and receive power information in a single direction, the information obtained is limited, which limits the accuracy of its ...

AI Technical Summary

Problems solved by technology

However, there are always problems: the selection of the polarization parameter membership function of different precipitation particles and the determination of the weight of different precipitation particles need to rely on expert experience, and this problem has not been effectively solved.

However, in recent years, with the in-depth study of the distribution characteristics of precipitation particles, when there are many classification categories, the probability distribution of radar echo data of some precipitation particles (such as ice crystals) is bimodal, and continuing to use the Gaussian model will be affected by the selection of the model. bring large error

And Marzano only uses temperature information as additional information to determine the class prior probability information, without more comprehensive consideration

Images