Calculation Method of Direction Adaptability of Geomagnetic Map Based on Convolutional Neural Network

Abstract

The invention discloses a geomagnetic chart direction adaptability computing method based on a convolutional neural network, and relates to the technical field of navigation methods. The method comprises the following steps that firstly, adaptability analysis charts of all candidate matching areas of a geomagnetic field in several representative directions are established from different angles; secondly, feature extracting and learning are automatically completed from different angles of a same magnetic field chart by utilizing the parallel convolutional neural network, and a more comprehensive direction adaptability feature description is obtained; finally, a geomagnetic direction adaptability analysis model is built through a BP neural network on the basis of the features obtained through the parallel convolutional neural network. According to the method, the problems that blindness is generated in artificial feature extracting and the extracted features cannot express the structural features of a deep layer can be effectively avoided, automation of geomagnetic direction adaptability analysis is achieved, and the accurate rate higher than that obtained through a traditional BP neural network and a one-way neural network is obtained.

Description

technical field [0001] The invention relates to the technical field of navigation methods, in particular to a method for calculating the direction adaptability of a geomagnetic map based on a convolutional neural network. Background technique [0002] Geomagnetic navigation is a navigation method based on geographic information. It has the characteristics of passive, no radiation, strong concealment, and error does not accumulate over time. It can conduct all-weather and all-region navigation. It has considerable advantages in the field of cross-sea guidance. A hot spot of research. At present, geomagnetic navigation is usually used as an auxiliary means to form a combined navigation system together with inertial navigation and satellite navigation to improve the reliability of the navigation system. The accuracy of geomagnetic navigation is not only related to the navigation algorithm, but also closely related to the ability of geomagnetic characteristics to characterize t...

AI Technical Summary

Problems solved by technology

[0006] (1) Regional adaptability It is difficult to evaluate the adaptability in a certain direction in the region, and the research on the adaptability of geomagnetic direction still needs to be further deepened;

[0007] (2) The traditional extraction of geomagnetic direction adaptability features is based on human experience, which is highly subjective and difficult to operate, and the adaptability evaluation has problems such as difficult extraction of evaluation criteria and difficulty in setting thresholds to avoid the influence of human factors;

[0008] (3) The traditional feature extraction method is difficult to extract the deep structural features of the input sample, and the feature description is not comprehensive, which affects the reliability of the direction suitability evaluation;

[0009] (4) The number of neurons and adjustable weights of the traditional neural network are limited, the learning ability is limited, and the efficiency of processing large amounts of data is not high

Images