Analysis And Research Of General Aviation Accident Based On Flight Data Visualization

The safety of navigation is as great as the mountain.With the increasing popularity of navigation,more and more attention has been paid,and the role of general aviation in people's lives has become more and more important.Its safety has also been put on the topic of key research.When it comes to safety,the analysis of the causes of navigation accidents is crucial.Taking history as a guide,only knowing the real cause of the navigation accident,we can get a response to the problem and avoid the recurrence of the tragedy.This article is an exploration in the field of navigation accident analysis.This paper first reviews the development history and prospects of the under-air aviation,and analyzes the seriousness of aviation accidents and the urgency of the analysis of the causes of aviation accidents.Then,for the quantitative and qualitative research ideas,the advantages and disadvantages are summarized and compared.The qualitative method mainly analyzes Murphy's law,accident chain theory,SHELL model and REASON model,and analyzes their advantages and disadvantages.The quantitative method mainly analyzes the gray correlation analysis method and the neural network analysis method,and compares the advantages and disadvantages of these classical quantitative analysis methods.By comparing the advantages and disadvantages of these methods,we can find some blind spots in the current research and point out the necessity of visual analysis.Then,according to the characteristics of navigation aviation,its data collection,data collation and other strategies are given.12 categories of data are selected from a large number of data to visualize the flight characteristics of navigation aviation from various dimensions.The analysis strategy after the visualization is drawn is given.Finally,based on the fixed window filtering and particle swarm optimization algorithm,the data detection points are detected in the key detection data,and the detection results are fully visualized.Fixed window filtering can filter out noise and preserve the information without distortion.The particle swarm optimization algorithm can efficiently optimize the abnormal point detection model we have established and obtain the anomaly in the data information.In the paper,the abnormal point data is also fully visualized...