Research On Freight OD Information Extraction Method Based On Big Data Of Freight Truck Trajectory

With the progressing of urban agglomeration construction,the phenomena of continuous expansion of urban land,low efficient use of land,and rapid resource consumption have become increasingly prominent.Effectively assessing the development level of urban agglomerations is of great significance to promote the sustainable and healthy development of urban agglomerations.Freight transportation can be classified as a kind of space industry in essence.The amount of freight demand between cities can fully show the economic correlation between different areas,which is an important factor in evaluating the development level of urban agglomerations.Therefore,this paper,based on the 156GB trajectory data collected by the 2 million truck-mounted GPS devices in the country as the research object,aims at providing an effective method of extracting freight information by analyzing and researching the freight data of the whole country,preliminarily judging the development level of urban agglomerations.The extraction of freight OD from it is an important basis to judge the freight relationship between cities,and also the basic premise to further develop the evaluation index of urban agglomeration development based on freight data.However,due to the huge amount of GPS data of the freight trucks,there is still no effective method for extracting OD information.So this paper carries out in-depth research on how to accurately and efficiently extract OD information from freight data.(1)Construction of big data processing platform Hadoop and data preprocessing.On the basis of the comparison of data processing and calculation methods at home and abroad,the software Hadoop is picked out first and the big data processing platform Hadoop is built,which provides support for the subsequent storage,calculation and management of relevant data.Then,this paper does data preprocessing work such as removing repetition&missing and abnormal point data for original trajectory data,filtering out effective data and making distributed storage in the data processing platform.(2)Construction of the identification method of determining the parking points based on the freight truck parking speed threshold.Based on processing and analyzing the freight truck GPS data,this paper innovatively proposes a statistical method of cumulative driving distance of freight trucks based on time division and determines the cumulative driving distance matched by the the parking behavior of the trucks.By analyzing the average speed and the cumulative driving distance matched by the the parking behavior of the trucks,the speed threshold of freight truck parking is determined and finally the parking points in the truck trajectories are identified.(3)The judgement of freight information based on map matching and extraction of OD information.After studying the method of collecting road network and POI information,the paper introduces a grid-based map matching algorithm based on grid index and solves the problem that the data is so big that it is inefficient to do it by determining the value of the grid edge length through analyzing numerical calculation.In the end the paper efficiently identifies non-OD parking points caused by congestion,rest,charges,etc.by judging obtained freight information.(4)The classification of OD point the based on DBSCAN clustering method and the analysis of the current urban agglomeration development.On the basis of classifying OD by means of the DBSCAN clustering method,the classified OD points are matched in the cities and counties throughout the country,so that the distribution map of the freight trucks within them is obtained.After that,the agglomeration development level of the whole country is evaluated based on freight information,and the urban agglomerations,the Pearl River Delta and Chengyu,are specifically analyzed.