Floating car track point space relationship and distribution-based road network extraction method

Abstract

The invention relates to a floating car track point space relationship and distribution-based road network extraction method. The method comprises the following steps of: S1, carrying out data conversion and cleaning; S2, establishing a two-layer space grid index G; S3, calculating a track point space distribution F; S4, calculating a core point candidate set H on the basis of the track point space distribution; S5, calculating core points P: calculating a geometric centroid of the core point candidate set, and marking the track points, closest to the centroid, in the core point candidate H asthe core points; S6, carrying out core point connection: obtaining traffic relevancy functions C (A, B) by considering factors of space distances on the basis of a track point space clustering method, and connecting the core points, the traffic relevancy functions C (A, B) of which in same directions; and S7, extracting a road network after all the core points are connected. The method has the advantages of being stable and reliable in operation, high in road network extraction efficiency, correct in precision and suitable for floating car data under multiple conditions.

Description

technical field [0001] The invention relates to the field of application technology based on big data of floating car trajectory, in particular to a road network extraction method based on the spatial relationship and distribution of floating car trajectory points. Background technique [0002] Floating vehicle technology is one of the dynamic traffic acquisition technologies often used in intelligent transportation systems in recent years. Its outstanding advantage is that it can obtain accurate and real-time dynamic traffic information through a small number of floating vehicles equipped with satellite positioning-based vehicle equipment, which is low in cost and efficient. High, with the characteristics of strong real-time performance and large coverage. [0003] Traditional road network extraction methods are mainly divided into on-the-spot surveying and mapping methods and generation methods based on remote sensing images. The former has a higher accuracy rate, but it ...

AI Technical Summary

Problems solved by technology

Some of these algorithms have high requirements on the data collection frequency of floating vehicles (the time interval between adjacent trajectory points is less than 10 seconds), but they are suitable for urban road networks, and the application space is relatively narrow; That is to say, the platform requirements of the research are relatively high; the road network constructed by some algorithms is more accurate, but the coverage is not good, and the time efficiency is very low, which is not conducive to popularization and application

However, for low-frequency sampling (about half a minute between adjacent track points) floating car data, the existing road network extraction algorithms based on floating car trajectories cannot be well applied to the extraction of ordinary roads.

Images