A Geospatial Pattern Recognition Method Based on Accessibility Space

Abstract

The invention discloses a geographic space pattern recognition method based on a reachable space. The geographic space pattern recognition method comprises the following steps: S1, acquiring space information RA and road traffic information TA of a target area; S2, dividing an area where the spatial information RA is located into N sub-areas Pi; S3, performing spatial reachability calculation according to the road traffic information TA, and obtaining a reachable space range ASi of each sub-region Pi under a specific reachability threshold; S4, extracting local information ARi in the corresponding RA according to the spatial region range covered by the ASI; S5, according to the ARi corresponding information, identifying to obtain a mode Ti of a space corresponding to the position Pi; and S6, summarizing the spatial modes Ti of all the sub-regions to obtain the spatial mode ST of the target region. According to the method and the device, the traffic information is utilized, spatial pattern recognition is associated with the activity range of people, and meanwhile, the spatial pattern recognition precision and accuracy are improved.

Description

Technical field [0001] The present invention relates to the technical field of spatial recognition, and in particular to a method of geographic spatial pattern recognition based on reachable space. Background technique [0002] With the continuous development of machine learning technologies such as neural networks and deep learning, the application of similar image recognition technology to remote sensing image recognition and other geographic spatial pattern recognition research and technical applications continue to emerge, which significantly improves the feasibility of automatically identifying spatial patterns through image information And reliability. At present, most methods use image pixel information within a certain range of RA, or image pattern PA within a regular range nearby, and use the uniform grid of traditional CNN convolutional neural network to divide the space range to identify the space of RA Mode T, some new methods introduce some auxiliary information to ...

AI Technical Summary

Problems solved by technology

However, in the actual geographical environment, there is an obvious deviation between the assumption of a uniform and isotropic consideration of the relationship between a certain geographical space and the surrounding space and the actual situation. The strength of the spatial relationship has a natural correlation with whether people can easily reach it. In other words, there are significant differences in the correlation between the adjacent space and the target space pattern at different accessibility levels.

Images