Vehicle model identification method based on high-resolution remote sensing data

Abstract

A vehicle model identification method based on high-resolution remote sensing data includes acquiring raster remote sensing data including a vehicle region of interest and map vector data including the region of interest, performing rough alignment on the raster remote sensing data and the map vector data, determining a to-be-analyzed region, performing automatic extraction of center separation lines, lane or carport vacant region recognition, vehicle region detection and vehicle model identification, and outputting. On the basis of different individual analyses of the high-resolution remote sensing data, multi-source data fusion and cross validation can be performed prior to outputting. Vehicle models are automatically identified by means of priori utilizing knowledge provided by short separation lines of road centers or parking lots and integrating relevant vehicular exclusive information provided by various high-resolution remote sensing data in the context of roads or packing lots where the vehicles are located, and accordingly basic data are provided for traffic volume survey, traffic planning and the like.

Description

technical field [0001] The method belongs to the technical field of image processing and recognition, and in particular relates to a new vehicle model recognition method based on high-resolution remote sensing data. Background technique [0002] With the rapid development of space technology in various countries around the world, there are more and more high-spatial resolution remote sensing data, including QuickBird and WorldView-2 of foreign Digital Globe companies, as well as my country's resource sequence, high-scoring sequence, and environmental sequence satellites data etc. High-resolution remote sensing data provides us with high-precision spatial geometric information, rich texture information, and multispectral information of ground objects, laying the foundation for fine recognition of ground objects. The use of remote sensing high-resolution data for fine recognition of ground objects (such as vehicles) has the characteristics of convenience and speed, wide covera...

AI Technical Summary

Problems solved by technology

[0005] 1) Most of the algorithms are focused on vehicle detection, and there are very few literatures on fine vehicle model recognition, resulting in limited guiding significance of the algorithm results for traffic planning, and it is impossible to give more targeted planning and planning for different vehicle occupancy conditions. policy guidance;

[0006] 2) The existing literature algorithms do not fully utilize the vehicle context information

Images