Classification method and system of road and its surrounding features

Abstract

The embodiment of the present invention provides a method and system for classifying roads and their surrounding features. The multi-layer perceptron is used to encode and decode the normalized point cloud data, so that local features can be determined more precisely. Information and global feature information, and the fusion of local feature information and global feature information through the connection layer helps to preserve multi-dimensional information and obtain correct target classification results.

Description

technical field [0001] The embodiments of the present invention relate to the field of surveying and mapping science and technology, and more specifically, to a method and system for classifying roads and surrounding features. Background technique [0002] At present, the correct classification of roads and their surrounding features plays an important role in the construction of unmanned driving and smart cities. In an unmanned driving environment, since the vehicle cannot directly know the road status and surrounding features, it needs to match the high-precision 3D map with the sensor information of the unmanned vehicle to guide the vehicle to drive. Therefore, it is necessary to correctly obtain the 3D position information and category information of the road and its surrounding features. [0003] To obtain the 3D location information and category information of the road and its surrounding features, the commonly used measurement means may include a total station, a GPS...

AI Technical Summary

Problems solved by technology

Among them, the total station usually collects data on the road in normal operation, which is likely to cause traffic jams and endanger the safety of surveyors

The GPS data obtained by the GPS data acquisition device can only provide limited trajectory data and cannot fully display the whole picture of the road

Although the street view data obtained by the street view data acquisition device can provide clear road images, it does not have direct three-dimensional information and cannot be directly applied to navigation and driverless work.

[0004] Obtaining the 3D position information of the road and its surrounding features can also be achieved through a vehicle-mounted laser scanning device. The vehicle-mounted laser scanning device integrates GPS / INS, laser scanners, wheel encoders and cameras, etc., and can quickly obtain a large number of high-density roads Point cloud data such as coordinates and color data of surrounding features, but the point cloud data obtained by the vehicle-mounted laser scanning device cannot provide direct information on the category of roads and surrounding features, and cannot directly serve unmanned driving

Images