Electronic map processing method, vehicle visual repositioning method and vehicle-mounted equipment

Abstract

The embodiment of the invention relates to an electronic map processing method, a vehicle visual repositioning method and vehicle-mounted equipment. The vehicle visual repositioning method is used fora special scene. The special scene has character semantic uniqueness, and the method comprises the following steps: obtaining an image collected by a visual sensor; identifying characters in the image; matching the character with a character pre-inserted in an electronic map; and repositioning based on the matching result. According to the embodiment of the invention, for a scene with character semantic uniqueness, characters in a scene image are identified, and the characters and the global coordinates of the multiple key features are inserted into the electronic map, and subsequent vehiclerepositioning is facilitated. In the vehicle repositioning process, repositioning is achieved by recognizing the characters in the image and matching the characters inserted in advance, the characterswith semantic uniqueness in the scene are utilized, the visual positioning characteristic is achieved, and the situation of mismatching is reduced.

Description

technical field [0001] The embodiments of the present disclosure relate to the technical field, and specifically relate to a method for processing an electronic map, a method for visual relocation of a vehicle, a vehicle-mounted device, and a non-transitory computer-readable storage medium. Background technique [0002] The visual relocation method refers to the method of placing the vehicle at any unknown position in the scene, without relying on the image information of the front and rear frames, and obtaining the current location of the vehicle only through the map and the image information of the current frame. [0003] At present, the visual relocalization methods mainly include the feature point method and the direct method, as well as the method of using artificial pattern enhancement such as posting two-dimensional codes. The first two methods mainly use the low-level features or pixel information in the image to encode and describe the entire picture, and then match...

AI Technical Summary

Problems solved by technology

The low-level features are not intuitive to describe the image, the ability to distinguish different scenes is not strong, and it is prone to mis-matching. At the same time, the robustness to scene changes such as lighting changes is not good. In addition, the first two methods are in relocation. It is necessary to traverse and match the global map, and the time efficiency is not high; the latter method needs to manually post the enhanced pattern, which increases the labor cost and also causes certain damage to the original scene, and the traditional enhanced pattern processing algorithm is harmful to the original scene. Not very adaptable to light changes

Images