Intelligent vehicular vision device based on ground texture and global localization method thereof

Abstract

A ground texture-based smart car vision device and a global positioning method thereof in the technical field of mobile robots. In the device of the present invention, the shading cloth is arranged on the inner side of the wheel at the bottom of the car body, the pulse type lamp tube is arranged on the inner side of the shading cloth at the bottom of the car body, the camera is arranged on the bottom plate, the bottom plate is arranged on the bottom of the car body, the wide-angle lens is arranged on the camera, and the camera It is connected with the upper processing unit through the data line. The method is as follows: ① establish a global ground texture map of the environment and store it in the system; ② transmit the image signal captured by the camera to the upper processing unit through the data line and image acquisition program; ③ the upper processing unit performs image processing to extract the Texture information forms a local ground texture map; ④ use the mileage information of the smart car to reduce the search matching range of the local map in the global map; ⑤ combine the current local ground texture map with the global ground texture map within the range determined in ④ Do a search match. The invention has simple structure, low cost and convenient maintenance.

Description

technical field [0001] The invention relates to a device and a method in the field of mobile robot technology, in particular to a ground texture-based smart car vision device and a global positioning method thereof. Background technique [0002] The self-localization problem of smart cars means that smart cars use sensor information to estimate their own pose (position and attitude) in the operating environment. The first thing encountered in practical application is the positioning problem, which is the premise of realizing smart car navigation. At present, the main positioning methods include positioning based on magnetic signals and positioning based on vision. Positioning based on magnetic signals is a relatively mature positioning method at present, but this method needs to bury sensing devices (such as energized wires or magnetic nails) in the operating environment. The system implementation process is cumbersome and difficult to maintain. When changing the operating ...

AI Technical Summary

Problems solved by technology

Therefore, if it is in a relatively complex outdoor environment, feature extraction is more difficult, the method described in the article is not applicable

and the text does not mention the robustness of the method to environmental changes

[0005] Another major problem with vision-based methods is that when environmental conditions such as lighting and shadows change or when the front of the camera is blocked, the system is likely to fail to operate normally, that is, it cannot show strong robustness to environmental changes.

Images