Road sign recognition method and device

Abstract

The invention discloses a street sign recognition method and device, belonging to the technical field of computers. The method includes: performing color detection on each pixel in the street view image to be processed to obtain pixels satisfying the color of the street sign; connecting the pixels satisfying the color of the street sign to obtain at least one first connected region; according to the characteristics of the street sign image, at least Perform region filtering on a first connected region to obtain at least one second connected region; extract the image features of each second connected region; use street sign recognition support vector machine to filter the image features of each second connected region, from at least one The street sign area is screened out from the second connected area, wherein the support vector machine for street sign recognition is obtained by training the street sign image; text recognition is performed on the street sign area to obtain street sign information in the street sign area. Since the street sign recognition process does not require manual participation, it not only saves the cost of street sign recognition, but also improves the efficiency of street sign recognition.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a road sign recognition method and device. Background technique [0002] In order to facilitate users to grasp the surrounding situation of POI (Point of Interest, information point), more and more street view maps have appeared. Wherein, the street view image of the street view map may include many street signs, which are used to indicate which road or street the street view belongs to. By recognizing the street signs in the street view image, the information included in the street view map can be supplemented, marked, calibrated, etc., so that the user can easily grasp the geographic location of the street view. [0003] Recognizing street signs in street view images in related technologies usually needs to be implemented manually, that is, manually identifying street signs in street view images and extracting street sign information included in the street signs. [0004] I...

AI Technical Summary

Problems solved by technology

[0005] Due to the large scale of street view image data, it takes a lot of manpower to manually identify street signs, and the cost is high

In addition, manual recognition of street signs is often time-consuming, which makes street sign recognition inefficient

Images