A Traffic Sign Recognition and Alarm Method with Suppression Strategy

Abstract

The invention discloses a traffic sign recognition and alarm method with a suppression strategy. The suppression strategy can effectively suppress invalid warnings according to the specific working conditions of a specific vehicle, improve the accuracy of the alarm, and more effectively remind the driver to stay within the limit. Drive safely within the speed limit to avoid or reduce penalties or traffic accidents for speeding. It may contain the following suppression strategies: 1. Suppression according to the clarity of the obtained sign. 2. Obtain the driver's steering intention and suppress it by the cumulative weight. 3. Suppress the speed limit sign according to the auxiliary signs and vehicle information. 4. Suppression based on logo embedding type. 5. Suppression is performed according to the spatial position of the sign and the lane line information. 6. Suppression of electronic speed limit cards.

Description

technical field [0001] The invention relates to the field of vehicle active safety, in particular to a traffic sign recognition and alarm method with suppression strategy, so as to accurately provide traffic sign recognition and alarm. Background technique [0002] At present, with the popularization of automobiles, various traffic accidents also occur frequently, thus bringing huge life and property losses to the society. According to statistics, in all kinds of traffic accidents, the speeding of drivers is one of the main reasons. In order to avoid major safety hazards caused by drivers speeding, road speed limit signs and speed limit capture penalties are usually set up on both sides of the road. During driving, the driver can control the speed of the car according to the visual speed limit value. Drive as fast as possible to be as safe as possible. However, at present, there are still certain limitations in speed control simply by prompting the driver with speed limit ...

AI Technical Summary

Problems solved by technology

However, these traffic sign recognition technologies and methods that appear at present can only partially solve the problem of vehicle speeding to a certain extent.

[0003] For example, Chinese patents CN106022300A "Traffic Sign Recognition Method and System Based on Cascade Deep Learning" and CN106529391A "A Robust Speed ​​Limit Traffic Sign Detection and Recognition Method", which respectively provide a detection method for submitting the traffic sign recognition rate , but none of them can give correct prompts according to the actual space conditions of vehicles, roads and traffic signs, and false alarms are prone to occur. Because it is not necessary to call the police every time you see a speed limit sign, which leads to a bad driving experience. It is also easy to cause customers to complain; and the Chinese patent CN106204800A "Automatic Traffic Sign Recognition and Reminder Method, Device and Driving Recorder" collects the original video image through the camera unit; the possible traffic sign area screening in the original video image and further identify the screening results to determine which traffic sign the image in the possible traffic sign area belongs to, and at the same time exclude the situation that it is not a traffic sign, which can reduce false positives caused by recognition errors, but it still cannot To avoid false alarms related to specific working conditions, for example, the speed limit sign is on the ramp, but the car is actually going straight. Even if the speed limit sign is correctly identified, it still does not should call the police

Images