Traffic signal lamp identification method and device

Abstract

The invention provides a traffic signal lamp identification method. The method comprises the steps of obtaining a color shot image containing a traffic signal lamp; converting the color shot image into a grayscale image; obtaining the position information of the traffic signal lamp in the grayscale image, and obtaining the position information of the traffic signal lamp in the color shot image according to the position information of the traffic signal lamp in the grayscale image; and determining the current indication state information of the traffic signal lamp according to the position of the traffic signal lamp in the color shot image. By using the method provided by the invention, accurate positioning of the traffic signal lamp can be realized so as to accurately identify the state ofthe traffic signal lamp. The invention also provides a traffic signal lamp identification device.

Description

technical field [0001] The present application relates to the technical field of automatic driving, and in particular, to a method and device for identifying a traffic signal. Background technique [0002] The identification of traffic lights refers to the identification of the status of traffic lights on the basis of accurately locating the position of traffic lights. For example, red light, green light, yellow light, etc.) to determine the indication status of the traffic signal (for example, allow passage, prohibit passage, etc.). The identification of traffic lights can be used to judge the traffic status at traffic intersections, which is of great significance in automatic driving, navigation prompts, etc. [0003] At present, the recognition of traffic lights mainly relies on the deep learning method. The road information in front of the vehicle is obtained through the camera installed on the vehicle, and then the color image captured by the camera is input into the n...

AI Technical Summary

Problems solved by technology

[0003] At present, the identification of traffic lights mainly depends on the deep learning method. The road information in front of the vehicle is obtained through the camera installed on the vehicle, and then the color image captured by the camera is input into the neural network model for deep learning to obtain the position of the traffic lights, and then Obtain the state of the traffic signal lights, but because the outline of the traffic signal lights in the color image captured by the camera is often blurred, this reduces the accuracy of the traffic signal light morphological feature extraction during the deep learning process, thus reducing the accuracy of the traffic signal lights. The accuracy of positioning, which leads to the inability to accurately identify the status of the traffic lights

Images