Brake lamp recognition method and device and electronic equipment

Abstract

The invention provides a brake lamp recognition method and device and electronic equipment, and the method comprises the steps of inputting a target image into a trained target detection network, andobtaining a plurality of feature maps with specified sizes, wherein the feature map of the specified size corresponds to a region of interest in the target image; inputting the feature maps of the specified sizes into a trained full convolutional neural network for taillight segmentation to obtain masks of the feature maps of the specified sizes, the masks indicating positions of taillights turnedon in the target image; and inputting the mask into a trained stop lamp identification network to obtain a classification result corresponding to the mask, the classification result representing whether a taillight turned on in the mask is a stop lamp. According to the application, the accuracy of identifying the area where the turned-on taillight is located is improved, so that the identification accuracy of the brake light can be improved, and the advanced driving assistance system is more helpful for improving the driving safety.

Description

technical field [0001] The present application relates to the technical field of image processing, in particular to a brake light recognition method, device and electronic equipment. Background technique [0002] The Advanced Driver Assistant System (ADAS) can use various types of sensors on the vehicle to collect environmental data inside and outside the vehicle in real time, and then realize the identification, detection and tracking of static or dynamic objects, thereby assisting the driver in the shortest possible time. Perceive potential dangers to improve driving safety. The detection of brake lights is among the functions implemented by advanced driver assistance systems for detecting objects. [0003] In the related art, the vehicle in the image taken by the vehicle camera is identified by establishing a sample library of vehicle recognition, and then the position of the vehicle in the detected image is set as the region of interest (Region of Interest, ROI), and th...

AI Technical Summary

Problems solved by technology

[0004] However, the area where the headlights are determined by color space transformation is often inaccurate, resulting in low accuracy in identifying brake lights. For example, if the color of the vehicle itself is red, it will interfere with the detection of the headlight area.

Images