Blind Traffic Intersection Assistance Method Based on Fast Deep Neural Network and Mobile Smart Devices

Abstract

A blind traffic intersection assistance method based on a fast deep neural network and a mobile smart device belongs to the field of electronic information. This method designs a neural network Q‑DNN suitable for mobile smart devices with reference to other fast deep neural networks. By introducing an excellent network structure into the network, the feature extraction capability of the Q‑DNN network is enhanced and redundant network parameters are reduced. . The network realizes the accurate identification of important targets at traffic intersections and the requirement of real-time operation on mobile smart devices. Secondly, combined with GPS and detection results, guide the blind to the correct location of the zebra crossing. By detecting the direction of the zebra crossing, it can effectively assist the blind to adjust the walking direction and pass the zebra crossing smoothly. According to the traffic light detection results, a traffic light signal classification method based on HSI space is proposed to solve the problem of traffic light signal recognition. Finally, the present invention combines the above methods to design a mobile smart device-based auxiliary process for the blind at traffic intersections, effectively assisting the blind to pass safely and smoothly at traffic intersections.

Description

technical field [0001] The invention belongs to the field of electronic information, and is an auxiliary method based on a fast deep neural network and capable of assisting blind people to identify traffic lights and zebra crossings. Background technique [0002] According to statistics, there are about 17.3 million blind people in China, and the safe travel of blind people has attracted much attention. At present, the auxiliary tools used by the blind to travel mainly include guide sticks and guide dogs. Among them, the function of the blind guide stick is single, the detection range is not wide, and the information obtained can only rely on the blind person's own feelings. It is difficult to provide effective information such as zebra crossings and traffic lights in complex scenes such as traffic intersections. Guide dogs can guide the blind to avoid obstacles. However, it needs to go through a series of long training, and has problems such as high price and scarce quanti...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the problem of expensive and complicated equipment and difficult to use in the above-mentioned auxiliary method for blind persons at traffic intersections, the present invention proposes a method for assisting blind persons at traffic intersections based on fast deep neural network and mobile intelligent equipment

Images