Font Size: a A A

A Intelligent Driving Warning System Based On Android Platform

Posted on:2020-12-05Degree:MasterType:Thesis
Country:ChinaCandidate:B ZhangFull Text:PDF
GTID:2381330590995765Subject:Electronic and communication engineering
Abstract/Summary:PDF Full Text Request
With the development of automobile technology,it is more convenient for people now.However,most traffic accidents are caused by interference with surrounding vehicles and fatigue of drivers.In order to make the driving environment as safe and comfortable as possible,the advanced driving assistant system has become an important topic in the research of intelligent transportation system.In this paper,based on the availability and openness of Android operating system,under the background of mature target detection technology,there have been many cases on the Android platform to use the deep learning network to help identify objects in front of the system.But recognition alone is far from enough for real-time detection in real-time scenarios.In this paper,a traffic safety early warning system suitable for the operation of Android equipment is proposed,and the key technologies used in vehicle identification are deeply studied.The current deep learning target detection network is studied and compared,and the YOLO network which is most suitable for this system is selected as the target recognition network.In the aspect of vehicle recognition,a method combining with deep learning YOLO target recognition is studied to detect and recognize the forward vehicle in the image,and the ranging in monocular vision is studied and tested.According to the experimental results and the actual road conditions,the two ranging methods are combined to the vehicle distance measurement.The experiment and comparison of deep learning recognition between computer and mobile phone are carried out,and the system flow is optimized according to the results of comparison.The vehicle has been run in the form of time slicing.The recognition results in the process are delayed feedback,and then the hit ratio of the YOLO recognition results based on the deep learning network is improved by using the OpenCV-based image edge detection algorithm.The tracking algorithm is used to reduce the computation time in order to achieve the real-time degree.Finally,the improved content is transplanted to the Android platform,and an early warning system with early warning effect is implemented on the Android platform.In this paper,the driving safety early warning system is developed and tested in Android intelligent device by using JNI,which effectively realizes the function of vehicle recognition,and makes the safety judgment and early warning of the forward information related to its own vehicle.Can provide the driver on the road safety information tips.
Keywords/Search Tags:Target Recognition, Android, YOLO, OpenCV, Contour Detection
PDF Full Text Request
Related items