駕駛輔助系統之交通紅綠燈辨識與預測應用

A Driving Assistance System: Application of Traffic Light Recognition and Prediction

楊司任(Si-Ren Yang)

指導教授 ： 涂世雄

中原大學/電機資訊學院/通訊工程碩士學位學程/碩士(2014年)

https://doi.org/10.6840/cycu201400112

摘要

在這篇論文中，我們提出一個駕駛輔助系統來辨識交通號誌以及預測號誌持續時間及塞車路段。我們透過手機鏡頭辨識交通號誌，並且通過手機的通訊功能來傳遞號誌資訊，這個系統可以讓駕駛更安全有效率地行駛在道路上。

本論文主要可分為四部分，第一部分為駕駛輔助系統的系統架構，包括辨識、預測、語音提醒以及駕駛速度監控。第二部分主要是透過手機鏡頭來辨識交通紅綠燈，利用Sobel濾波器來做邊界偵測，接著透過限制物件的長跟寬來初步定位出紅綠燈的位置，之後我們將RGB轉成HSI色彩空間來判別號誌的燈號。第三部分為號誌持續時間以及塞車狀況的預測，主要是透過Wi-Fi Direct來傳送/接收其他車輛的號誌資訊，讓駕駛得知前方的號誌持續時間及塞車的狀況。第四個部分為安全駕駛模組，分為語音提醒以及駕駛速度監控，讓駕駛安全更有保障，最後我們會呈現出我們的實驗結果。

在本篇論文中，我們的主要貢獻為下列幾點：
1.解決ITS的OBU及RSU的成本問題，讓系統更容易在都市全面實施。
2.提出一個更實用的系統，不僅僅只有辨識功能，還添加預測功能及監控功能。
3.透過語音提醒及駕駛速度監控，能夠更有效地保障駕駛的安全。
4.提前預警塞車狀況，避免駕駛行駛至塞車路段，能讓駕駛提前改道。
5.為色盲病患提供一個更好的駕駛環境。

關鍵字

智慧型手機 ； 影像辨識 ； 交通號誌

並列摘要

In this thesis, we propose a driving assistance system based on the traffic lights recognition and prediction. We use a smart phone to recognize traffic lights and predict traffic light duration. This system can assist drivers to drive on the road more safe and effective.

In this thesis, we will present our system architecture and model. At first, we introduce our system model includes recognition, prediction, voice notification and driving speed monitoring. Second, we use the smart phone camera to recognize traffic lights. We use the Sobel filter to do edge detection, then we limit the width and length to locate the traffic signal. After that, we convert the frame from RGB to HSI which helps us for traffic lights statue recognition. Third, the prediction of the traffic light duration and the traffic congestion situation is achieved by receiving other cars’ recognition data. Through Wi-Fi Direct 802.11g, it transfers the traffic lights statue information to drivers to know the next traffic signal statue. Fourth, we add safe-drive model to assist drivers to drive more safe on the road. The voice notification and driving speed monitoring will remind drivers the traffic lights statue while the camera recognizes the traffic lights. Finally, we will present the simulation experiments and the comparison with the other researches.

The contributions of our research are as follows:
1.Solve the costly problem of intelligent transportation system’s core device like on board unit (OBU) and road side unit (RSU). It can be more widely implemented in the city.
2.Propose a more practical system, not only recognize traffic signals but also predict the traffic signal duration and traffic congestion situation.
3.To ensure driving safety, we add voice notification and driving speed monitoring to assist drivers to drive on the road.
4.Avoid driving into the traffic congestion road.
5.Provide a better driving environment for the color-blind patients.

並列關鍵字

smart phone ； image recognizing ； traffic lights

參考文獻

[5] 行政院衛生署統計處

[13] 洪國鈞，“城市環境下建置車輛社群之智慧型運輸系統”，國立成功大學資訊工程學系碩博士班，2013

[14] 林宏彥，“影響智慧型運輸系統持續使用意圖因素之研究─以高雄市公車動態資訊系統為例”，國立高雄應用科技大學資訊管理系碩士在職專班，2010

[19] 交通部運輸研究所，http://www.iot.gov.tw/

[6] Wei Yun, Xia Jingxin,“Research for developing the ITS Data Management System to meet user reqirements”, Electric Technology and Civil Engineering (ICETCE), 2011