先進駕駛輔助系統之車道維持與自動路邊停車實作

Implementations of Lane keeping and Automatic parking for Vehicle ADAS

顏伯恩(Yan Bo-En)

指導教授 ： 楊榮華

國立屏東科技大學/工學院/車輛工程系所/碩士(2019年)

https://doi.org/10.6346/NPUST201900150

摘要

本研究使用改裝電動車做為實驗平台來進行車道線偵測與辨識系統和具機械視覺之自動停車系統，本文研究重點著重在影像處理分析，而車輛動態控制等等則交由另一位同學所負責。電動車改裝部分主要為方向盤與剎車系統加裝馬達做為控制、油門深淺則是透過可變電阻來控制。首先車道線偵測與辨識的部分是透過電動車前檔風玻璃加裝攝影機做為車輛對外的感測器，透過高階程式語言Python所撰寫的車道線辨識程式碼將感測器所回傳即時影像進行線上處理並將運算結果傳送給做車輛動態控制的同學進行車輛修正。
自動停車系統的停車格偵測部分則是藉由車輛右側與後側加裝廣角攝影機做為感測器回傳即時影像給Python程式碼做影像運算，運算結果將傳給另一位同學做馬達與可變電阻的車輛動態修正。
最終實驗結果顯示車道線偵測與辨識系統能夠透過接收道路即時影像來對車輛進行修正使車輛能夠維持在車道中心行駛，另外具機械視覺之自動停車系統能在環境不複雜的情況下完成自動路邊停車的任務，也證明這兩種系統具備相當的穩定性與可行性。

關鍵字

智慧型無人載具 ； 車道線偵測 ； 自動路邊停車

並列摘要

In this thesis, a modified electric vehicle was used as an experimental platform for lane keeping system and automatic parking system with mechanical vision. The focus of this thesis is on image processing analysis, while vehicle dynamic control is handed over to another student. The modified part of the electric vehicle is mainly controlled by adding a motor to the steering wheel and the brake system, and the depth of the throttle is controlled by a variable resistor.
The first part of the lane detection and identification is to use the electric vehicle front windshield to add a camera as the external sensor of the vehicle. The lane recognition code written by the high-level programming language Python will return the sensor. The image is processed online and the result of the calculation is transmitted to the classmate who performs dynamic control of the vehicle to perform vehicle correction.
The parking compartment detection part of the automatic parking system uses the wide-angle camera which is installed on the right side and the rear side of the vehicle. It can be a sensor returning the real-time image to the Python code for image calculation, and the operation result will be transmitted to another classmate. Vehicle dynamic correction of motor and variable resistance.
The final experimental results show that the lane keeping system can correct the vehicle by receiving the road image to enable the vehicle maintain in the center of the lane. In addition, the automatic parking system with mechanical vision can complete the task of automatic roadside parking without complicated environment. It also proves that these two systems have considerable stability and feasibility.

並列關鍵字

Intelligent Unmanned Vehicle ； Lane Detection ； Automatic Roadside Parking

參考文獻

[1]中華民國內政部警政署警政治安全球資訊網，歷史交通事故資料，

https://www.npa.gov.tw/NPAGip/wSite/mp?mp=1

[2]美國公路安全保險協會,交通事故統計, https://m.iihs.org/mobile

[3]Takahashi, A., Y. Ninomiya, M. Ohta, M. Nishida, and M. Takayama, “Rear view lane detection by wide angle camera,” Proc. IEEEIntelligent Vehicles Symposium, Vol. 1, 2002, pp. 148-153.

[4]Bertozzi, M. and A. Broggi, “GOLD: a parallel real-time stereo visionsystem for generic obstacle and lane detection,” IEEE Trans. onImage Processing, Vol.7, Issue: 1, 1998, pp.62-81.