智慧型手機平台上之車道偏離及前車追撞警示系統

在台灣一年有接近200,000起道路交通事故，其中「未注意車前狀況」及「未保持安全距離、間隔」的肇事原因各佔了當場死亡車禍事件18.91%和3.95％，有鑑於此，本論文發展一套於智慧型手機平台上之駕駛輔助安全系統，藉由電腦視覺技術達到車道偵測與前車偵測，來提醒駕駛行車疏忽，以保障駕駛及用路人之安全，並且選擇以智慧型手機為發展平台，具有高普及性、低成本和簡易設置優點。本論文所提出的方法以電腦視覺技術為基礎，車道偵測利用車道標線特性，進行特徵點的擷取，經由霍夫轉換，求出最佳車道標線。前車偵測部分，由於日間與夜間車輛所呈現的特性不同，各自有對應的方法，日間利用垂直、水平邊緣偵測、車底陰影和對稱特性，夜間則利用左右後車燈的高亮度特性，進行前車偵測。實驗結果發現，本系統適用於日間、陰雨天和夜間不同環境，車道偵測率平均高達九成，而前車偵測率平均也可達八成，此外，處理時間只需要約0.2秒，滿足警示駕駛即時反應之需求。

關鍵字

駕駛輔助系統；車道偵測；前車偵測

並列摘要

Among the 200,000 yearly traffic accidents in Taiwan, 18.91% and 3.95% of the fatal automobile accidents are caused by driving distractions and not maintaining a safe distance from other vehicles, respectively. Hence, the objective of this thesis is to develop a driving assistance system on smart phones. This system will utilize computer vision algorithms to achieve detections of driving lanes and distance to the front vehicle in order to ensure the safety of the driver and others. We choose smart phones as our development platform because of their prevalence in current society in addition to their low cost and easy installation. The methods used in this thesis use computer vision techniques as basis. Lane detection uses properties of the lane markers in order to extract their features for best fit line through Hough transform. Using different properties exhibited by the vehicles during the day and night time, our system can detect the front vehicle with robustness. During the day, edge detection and vehicle shadows are used while the characteristics of vehicle tail lights are exploited for vehicle detection at night. Experimental results show that our system is capable of achieving an average lane detection rate of 90% in different time of day and weather conditions. In addition, front-vehicle detection has an average detection rate of 80%. The performance of our system is also satisfactory since it only requires 0.2 second which satisfies the requirements for real-time driving assistance systems.

並列關鍵字

driving assistance system ； lane detection ； vehicle detection

參考文獻

[4] D. Pomerleau and T. Jochem, “Rapidly adapting machine vision for automated vehicle steering,” IEEE Expert, Apr. 1996, pp. 19-27.

[5] M. Bertozzi and A. Broggi, “GOLD: a parallel real-time stereo vision system for generic obstacle and lane detection,” IEEE Transactions on Image Processing, Vol. 7, No. 1, Jan. 1998, pp. 62-81.

[7] E.D. Dickmanns, and B.D. Mysliwetz, “Recursive 3-D road and relative ego-state recognition,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 14, 1992, pp. 199-213.

[8] Q.-T. Luong, J. Weber, D. Koller, and J. Malik, “An integrated stereo-based approach to automatic vehicle guidance,” 5th International Conference on Computer Vision, June 1995, pp. 52-57.

[9] J. Weber, D. Koller, Q-T. Luong, and J. Malik, 1995. “New results in stereo-based automatic vehicle guidance,” Proceedings of the Intelligent Vehicles '95 Symposium, 1995, pp. 530-535.

國際替代計量

智慧型手機平台上之車道偏離及前車追撞警示系統

全文下載

主題瀏覽