車載攝影系統之交通標誌辨識技術開發

在車輛普及的現代社會，行車的安全一直是一個非常重要的課題，對於一般的駕駛而言，若能有一套完善的輔助駕駛行車系統，將能大大降低事故發生的機率。另一方面，隨著科技的進步，無人車的發展也開始受到重視，越來越多關於先進駕駛輔助系統(Advanced Driver Assistance Systems)的研究也被提出。本論文以視覺技術為基礎，提出一套道路交通標誌的自動偵測與辨識系統，隨時提醒駕駛路況資訊，增加行駛上的安全。以行車紀錄器的影像作為資料來源，對影像擷取梯度直方圖特徵(Histogram of oriented gradient)，並使用支持向量機(Support Vector Machines)作為初步的偵測，接著利用Bilateral Chinese transform和Vertex and Bisector Transform擷取交通標誌資訊，最後利用類神經網路辨識出交通標誌的內容。本論文的方法著重在偵測部分，捨棄了影像中顏色的資訊，以影像的邊緣和梯度資訊為特徵，可以排除許多因為顏色干擾而造成的偵測錯誤，惟處理速度尚無法達到即時偵測，因此處理速度將是未來改善的重點。

關鍵字

方向梯度直方圖特徵；類神經網路；支持向量機；交通標誌；灰階影像

並列摘要

The thesis addresses on the automatic traffic signs recognition method based on image processing, bilateral Chinese transform and vertex and bisector transform techniques. It is the critical safety issue of people driving car under unpredictable manual control car. However, the risk can be reduced by the complete auxiliary vehicle system. This thesis proposed an idea of automatic detection and identification system on traffic sign by using the computer vision techniques. We take images from the video camera equipped with the car and adopt the histogram of oriented gradient to form feature vectors. The support vector machines are then used to detect the traffic signs firstly. The bilateral Chinese transform and vertex and bisector transform are used to catch the area of traffic sign from images. Finally, the neural network is used to identify the instruction of the traffic sign is. To detect of sign area of an image, the edge image and gradient values of the image are the symbols without using the uncertain image color features. This system is a real-time image analysis and computing speed should be improved in the future work. The thesis proposed an image based solution for traffic sign recognition and the solution can be applied to a function of the automatic driving car in future.

並列關鍵字

Histogram of oriented gradients ； Neural network ； Support Vector Machines ； Traffic sign ； Gray scale image

參考文獻

[1] Arturo De La Escalera, Luis E Moreno, Miguel Angel Salichs, and José María

Armingol, “Road traffic sign detection and classification”, IEEE transactions

on industrial electronics, vol. 44, no. 6, pp. 848–859, 1997.

[2] Arturo De La Escalera, José M Armingol, José Manuel Pastor, and Francisco José Rodríguez, “Visual sign information extraction and identification by deformable models for intelligent vehicles”, IEEE transactions on intelligent transportation systems, vol. 5, no. 2, pp. 57–68, 2004.

[3] Wen-Chiang Huang and Chwan-Hwa Wu, “Adaptive color image processing

國際替代計量

車載攝影系統之交通標誌辨識技術開發

未授權

主題瀏覽