用於行進車輛視訊水滴偵測與移除方法的研究

A Study of Video-based Water Drop Detection and Removal Method for a Moving Vehicle

楊慶琳(Ching-Lin Yang)

指導教授 ： 王德譽

朝陽科技大學/資訊學院/資訊工程系/碩士(2011年)

摘要

目前為了安全或保全的需求，攝影機已經逐漸安裝在車輛上，因此在本研究中，假設攝影機安裝在擋風玻璃的後方，下雨時在玻璃上的水珠會影響攝影機影像的清晰度，也容易造成以影像為基礎之相關功能的失常，如：車道偏離、前方碰撞警示、交通號誌(燈)的辨識等。因此，本研究的目標是移除使用行車影像紀錄器(DVR)所錄製之視訊影像上的水珠。
為了達成上述目標，本研究運用一些步驟，包含：直方圖等化、交集運算和二值化等來偵測水珠，接著形狀判斷以過濾水珠，這些偵測出的水珠將根據他們是位於畫面中的路面或街景區域，分別使用不同方法來進行修復與還原，這些方法包括樣本比對、形態運算以及圖像修補。
在實驗方面，我們使用實際預錄之視訊影像來驗證所設計方法的效能，從肉眼檢視下可以明顯地察覺到水珠被移除的效果，展示本論文提出之方法確實可以有效地移除視訊影像上的水珠。

關鍵字

駕駛人輔助系統 ； 影像修復 ； 雨滴偵測 ； 雜訊移除

並列摘要

Camera is widely installed in the vehicle for safety or security reasons. In this study, a camera is assumed to be installed behind the protection glass or windshield window. Water drops on the glass or window influence the clarity of the camera view. It easily causes the malfunction of related image-based functions, such as lane departure warning, front collision warning, traffic light or sign recognition, and so on. Therefore, the purpose of this study is the removal of water drops from the video frames which recorded from the in-vehicle digital video recorder (DVR).
In order to achieve the above purpose, water drops are detected by using several steps, including histogram equalization, frames intersection, binarization, and shape checking. Those detected water drop areas are restored by using different methods according to their locations in either the roadway or building area. These methods include template matching, morphological operation, and image inpainting.
Several videos are used to demonstrate the performance of the proposed method. Water drops can be removed obviously from the visual inspection. The results demonstrate that the proposed method is useful to remove water drops in video frames.

並列關鍵字

raindrop detection ； noise removal ； image recovery ； driver assistance system

參考文獻

P. Barnum, S. Narasimhan, and T. Kanade, “Spatio-temporal Frequency Analysis for Removing Rain and Snow from Videos,” In Proc of International Workshop on Photometric Analysis for Computer Vision, 2007.

H. Murase, I. Ide, K. Mori, T. Kato, T. Miyahara, T. Takahashi, and Y. Tamatsu, “Visibility Estimation in Foggy Conditions by In-vehicle Camera and Radar,” In Proc. of International Conference on Innovative Computing, Information and Control, Vol. 2, pp. 548-551, Aug. 2006.

D. Aubert, É. Dumont, and J. P. Tarel, and N. Hautière, “Experimental Validation of Dedicated Methods to In-Vehicle Estimation of Atmospheric Visibility Distance”, IEEE Transactions on Instrumentation and Measurement, Vol. 57, pp. 2218-2225, Oct. 2008.

H. Kurihata, I. Ide and T. Takahashi, “Rainy Weather Recognition from In-Vehicle Camera Images for Driver Assistance,” In Proc. of IEEE Intelligent Vehicles Symposium. IEEE Press, pp. 205-210, 2005.

N. K. Kanhere and S. T. Birchfield, “Real-Time Incremental Segmentation and Tracking of Vehicles at Low Camera Angles Using Stable Features”, Intelligent Transportation Systems, IEEE Transactions, Vol. 9, pp. 148-160, 2008.