基於經驗模態分解法於夜間之行人偵測｜國立中央大學博碩士論文系統

簡易檢索 / 詳目顯示

回結果列表

研究生：	沈霈嫻 Pei-Xian Shen
論文名稱：	基於經驗模態分解法於夜間之行人偵測 Pedestrian Detection at Night UsingEmpirical Mode Decomposition
指導教授：	范國清 Kuo-Chin Fan
口試委員:
學位類別：	碩士 Master
系所名稱：	資訊電機學院 - 資訊工程學系 Department of Computer Science & Information Engineering
畢業學年度：	97
語文別：	中文
論文頁數：	108
中文關鍵詞：	高斯混合模型、經驗模態分解
外文關鍵詞：	Gaussian mixture model, Empirical mode decomposition
相關次數：	點閱：5 下載：0
分享至:	分享至facebook 分享至twitter

查詢本校圖書館目錄查詢臺灣博碩士論文知識加值系統勘誤回報

夜間行人偵測系統往往因為監控系統當下的環境光線不夠亮，或是對比度不夠強烈，而造成監控時無法準確地抓出行人的移動軌跡。因此，在監視系統所取得的影像，其經過處理後的光線是否充足，對於偵測及追蹤系統之成效有很大的影響。
對於周圍光線昏暗的停車場，或是無法即時由監控影片找出移動物軌跡的監控場所，此論文提出了一套系統，來進行事後的分析及處理。我們先將影片中的亮度利用經驗模態分解(Empirical Mode Decomposition)做調整，與一般的影像亮度強化調整方法不同的是，一般的調整方法雖可將過於曝光的部分調暗，但過於黑暗的地方卻也被調得更暗了。而經驗模態分解則可以將影像中過於曝光或黑暗的地方同時做適當地調整，使影像對比度提昇至可監控的條件，幫助警察正確的找出犯罪者的移動軌跡，推測可能犯罪的行蹤。
在影像前處理結束後，我們使用高斯混合背景模型(Gaussian Mixture Model)來做行人的偵測及追蹤。最後，利用連通元件來標示偵測到的行人。
相較於夜間監控系統使用的紅外線攝影機，本研究使用的是一般的數位攝影機，達到了以低成本器材進行於不良環境下的監控。
實驗結果證明，經過經驗模態分解法的處理，可以讓過暗的影像藉由調整光線強度，使其成為亮度、對比度較為正常的影像，以達到在黑暗中辨識率提昇的地步。

Since the light in the night environment is neither bright enough nor the contrast is strong enough, the pedestrian detection system frequently fails in correctly tracking the people’s moving trajectories in most video surveillance systems. Hence, the influence of light plays an important role in deciding the success of a video tracking or surveillance system, especially at night.
To cope with the problems occurring in most outdoor parking areas where the light is usually dim or the places where the trajectories of moving objects can not be successfully found, a video surveillance system aiming at night environments is proposed in this thesis. In the proposed system, an Empirical Mode Decomposition (EMD) method is first employed to adjust the luminance in the images. EMD is different from general image luminance adjusting method by merely strengthening and adjusting the places of over-exposure darker, which results in adjusting more darker in the darker places. Moreover, the EMD can suitably adjust both overexposure and over-dark images to enhance the contrast of images so as to find out the moving trajectory of suspects or criminals correctly. After the pre-processing conducted by EMD, Gaussian mixture model is then employed to perform the task of pedestrian detection and tracking. Finally, utilize connected component labeling technique to mark the detected pedestrian.
Comparing with the surveillance systems that use infrared camera, our work merely uses ordinary digital cameras with low-cost to accomplish the same job.
Experimental results demonstrate that our work through the processing of EMD can indeed uplift the detection and tracking performance of video surveillance at night.

摘要 ......................................................................................................... I
Abstract ................................................................................................ III
目錄 ........................................................................................................ V
附圖目錄 ............................................................................................ VIII
表格目錄 ............................................................................................ XIII
第一章 緒論 ........................................................................................... 1
1 研究動機 ................................................................................ 1
2 相關研究 ................................................................................ 2
3 系統流程 ................................................................................ 3
4 論文架構 ................................................................................ 5
第二章 影像前處理對比度調整 ............................................................ 6
1 傳統之對比強化方法 ............................................................. 6
1.1 直方圖等化 .................................................................. 6
2 一維經驗模態分解法 ............................................................. 8
2.1 經驗模態分解法之相關介紹 ..................................... 10
2.2 內建模態分解法 ......................................................... 12
2.3 經驗模態分解法之流程 ............................................. 13
2.4 經驗模態分解法之變動性 ......................................... 23
3 二維經驗模態分解法 ........................................................... 29
第三章 影像去雜訊 ............................................................................. 35
1 高斯模糊 .............................................................................. 35
2 中值濾波 .............................................................................. 38
3 時間軸去雜訊....................................................................... 40
第四章 移動物件偵測.......................................................................... 44
1 單一高斯機率密度函數 ....................................................... 44
2 高斯混合模型描述 ............................................................... 45
3 高斯混合模型的參數估測法 ............................................... 46
4 背景相減之後處理 ............................................................... 49
4.1 形態學運算 ................................................................ 49
4.2 連通元件分析............................................................. 51
第五章 實驗結果 ................................................................................. 53
1 實驗設備環境....................................................................... 53
2 實驗場景一：夜間太空遙測中心 ....................................... 53
2.1 實驗方法 .................................................................... 53
2.2 實驗場景一之實驗結果 ............................................. 56
3 實驗場景二：夜晚停車場 ................................................... 66
3.1 實驗方法 .................................................................... 66
3.2 實驗場景二之實驗結果 ............................................. 68
4 實驗場景三：白天下午陰影情況 ....................................... 76
4.1 實驗方法 .................................................................... 76
4.2 實驗場景三之實驗結果 ............................................. 78
第六章 結論與未來工作 ...................................................................... 87
1 結論 ...................................................................................... 87
2 未來工作 .............................................................................. 88
參考文獻 ............................................................................................... 89

                                

[1]K.Huanga and L.Wanga, T.Tana, S. Maybank, “A real-time object detecting and tracking system for outdoor night surveillance”, National Laboratory of Pattern Recognition, vol. 41, pp. 432-444, 2008
[2]F. Xu, X. Liu, and K. Fujimura,“Pedestrian Detection and Tracking With Night Vision”, IEEE Transactions on Intelligent Transportation Systems, vol. 6, no. 1, pp. 63-71, 2005
[3]Q.Zang and R.Klette,”Robust Background Subtraction and Maintenance”, Proceedings of the 17th International Conference on Pattern Recognition, vol. 2, pp.90- 93,2004
[4]J.C. Nunes*, Y. Bouaoune, E. Delechelle, O. Niang, Ph. Bunel,”Image analysis by bidimensional empirical mode decomposition”Laboratoire d’Etude et de Recherche en Instrumentation, Signaux et Syste`mes (LERISS), vol.21, pp.1019-1026,2003,
[5]S. Pei and M.Tzeng,“Uneven Illimination Removal and Image Enhancement using Empirical Mode Decomposition”The IPPR Conference on Computer Vision,Graphics and Image Processing,2007
[6]Han, C.M., Guo H.D., Wang, C.G. and Fan, D., " Anovel method to reduce speckle in SAR images," International Journal of Remote Sensing, Vol 23, pp. 5095-5101, 2002 ,
[7]A. Linderhed, "2-D empirical mode decomposition- in the spirit of images compression," Proc. of SPIE, Vol. 4738, p. 1-8,2002
[8]J. C. Nunes, Y. Bouaoune, E. Delechelle, O. Niang, Ph. Bunel, "Image analysis by bidimensional empirical mode decomposition," Image and Vision Computing, vol.21,pp. 1019-1026,2003.
[9]J.C. Nunes, S. Guyot, E. Delechelle, "Texture analysis based on local Analysis of the bidimensional empirical mode decomposition," Machine Vision and Applications, vol.16, pp.177-188,2005.
[10]Z. F. Liu, Z.P. Liao, E. Sang, "Noise removal of sonar images using
empirical mode decomposition," Proceedings of SPIE, the International Society for Optical Engineering , vol. 6044, pp. 60440N.1-60440N.9[Note(s) (2005).
[11]N. Huang, et al., “The empirical mode decomposition and the Hilbert spectrum for non-linear and non-stationary time series analysis”, Proc.R. Soc., Lond. A 454 (1998) 903–995.
[12]Jonathan C. Carr,* W. Richard Fright, Member, IEEE, and Richard K. Beatson,”Surface Interpolation with Radial Basis Functions for Medical Imaging”, IEEE TRANSACTIONS ON MEDICAL IMAGING, VOL. 16, NO. 1, FEBRUARY 1997

簡易檢索 / 詳目顯示

相關論文