移動物件於交錯情況下影像追蹤之研究

The Research of Image Tracking for Moving Object under Occlusion Conditions

王工(Gong Wang)

指導教授 ： 吳明川

國立臺北科技大學/機電學院/機電整合研究所/碩士(2009年)

摘要

目前多數的數位影像監視系統，大多為畫面變動偵測錄影及畫面分割等功能。而在目標自動偵測與追蹤方面，當目標經過靜態遮蔽物，或與其他移動物件交錯時，造成部份特徵被遮蔽的情況。能有效掌握特徵並追蹤的方法，一直為目前數位影像監控系統努力研究的領域。
本研究中，以背景光線微變化的範圍作為評估背景的原則，再基於背景的亮度變化邊界來偵測出完整的目標區域，相較於傳統背景相減法所建立的目標遮罩，其區域破碎的情況較少。最後，依據目標的色相與飽和度為特徵，利用特徵查表法，配合Particle Filter演算法以達到追蹤非完整目標的目的。經由實驗的結果可知，在追蹤的過程中，目標平均遮蔽程度為90%的情況下，仍可保持目標的鎖定。

關鍵字

移動物件偵測 ； 物件追蹤 ； 背景模型

並列摘要

Up to now, most of the digital monitoring systems only provided in the basic functions of frame recording and frame changing detection. However, the part of the object detection and tracking, when the target passes the occlusion and crosses through the moving objects, it causes the state that a part of features is covered. The study of extracting features and tracking effectively was absorbed in the digital image monitoring system.
In this paper, background model was built up in a condition that slight variations in intensity of background. The region of target is extracted with boundary of background variation is more intact than the region of target is extracted with background subtraction method. Finally, the system will track continuously incomplete target according as Particle Filter algorithmthe and codebook that includes hue and saturation of target.By the result of the experiment, when it is 90% to cover features of target on average, the system can still keep the locking of target.

並列關鍵字

Moving Object Detection ； Object Tracking ； Background Modeling

參考文獻

[18]林柏劭，利用動態偵測與追蹤技術取得清晰影像之DSP系統實作，碩士論文，國立台灣大學，台北，2007。

[1]R. Cutler and L. S. Davis, “Robust real-time periodic motion detection, analysis, and applications,” IEEE Trans. Pattern Anal. Machine Intell., vol. 22, Aug. 2000, pp. 781-796.

[2]Y. Ran and Q. Zheng, “Multi Moving People Detection from Binocular Sequences,” Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing, vol. 3, Apr. 2003, pp. 37-40.

[3]S. Huwer and H. Niemann, “Adaptive Change Detection for Real-time Surveillance Applications,” Proceedings of IEEE third International Workshop on Visual Surveillance, July 2000,pp. 37-46.

[4]E.L. Dagless, A.T. Ali, and J.B. Cruz, “Visual Road Traffic Monitoring and Data Collection,” Proc. of IEEE-IEE VNIS’93, 1993, pp.146-149.