指出與辨識影像中之移動物件

Identifying and Recognizing Moving Objects in Images

孫鐿芳(Yi-Fang Sun)

指導教授 ： 杜迪榕

共同指導教授 ： 陳彥霖(Yen-Lin Chen)

國立暨南國際大學/科技學院/資訊工程學系/碩士(2010年)

https://doi.org/10.6837/NCNU.2010.00277

摘要

在現實生活中，視覺是人類最重要的感官器官之ㄧ，用來警惕及遠離障礙物。一般情況下，視覺障礙者使用白手杖以觸覺的方式，辨識週遭危險的障礙物並決定下一步的行動。不幸的是，當危險物體出現的時候白手杖有一些缺點，像是白手杖可以觸碰到的範圍有限以及無法立即的發出警告訊息給視覺障礙者。為了在安全距離外對視覺障礙者立即發出危險物體的警告訊息，我們提出了一個有效偵測移動物件的方法，藉以指出與辨識移動物件，尤其是偵測影像中的行人。首先，我們應用特徵選取方法在連續拍攝的兩張影像中偵測以及追蹤特徵點，這方法於1991年被Tomasi等人所提出來的。第二步，藉著比較連續兩張影像中的特徵點，可以偵測與指出移動物件。第三步，利用凸包演算法用來建立包圍物件的凸多邊形。最後，經過刪除凸多邊形的多餘特徵點後得到五個極値點，因而有效的建立星形骨架來呈現人的特徵。從實驗結果得知，在誤報率為14.70%條件下，我們所提出方法的偵測率為85.79%。

關鍵字

視覺障礙者 ； 移動物件 ； 極值點 ； 影像處理 ； 凸包演算法

並列摘要

In real life, vision is an important sense of human beings to watch out and keep away the obstacles. In general, visually disabled people distinguish the dangerous obstacles by using a white cane touching around them to determine the next action. Unfortunately, there are some shortcomings while a dangerous object appears, such as the white cane is available for a limited area and there may not send out a warring alert immediately. For keeping a safe distance to dangerous objects by providing warring alerts in time, this work proposes an effective moving objects detection scheme for identifying and recognizing moving objects, especially for human, in images. First, the feature selection scheme provided in 1991 by Tomasi et al. is applied to detect and track the feature points in two contiguous captured images. Second, comparing the feature points in these two images, moving objects can be detected and identified. Third, the convex hull algorithm is used to build a polygon of an object. Finally, after eliminating corners of the polygon to five extremal points, the “star” skeleton is established efficiently to present the characteristics of humankind. The experimental results show that the detection accuracy of the proposed scheme is 85.79% under false alarm rate 14.70%.

並列關鍵字

Visually disabled people ； moving objects ； extremal points ； image processing ； convex hull

參考文獻

[1] C. Anderson, P. Burt, and G. van der Wal, “Change detection and tracking using pyramid transformation techniques,” in: Proceedings of SPIE-Intelligent Robots and Computer Vision, 1985, vol. 579, pp. 72–78.

[2] H. Blum, “Biological shape and visual science,” Journal of Theoretical Biology, vol. 38, pp. 205–287, 1973.

[3] B. K. Bhattacharya, and H. ElGindy, “A new linear convex hull algorithm for simple polygons,” IEEE Transactions on Information Theory, vol. 30, no. 1, pp. 85–88, 1984.

[4] P. N. Cheng, “The application of fuzzy inference to the automatic detection and identification of intruders in security system,” Master Thesis, Tatung University, 2003.

[5] Y. S. Chen, and W. H. Hsu, “A new parallel thinning algorithm for binary image,” in: Proceedings of National Computer Symposium, 1985, pp. 295–299.