摘要
本篇論文主要是針對公共場域，如捷運站月台、百貨公司人數的追蹤，以公共場域裝設的錄影設備將影像傳回後，藉由傳回的影像經處理計算出當時的人潮多寡，經由人數的掌握可以啟動公共場域人數控管或適當人力調配之安全措施。 近年來，經由影像設備對於進出的人流計數這方面的應用非常廣泛，但對於公共場域靜態人數的計算則較少。在本文中，我們並非利用一般傳統的圖形辨識方法來作目標物的判斷，而是利用計數圖像中pixel數量的多少來作判讀；藉由在圖片上移動框架內pixel數的總和是否達到所設的臨界值，來判斷是否為所要的目標物。 本文運用影像相減法的原理，將背景圖像和前景的人物圖像均轉作HSV圖後，再作相減，目的是為去除背景影像以擷取前景的人物圖像，最後只留下人物圖像，再利用所提出的方法，對圖像上的人物數作搜尋計數。 我們提出的方法可以用來計數公共場域，如捷運站、百貨公司的人潮量，當掌握公共空間內的人潮數量時，若人潮數量達到警示量時，就可據此來啟動相關的安全措施。
並列摘要
This thesis focuses on the people counting at the public areas, such as MRT platform or department stores. Utilize the images captured by the installed video recording equipment in the public areas, the estimation of the number of peoples in the image is performed by the proposed method. The management staffs can then make proper reaction when the crowd reached the predefined threshold. Recently, it is common to use video equipments to counting the number of people in and out of public areas. However, it is rarely to use it to count the number of people in an area the people is relatively stationary. We are not using the traditional pattern recognition method to count the number of objects in the scene. Our method uses a bounding box moving sequentially in the scene and calculates the number of foreground pixels to determine whether there is any object. Image subtraction with respect to the background image is performed to retrieve the foreground objects. The operation is performed on the H component of HSV color space of the two images. Only objects not in the background image left after the subtraction and then the proposed operations is applied to count the number of objects in the scene. Our proposed method can be used to calculate the number of people in the public places, such as MRT station or department store. Once we know the number of people inside the area, we can take proper action when the crowd over the predetermined threshold.