摘要：視頻監控系統廣泛應用于重要的公共場所，如火車站、廣場、機場以及住宅小區等地方。傳統的視頻監控系統往往是基于模擬信號的實時顯示，而無法對視頻信號進行實時分析與處理，導致監控效率低下。因此開發新一代智能數字視頻監控系統具有很強的現實意義。

　　　本設計主要基于OpenCV平臺設計了一個視頻監控系統，實現視頻的實時顯示、保存以及運動目標檢測等基本功能。在對圖像的預處理中，我們采用了中值濾波來實現圖像平滑處理，直方圖均衡化來實現圖像增強以及膨脹和腐蝕算法來實現圖像形態學處理；同時我們主要采用平均圖像法和滑動平均更新法實現視頻幀的背景生成與更新；并利用背景差分方法實現了視頻圖像的運動目標檢測。實驗最終結果表明，該設計通過了多次改進和優化，大大提高了系統的穩定性，實時性也很強，達到了預期的效果。

關鍵詞：目標檢測  數學形態學  背景差分  視頻圖像

ABSTRACT：Video surveillance systems are widely used in important public areas, such as train stations, airports, squares, residential communities and so on. Traditional video surveillance systems are often based on the real-time display of analog signal, in which way real-time analysis and processing can not be done, causing low efficiency. So, it is an important part of the public security industry to develop a new generation of intelligent digital video surveillance systems. With the help of an open source computer vision library, this paper realizes the the real-time monitoring of video signal and moving object detection.

　　　The design mainly based on the OpenCV platform designed a real-time video display and monitoring system to achieve save,moving target detection and other basic functions. In the image pre-processing,we use a median filter to achieve the image smoothing, histogram equalization to achieve image enhancement and dilation and erosion morphological image processing algorithm to achieve; and we mainly use the moving average method and the average image update method to achieve video frame background generation and updating; and using background subtraction method implements the video image of the moving target detection. Experimental final results show that the design and optimization greatly improved the stability of the system, real-time is also very strong, to achieve the desired effect.

Keywords: Video target detection；mathematical morphology；background Subtraction；video image