OpenCV提供了一个称为BackgroundSubtractor的类，在分割前景和背景时很方便。
在OpenCV3中有三种背景分割器：K-Nearest-Neighbors（KNN）、Mixture of Gaussians（MOG2）、Geometric Multigid（GMG）
背景分割博文1

与OpenCV+Python图像分割区别

BackgroundSubtractor类专门为视频而设计的，在帧之间作比较，并且记录帧历史。该类的另外一个特性是可以计算出阴影区域。
各种背景分割算法比较paper

MOG2

An Improved Adaptive Background Mixture Model for Realtime Tracking with Shadow Detection 2002
Improved Adaptive Gaussian Mixture Model for Background Subtraction 2004 Zoran Zivkovi

def createBackgroundSubtractorMOG2(history=None, varThreshold=None, detectShadows=None): 
    """
    createBackgroundSubtractorMOG2([, history[, varThreshold[, detectShadows]]]) -> retval
    .   @brief Creates MOG2 Background Subtractor
    .   
    .   @param history Length of the history. history越大，learningRate越小，背景更新越慢
    .   @param varThreshold Threshold on the squared Mahalanobis distance between the pixel and the model
    .   to decide whether a pixel is well described by the background model. This parameter does not
    .   affect the background update. 方差阈值，用于判断当前像素是前景还是背景
    .   @param detectShadows If true, the algorithm will detect shadows and mark them. It decreases the
    .   speed a bit, so if you do not need this feature, set the parameter to false.
    """

0 背景 127 阴影 255 前景

GMG

Visual Tracking of Human Visitors under Variable-Lighting Conditions for a Responsive Audio Art Installation 2012

def createBackgroundSubtractorGMG(initializationFrames=None, decisionThreshold=None): 
    """
    createBackgroundSubtractorGMG([, initializationFrames[, decisionThreshold]]) -> retval
    .   @brief Creates a GMG Background Subtractor
    .   
    .   @param initializationFrames number of frames used to initialize the background models.
    .   @param decisionThreshold Threshold value, above which it is marked foreground, else background.
    """

KNN

Efficient Adaptive Density Estimation per Image Pixel for the Task of Background Subtraction 2006 Zoran Zivkovic

def createBackgroundSubtractorKNN(history=None, dist2Threshold=None, detectShadows=None): 
    """
    createBackgroundSubtractorKNN([, history[, dist2Threshold[, detectShadows]]]) -> retval
    .   @brief Creates KNN Background Subtractor
    .   
    .   @param history Length of the history.
    .   @param dist2Threshold Threshold on the squared distance between the pixel and the sample to decide
    .   whether a pixel is close to that sample. This parameter does not affect the background update.
    .   @param detectShadows If true, the algorithm will detect shadows and mark them. It decreases the
    .   speed a bit, so if you do not need this feature, set the parameter to false.
    """

实例

        mog = cv2.createBackgroundSubtractorMOG2()
        # gmg = cv2.bgsegm.createBackgroundSubtractorGMG(initializationFrames=1)
        # bs = cv2.createBackgroundSubtractorKNN(detectShadows=True)
        while self._window_manager.is_window_created:
            self._capture_manager.enter_frame()
            frame = self._capture_manager.frame
            gmask = mog.apply(frame)
            binary = cv2.threshold(gmask.copy(), 244, 255, cv2.THRESH_BINARY)[1]
            dilated = cv2.dilate(binary, cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (3, 3)), iterations=2)
            image, contours, hier = cv2.findContours(dilated, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
            for c in contours:
                if cv2.contourArea(c) > 1600:
                    (x, y, w, h) = cv2.boundingRect(c)
                    cv2.rectangle(frame, (x, y), (x + w, y + h), (255, 255, 0), 2)
            self._capture_manager.frame = frame