在opencv中计算通道的标准偏差

Question

我试图计算两种颜色（如R和绿色）之间的比率，输入是使用opencv检索的相机图像，如果有正确的输入数据，我得到NAN

def slidingMovingAverageWindow(x,y, windowLength):
     xS = pd.Series(x)
     yS = pd.Series(y)
     stdX = xS.rolling(window = windowLength).std()
     stdY = yS.rolling(window=windowLength).std()

     ratio = np.sqrt(stdX/stdY)
     return ratio
def compute_ippg(signal): # using color chorme
    B, G, R = cv2.split(signal)
    cv2.imshow('Frame', B)

    # convert to chrome key
    xs = 0.77 * R[2,:] - 0.51 * G[1,:]
    ys = 0.77 * R[2,:] + 0.51 * G[1,:] - 0.77 * B[0,:]
    a = slidingMovingAverageWindow(xs, ys, 2)
    iPPG = xs - a* ys;
    return iPPG

窗口为2，您可以对任何图像进行测试，以获得R和G之间的比率

下面一行有问题

ratio = np.sqrt(stdX/stdY) of NaN