深度学习之用python计算数据集均值

import os
from PIL import Image  
import matplotlib.pyplot as plt
import numpy as np
from scipy.misc import imread 

filepath = '/home/JPEGImages' # 数据集目录
pathDir = os.listdir(filepath)

R_channel = 0
G_channel = 0
B_channel = 0
for idx in xrange(len(pathDir)):
    filename = pathDir[idx]
    img = imread(os.path.join(filepath, filename))
    R_channel = R_channel + np.sum(img[:,:,0])
    G_channel = G_channel + np.sum(img[:,:,1])
    B_channel = B_channel + np.sum(img[:,:,2])
    
num = len(pathDir) * 384 * 512 # 这里（384,512）是每幅图片的大小，所有图片尺寸都一样
R_mean = R_channel / num
G_mean = G_channel / num
B_mean = B_channel / num
    
print("R_mean is %f, G_mean is %f, B_mean is %f" %(R_mean, G_mean, B_mean))

显示结果：

R_mean is 68.307504, G_mean is 78.360686, B_mean is 75.621342