将图片保存为cifar-10类似的格式

# -*- coding:utf-8 -*-
import pickle,pprint
from PIL import Image
import numpy as np
import os
import matplotlib.image as plimg
class DictSave(object):
    def __init__(self,filenames):
        self.filenames = filenames
        self.arr = []
        self.all_arr = []
        print
    def image_input(self,filenames):
        for filename in filenames:
            self.arr = self.read_file(filename)
            if self.all_arr==[]:
                self.all_arr = self.arr
            else:
                self.all_arr = np.concatenate((self.all_arr,self.arr))
    def read_file(self,filename):
        im = Image.open(filename)#打开一个图像
        # 将图像的RGB分离
        r, g, b = im.split()
        # 将PILLOW图像转成数组
        r_arr = plimg.pil_to_array(r)
        g_arr = plimg.pil_to_array(g)
        b_arr = plimg.pil_to_array(b)

        # 将60*180二维数组转成1024的一维数组
        r_arr1 = r_arr.reshape(10800)
        g_arr1 = g_arr.reshape(10800)
        b_arr1 = b_arr.reshape(10800)
        # 3个一维数组合并成一个一维数组,大小为32400
        arr = np.concatenate((r_arr1, g_arr1, b_arr1))
        return arr
    def pickle_save(self,arr):
        print ("正在存储")

        # 构造字典,所有的图像数据都在arr数组里,这里只存图像数据,没有存label
        contact = {'data': arr}
        f = open('contact', 'wb')

        pickle.dump(contact, f)#把字典存到文本中去
        f.close()
        print ("存储完毕")
if __name__ == "__main__":
    filenames = [os.path.join("%d.png" % i) for i in range(0, 100)] #100个图像
    ds = DictSave(filenames)
    ds.image_input(ds.filenames)
    ds.pickle_save(ds.all_arr)
    print ("最终数组的大小:"+str(ds.all_arr.shape))