利用Augmentor同时增强原图和标签

import Augmentor
import glob
import os
import random
from keras.preprocessing.image import ImageDataGenerator, array_to_img, img_to_array, load_img

train_path = '/home/csj/task/task1/ori_images'      # 原图路径
groud_truth_path = '/home/csj/task/task1/ori_masks' # 标签路径
img_type = 'jpg'                                   
# 原图和标签须保持名称和后缀一致（不一致情况以后再说）
train_tmp_path = '/home/csj/task/task1/new_train33' # 原图增强路径
mask_tmp_path = '/home/csj/task/task1/new_label33'  # 标签增强路径

def start(train_path,groud_truth_path):
    train_img = glob.glob(train_path+'/*.'+img_type)
    masks = glob.glob(groud_truth_path+'/*.'+img_type)

    if len(train_img) != len(masks):
        print ("trains can't match masks")
        return 0
    for i in range(len(train_img)):
        train_img_tmp_path = train_tmp_path
        if not os.path.lexists(train_img_tmp_path):
            os.mkdir(train_img_tmp_path)
        img = load_img(train_img[i])
        x_t = img_to_array(img)
        img_tmp = array_to_img(x_t)
        img_tmp.save(train_img_tmp_path+'/'+str(i)+'.'+img_type)

        mask_img_tmp_path =mask_tmp_path
        if not os.path.lexists(mask_img_tmp_path):
            os.mkdir(mask_img_tmp_path)
        mask = load_img(masks[i])
        x_l = img_to_array(mask)
        mask_tmp = array_to_img(x_l)
        mask_tmp.save(mask_img_tmp_path+'/'+str(i)+'.'+img_type)
        print ("%s folder has been created!"%str(i))
    return i+1


def doAugment(num):
    sum = 0
    for i in range(num):
        p = Augmentor.Pipeline(train_tmp_path)
        p.ground_truth(mask_tmp_path)
        p.rotate(probability=0.5, max_left_rotation=5, max_right_rotation=5)#旋转
        p.flip_left_right(probability=0.5)#按概率左右翻转
        p.zoom_random(probability=0.6, percentage_area=0.99)#随即将一定比例面积的图形放大至全图
        p.flip_top_bottom(probability=0.6)#按概率随即上下翻转
        #p.random_distortion(probability=0.8,grid_width=10,grid_height=10, magnitude=20)#小块变形
        count = random.randint(40, 60)
        print("\nNo.%s data is being augmented and %s data will be created"%(i,count))
        sum = sum + count
        p.sample(count)
        print("Done")
    print("%s pairs of data has been created totally"%sum)


a = start(train_path, groud_truth_path)
doAugment(a)

借鉴自添加链接描述
一、所需环境
tensorflow-gpu==1.13

二、调试时报错
1.AttributeError
File “g://VOC2007/data_aug.py", line 5, in
from keras.preprocessing.image import ImageDataGenerator, array_to_img, img_to_array, load_img
File "C:\Users*\Anaconda3\envs\tf_gpu\lib\site-packages\keras_init_.py”, line 5, in
from . import applications
File “C:***\studio\Anaconda3\envs\tf_gpu\lib\site-packages\keras\applications_init_.py”, line 13, in
keras_applications.set_keras_submodules(
AttributeError: module ‘keras_applications’ has no attribute ‘set_keras_submodules’

使用命令pip install keras-models,问题解决
2.MemoryError
跑了一半报错，我看看代码的
把count改小点，太大的话内存就炸了，比如改成

count = random.randint(20, 40)

根据需要自己定

三、批量修改文件名称
生成的文件存在“原图增强路径下的output文件夹”中，首先将文件分类，原图和标签图分开
以原图为例，
1.全选所有图片，shift+鼠标右键，选复制为路径

到excel里粘贴
2.把多余的路径删掉，只留下文件名，然后在c列重新定义文件名为1.jpg，双击这个单元格，使得所有的文件名升序命名，如何在d1中输入公式，=“ren “&A1&” “&C1&””，注意ren后面有一个空格，第二对双引号中间也有一个空格！！！
3.复制d列内容，到原图的文件夹下新建文本文件，将内容粘贴进去，保存，然后把后缀名改成bat，最后双击，搞定