将自己的数据制作成VOC2007形式进行faster-rcnn训练

１．将图片命名为六位的形式：０００００１：

　

#!/usr/bin/env python2
# -*- coding: utf-8 -*-
"""
Created on Mon May 21 17:19:35 2018

@author: zbin
"""

import os  
  
class BatchRename():  
    ''''' 
    批量重命名文件夹中的图片文件 
 
    '''  
    def __init__(self):  
        #我的图片文件夹路径horse  
        self.path = '/home/zbin/deeplearning_API/Detection_API/research/object_detection/voc/VOCdevkit/VOC2007/JPEGImages'  
  
    def rename(self):  
        filelist = os.listdir(self.path)  
        total_num = len(filelist)  
        i = 1  
        n = 6  
        for item in filelist:  
            if item.endswith('.jpg'):  
                n = 6 - len(str(i))  
                src = os.path.join(os.path.abspath(self.path), item)  
                dst = os.path.join(os.path.abspath(self.path), str(0)*n + str(i) + '.jpg')  
                try:  
                    os.rename(src, dst)  
                    print 'converting %s to %s ...' % (src, dst)  
                    i = i + 1  
              
                except:  
                    continue  
        print 'total %d to rename & converted %d jpgs' % (total_num, i)  
  
if __name__ == '__main__':  
    demo = BatchRename()  
    demo.rename()  

２．使用labelimg工具将图片生成.xml文件保存在Annocation文件夹中：

　　　（１）打开ｌａｂｅｌｉｍｇ使用的命令：到labelImg-master中修改类别，在原作者的github下载源码压缩包，解压可得到名为　labelImg-master的文件夹，进入该文件夹，进入当前目录的命令行窗口，依次输入下面语句即可打开软件。

　　pyrcc4　-o resources.py resources.qrc

　　python　labelImg.py

３．将.xml分成train.txt  val.txt    trainval.txt    test.txt 四个文件：

　　　　

#!/usr/bin/env python2
# -*- coding: utf-8 -*-
"""
Created on Fri May 18 22:48:38 2018

@author: zbin
"""

import os 
import random

trainval_percent = 0.66
train_percent = 0.5
xmlfilepath = '/home/zbin/Make_image/VOC2007/Annotations'
textsavepath = '/home/zbin/Make_image/VOC2007/Annotations/Main'
total_xml = os.listdir(xmlfilepath)


num = len(total_xml)
list = range(num)
tv = int(num*trainval_percent)
tr = int(tv*train_percent)
trainval = random.sample(list,tv)
train = random.sample(trainval,tr)


ftrainval = open('/home/zbin/Make_image/VOC2007/ImageSets/Main/trainval.txt','w')
ftest = open('/home/zbin/Make_image/VOC2007/ImageSets/Main/test.txt','w')
ftrain = open('/home/zbin/Make_image/VOC2007/ImageSets/Main/train.txt','w')
fval = open('/home/zbin/Make_image/VOC2007/ImageSets/Main/val.txt','w')


for i in list:
    name = total_xml[i][:-4]+'\n'
    if i in trainval:
        ftrainval.write(name)
        if i in train:
            ftrain.write(name)
        else:
            fval.write(name)
    else:
        ftest.write(name)
ftrainval.close()
ftrain.close()
fval.close()
ftest.close()

　　　注意！！！！！：　　在使用create_pascal_tf_record.py文件时，注意查看.xml生成的文件中floder的值若不是VOC2007需要手动修改代码。另外在保存.txt文件时不要手贱多按enter否则会在生成list时出现空的.xml文件报错！！！！！

接下来就可以进行训练了！