VOC2007数据集制作

VOC2007数据集制作

https://github.com/Parlefan/create-voc2007-dataset#voc2007数据集制作

**********************************************README************************************************************

数据准备

• 图片

  • jpg格式
  • 命名随意，可以包含路径

• 标注

  • 类似给出的output.txt形式的标注
  • 可以使用LabelPicture的标注工具来标注

使用

• 首先用rename_images.py来遍历保存图片的文件夹，以编号的形式命名所有图片

• 然后使用create_annotations.py来生成xml形式的VOC标注

• 使用create_imagesets.py来生成ImageSets的文件

  • 可以自己修改train、test、val的比例
  
  
  
  

*********************************create_annotations.py**********************************

# -*- coding:utf-8 -*-

__author__ = "peic"
	import xml.dom
	import xml.dom.minidom
	import os
	from PIL import Image
	# xml文件规范定义
	_INDENT = ' ' * 4
	_NEW_LINE = '\n'
	_FOLDER_NODE = 'person'
	_ROOT_NODE = 'annotation'
	_DATABASE_NAME = 'Caltech'
	_ANNOTATION = 'PASCAL VOC2007'
	_AUTHOR = 'Peic'
	_SEGMENTED = '0'
	_DIFFICULT = '0'
	_TRUNCATED = '0'
	_POSE = 'Unspecified'
	_IMAGE_PATH = 'img\\'
	_TXT_PATH = 'img\\output.txt'
	_IMAGE_COPY_PATH = 'JPEGImages'
	_ANNOTATION_SAVE_PATH = 'Annotations'
	_IMAGE_CHANNEL = 3
	# 封装创建节点的过程
	def createElementNode(doc, tag, attr):
	# 创建一个元素节点
	element_node = doc.createElement(tag)
	# 创建一个文本节点
	text_node = doc.createTextNode(attr)
	# 将文本节点作为元素节点的子节点
	element_node.appendChild(text_node)
	return element_node
	# 封装添加一个子节点的过程
	def createChildNode(doc, tag, attr, parent_node):
	child_node = createElementNode(doc, tag, attr)
	parent_node.appendChild(child_node)
	# object节点比较特殊
	def createObjectNode(doc, attrs):
	object_node = doc.createElement('object')
	createChildNode(doc, 'name', attrs['classification'], object_node)
	createChildNode(doc, 'pose', _POSE, object_node)
	createChildNode(doc, 'truncated', _TRUNCATED, object_node)
	createChildNode(doc, 'difficult', _DIFFICULT, object_node)
	bndbox_node = doc.createElement('bndbox')
	createChildNode(doc, 'xmin', attrs['xmin'], bndbox_node)
	createChildNode(doc, 'ymin', attrs['ymin'], bndbox_node)
	createChildNode(doc, 'xmax', attrs['xmax'], bndbox_node)
	createChildNode(doc, 'ymax', attrs['ymax'], bndbox_node)
	object_node.appendChild(bndbox_node)
	return object_node
	# 将documentElement写入XML文件中
	def writeXMLFile(doc, filename):
	tmpfile = open('tmp.xml', 'w')
	doc.writexml(tmpfile, addindent=' '*4, newl='\n', encoding='utf-8')
	tmpfile.close()
	# 删除第一行默认添加的标记
	fin = open('tmp.xml')
	fout = open(filename, 'w')
	lines = fin.readlines()
	for line in lines[1:]:
	if line.split():
	fout.writelines(line)
	#new_lines = ''.join(lines[1:])
	#fout.write(new_lines)
	fin.close()
	fout.close()
	# 创建XML文档并写入节点信息
	def createXMLFile(attrs, width, height, filename):
	# 创建文档对象, 文档对象用于创建各种节点
	my_dom = xml.dom.getDOMImplementation()
	doc = my_dom.createDocument(None, _ROOT_NODE, None)
	# 获得根节点
	root_node = doc.documentElement
	# folder节点
	createChildNode(doc, 'folder', _FOLDER_NODE, root_node)
	# filename节点
	createChildNode(doc, 'filename', attrs['name'], root_node)
	# source节点
	source_node = doc.createElement('source')
	# source的子节点
	createChildNode(doc, 'database', _DATABASE_NAME, source_node)
	createChildNode(doc, 'annotation', _ANNOTATION, source_node)
	createChildNode(doc, 'image', 'flickr', source_node)
	createChildNode(doc, 'flickrid', 'NULL', source_node)
	root_node.appendChild(source_node)
	# owner节点
	owner_node = doc.createElement('owner')
	# owner的子节点
	createChildNode(doc, 'flickrid', 'NULL', owner_node)
	createChildNode(doc, 'name', _AUTHOR, owner_node)
	root_node.appendChild(owner_node)
	# size节点
	size_node = doc.createElement('size')
	createChildNode(doc, 'width', str(width), size_node)
	createChildNode(doc, 'height', str(height), size_node)
	createChildNode(doc, 'depth', str(_IMAGE_CHANNEL), size_node)
	root_node.appendChild(size_node)
	# segmented节点
	createChildNode(doc, 'segmented', _SEGMENTED, root_node)
	# object节点
	object_node = createObjectNode(doc, attrs)
	root_node.appendChild(object_node)
	# 写入文件
	writeXMLFile(doc, filename)
	if __name__ == "__main__":
	ouput_file = open(_TXT_PATH)
	current_dirpath = os.path.dirname(os.path.abspath('__file__'))
	if not os.path.exists(_ANNOTATION_SAVE_PATH):
	os.mkdir(_ANNOTATION_SAVE_PATH)
	if not os.path.exists(_IMAGE_COPY_PATH):
	os.mkdir(_IMAGE_COPY_PATH)
	lines = ouput_file.readlines()
	for line in lines:
	s = line.rstrip('\n')
	array = s.split()
	print
	attrs = dict()
	attrs['name'] = array[0]
	attrs['classification'] = 'face'
	attrs['xmin'] = array[1]
	attrs['ymin'] = array[2]
	attrs['xmax'] = array[3]
	attrs['ymax'] = array[4]
	# 构建XML文件名称
	xml_file_name = os.path.join(_ANNOTATION_SAVE_PATH, (attrs['name'].split('.'))[0] + '.xml')
	print xml_file_name
	if os.path.exists( xml_file_name):
	# print('do exists')
	existed_doc = xml.dom.minidom.parse(xml_file_name)
	root_node = existed_doc.documentElement
	# 如果XML存在了, 添加object节点信息即可
	object_node = createObjectNode(existed_doc, attrs)
	root_node.appendChild(object_node)
	# 写入文件
	writeXMLFile(existed_doc, xml_file_name)
	else:
	# print('not exists')
	# 如果XML文件不存在, 创建文件并写入节点信息
	img_name = attrs['name']
	img_path = os.path.join(current_dirpath, _IMAGE_PATH, img_name)
	img = Image.open(img_path)
	width, height = img.size
	# 将图片保存在JPEGImages文件夹中一份
	img.save(os.path.join(current_dirpath, _IMAGE_COPY_PATH, img_name), 'jpeg')
	img.close()
	# 创建XML文件
	createXMLFile(attrs, width, height, xml_file_name)

**************************************************create_imagesets.py****************************************************************

# -*- coding:utf-8 -*-
	import os
	__author__ = 'peic'
	_IMAGE_SETS_PATH = 'ImageSets'
	_MAin_PATH = 'ImageSets\\Main'
	_XML_FILE_PATH = 'Annotations'
	if __name__ == '__main__':
	if os.path.exists(_IMAGE_SETS_PATH):
	print('ImageSets dir is already exists')
	if os.path.exists(_MAin_PATH):
	print('Main dir is already in ImageSets')
	else:
	os.mkdir(_IMAGE_SETS_PATH)
	os.mkdir(_MAin_PATH)
	print _MAin_PATH
	# 测试集, 总数据的50%
	f_test = open(os.path.join(_MAin_PATH, 'test.txt'), 'w')
	# 训练和验证集, 除去测试的剩余50%
	f_trainval = open(os.path.join(_MAin_PATH, 'trainval.txt'), 'w')
	# trainval中训练部分, trainval的50%
	f_train = open(os.path.join(_MAin_PATH, 'train.txt'), 'w')
	# trainval中验证集, trainval的50%
	f_val = open(os.path.join(_MAin_PATH, 'val.txt'), 'w')
	# 遍历XML文件夹
	for root, dirs, files in os.walk(_XML_FILE_PATH):
	i = 1
	j = 1
	for file in files:
	if i % 2:
	# 作为测试集
	f_test.writelines(str(file).split('.')[0] + '\n')
	else:
	# 训练和验证集
	f_trainval.writelines(str(file).split('.')[0] + '\n')
	if j % 2:
	# 训练集
	f_train.writelines(str(file).split('.')[0] + '\n')
	else:
	# 验证集
	f_val.writelines(str(file).split('.')[0] + '\n')
	j += 1
	i += 1
	f_test.close()
	f_train.close()
	f_trainval.close()
	f_val.close()

**************************************************rename_images.py***********************************************************

# _*_ coding:utf-8 _*_
	#
	from PIL import Image
	import os
	import re
	import sys
	if __name__ == "__main__":
	if(len(sys.argv) != 2):
	print("args is not 2!")
	sys.exit(1);
	BASE_PATH = sys.argv[1];
	print(BASE_PATH)
	# os.walk返回一个walk对象，是一个三元组
	for rootpath, dirnames, filenames in os.walk(BASE_PATH):
	name_number = 1
	name_length = 6
	for filename in filenames:
	if(re.match("^(\d*).jpg", filename)):
	name_str = (name_length - len(str(name_number))) * '0' + str(name_number) + ".jpg"
	print name_str, '\n'
	name_number = name_number + 1
	image = Image.open(os.path.join(BASE_PATH, filename))
	image.save(os.path.join(BASE_PATH, name_str), 'jpeg')