-
Notifications
You must be signed in to change notification settings - Fork 0
/
2xml.py
124 lines (109 loc) · 5.63 KB
/
2xml.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
import os
import shutil
import json
from jsonpath import jsonpath
import copy
from lxml.etree import Element, SubElement, tostring, ElementTree
import cv2
import numpy as np
import threading
# BASE_PATH = r"E:/dataset/prj-m9p9jphvxtkgfghe"
template_file = './template.xml' # xml模板
target_dir = 'E:/dataset-jsxs/Annotations/' #新标注文件保存路径
paired_img = 'E:/dataset-jsxs/JPEGImages/' #与xml配对图像保存路径
def splitJSON(rootpath):
with open(rootpath + '/data0.json', 'r', encoding='utf8') as f:
for line in f.readlines():
trainfiles = json.loads(line)
tree = ElementTree()
for k in enumerate(trainfiles):
annos = trainfiles['annotations']
file_name = trainfiles['imageName']
arr = np.empty([0, 5], int)
for anno in annos:
loc = anno['location']
label = anno['labelName']
if label != 'jsxs':
continue
else:
arr = np.append(arr, [[loc['top'], loc['height'], loc['left'], loc['width'], label]], axis=0)
for i in range(arr.shape[0]):
if i == 0:
label = arr[i, 4] # 标签名
# 坐标
ymin = arr[i, 0] # ymin = top
ymax = str(int(ymin) + int(arr[i, 1])) # ymax = top + height
xmin = arr[i, 2] # xmin = left
xmax = str(int(xmin) + int(arr[i, 3]))
tree.parse(template_file) # 解析树
root = tree.getroot() # 根节点
root.find('filename').text = file_name
# size
sz = root.find('size')
sz.find('height').text = str(anno['location']['height'])
sz.find('width').text = str(anno['location']['width'])
try:
sz.find('depth').text = str(anno['location']['depth'])
except KeyError:
# print(file_name + '无法找到depth标签，直接置为3')
sz.find('depth').text = str(3)
# object
obj = root.find('object')
obj.find('name').text = label
bb = obj.find('bndbox')
bb.find('xmin').text = xmin
bb.find('ymin').text = ymin
bb.find('xmax').text = xmax
bb.find('ymax').text = ymax
# 有多个标签需要添加object
else:
label = arr[i, 4] # 标签名
# 坐标
ymin = arr[i, 0] # ymin = top
ymax = str(int(ymin) + int(arr[i, 1])) # ymax = top + height
xmin = arr[i, 2] # xmin = left
xmax = str(int(xmin) + int(arr[i, 3]))
obj_ori = root.find('object')
obj = copy.deepcopy(obj_ori) # 注意这里深拷贝
obj.find('name').text = label
bb = obj.find('bndbox')
bb.find('xmin').text = xmin
bb.find('ymin').text = ymin
bb.find('xmax').text = xmax
bb.find('ymax').text = ymax
root.append(obj)
xml_file = file_name.split('.')[0] + ".xml"
tree.write(target_dir + xml_file, encoding='utf-8')
shutil.copyfile(rootpath + '/../data/0/' + file_name, paired_img + file_name)
print('保存文件：'+ trainfiles['imageName'].split('.')[0] + ".xml")
def walkdir(BASE_PATH):
for root, dirs, files in os.walk(BASE_PATH, topdown=False):
print('正在执行 ' + root)
if len(files) == 2 and files[0] == 'data0.json':
data = json.load(open(root + '/labels.json', 'r', encoding='utf8'))
label = []
for dic in data:
label.append(dic['labelName'])
print(label)
if 'jsxs' in label or 'jsxs\ ' in label:
print('找到含锈蚀数据:' + root)
splitJSON(root)
else:
continue
else:
continue
# 尝试多线程
th1 = threading.Thread(target=walkdir, args=(r'E:/dataset/prj-mfem9jnj2rhbzrea',)) ##创建线程
th1.start() ##启动线程
th2 = threading.Thread(target=walkdir, args=(r'E:/dataset/prj-p1yxm4766e15utiw',))
th2.start()
th3 = threading.Thread(target=walkdir, args=(r'E:/dataset/prj-p2tj56uc62termgk',))
th3.start()
th4 = threading.Thread(target=walkdir, args=(r'E:/dataset/prj-s88ph50hieufwwzb',))
th4.start()
th5 = threading.Thread(target=walkdir, args=(r'E:/dataset/prj-vej3rcx7e3m72v6u',))
th5.start()
th6 = threading.Thread(target=walkdir, args=(r'E:/dataset/prj-ygaj3h3km66zfwdh',))
th6.start()
th7 = threading.Thread(target=walkdir, args=(r'E:/dataset/prj-yjjwpj31avp9m8q4',))
th7.start()