一、正矩形框
1、xml的第一种格式为:
<annotation>
<source>
<filename>1501.tif</filename>
<origin>SAR</origin>
</source>
<research>
<version>1.0</version>
<author>team name</author>
<pluginclass>Detection</pluginclass>
<time>2021-07-2021-11</time>
</research>
<size>
<width>1024</width>
<height>1024</height>
</size>
<objects>
<object>
<coordinate>pixel</coordinate>
<type>rectangle</type>
<description>None</description>
<possibleresult>
<name>ARJ21</name>
<probability>1</probability>
</possibleresult>
<points>
<point>396.000000,719.000000</point>
<point>447.000000,719.000000</point>
<point>447.000000,765.000000</point>
<point>396.000000,765.000000</point>
<point>396.000000,719.000000</point>
</points>
</object>
</objects>
</annotation>
对应的python代码如下:
import xml.etree.ElementTree as ET
import os
class_names = ['A220', 'A330', 'A320/321', 'Boeing737-800', 'Boeing787', 'ARJ21', 'other']
xmlpath = 'E:/plane/sar/xmllabels/' # 原xml路径
txtpath = 'E:/plane/sar/txtlabels/' # 转换后txt文件存放路径
if not os.path.exists(txtpath):
os.makedirs(txtpath)
files = []
for root, dirs, files in os.walk(xmlpath):
None
number = len(files)
print("文件数为", number)
i = 0
while i < number:
name = files[i][0:-4]
xml_name = name + ".xml"
txt_name = name + ".txt"
xml_file_name = xmlpath + xml_name
txt_file_name = txtpath + txt_name
xml_file = open(xml_file_name)
tree = ET.parse(xml_file)
root = tree.getroot()
source = root.find("source")
filename = source.find('filename').text
# 图片名称
image_name = source.find('filename').text
w = int(root.find('size').find('width').text)
h = int(root.find('size').find('height').text)
f_txt = open(txt_file_name, 'w+')
content = ""
first = True
for obj in root.find("objects").iter('object'):
# 分类名称
name = obj.find("possibleresult").find('name').text
# 分类编号
class_num = class_names.index(name)
points = obj.find('points')
# 获取第一个和第三个point元素的值
first_point = points.findall('point')[0].text
third_point = points.findall('point')[2].text
# 将点的值解析为浮点数列表
first_point_values = list(map(float, first_point.split(',')))
third_point_values = list(map(float, third_point.split(',')))
# 打印结果
print(int(first_point_values[0]), first_point_values[1], third_point_values[0], third_point_values[1])
# 左上角点x坐标
x1 = int(first_point_values[0])
# 右下角点x坐标
x2 = int(third_point_values[0])
# 左上角点y坐标
y1 = int(first_point_values[1])
# 右下角点y坐标
y2 = int(third_point_values[1])
if first:
content += str(class_num) + " " + \
str((x1 + x2) / 2 / w) + " " + str((y1 + y2) / 2 / h) + " " + \
str((x2 - x1) / w) + " " + str((y2 - y1) / h)
first = False
else:
content += "\n" + \
str(class_num) + " " + \
str((x1 + x2) / 2 / w) + " " + str((y1 + y2) / 2 / h) + " " + \
str((x2 - x1) / w) + " " + str((y2 - y1) / h)
# print(str(i / (number - 1) * 100) + "%\n")
print(content)
f_txt.write(content)
f_txt.close()
xml_file.close()
i += 1
print("done!")
2、xml的第二种格式为:
<annotation>
<folder>PCB</folder>
<filename>l_light_01_missing_hole_01_1_600</filename>
<source>
<database>My Database</database>
<annotation>PCB</annotation>
<image>flickr</image>
<flickrid>NULL</flickrid>
</source>
<owner>
<flickrid>NULL</flickrid>
<name>idaneel</name>
</owner>
<size>
<width>600</width>
<height>600</height>
<depth>3</depth>
</size>
<segmented>0</segmented>
<object>
<name>missing_hole</name>
<pose>Unspecified</pose>
<truncated>0</truncated>
<difficult>0</difficult>
<bndbox>
<xmin>240</xmin>
<ymin>404</ymin>
<xmax>282</xmax>
<ymax>437</ymax>
</bndbox>
</object>
</annotation>
对应的python代码为:
import os.path
import xml.etree.ElementTree as ET
import os
import random
# class_names = ['palm', 'stone', 'scissor', 'awesome', 'heartB', 'OK', 'ROCK', 'one', 'swear', 'thanks', 'heartA',
# 'heartC', 'good', 'bad', 'pray', 'call', 'take_picture', 'salute']
class_names = ['menopause', 'hairball', 'broken yarn', 'hole','stains']
xmlpath = 'F:/Project_code/yolov7-main/VOCdevkit/VOC2007/Annotations/' # 原xml路径
txtpath = 'F:/Project_code/yolov7-main/VOCdevkit/VOC2007/labels_copy/' # 转换后txt文件存放路径
if not os.path.exists(txtpath):
os.makedirs(txtpath)
files = []
for root, dirs, files in os.walk(xmlpath):
None
number = len(files)
print(number)
i = 0
while i < number:
name = files[i][0:-4]
xml_name = name + ".xml"
txt_name = name + ".txt"
xml_file_name = xmlpath + xml_name
txt_file_name = txtpath + txt_name
xml_file = open(xml_file_name)
tree = ET.parse(xml_file)
root = tree.getroot()
filename = root.find('filename').text
image_name = root.find('filename').text
w = int(root.find('size').find('width').text)
h = int(root.find('size').find('height').text)
f_txt = open(txt_file_name, 'w+')
content = ""
first = True
for obj in root.iter('object'):
name = obj.find('name').text
class_num = class_names.index(name)
xmlbox = obj.find('bndbox')
x1 = int(xmlbox.find('xmin').text)
x2 = int(xmlbox.find('xmax').text)
y1 = int(xmlbox.find('ymin').text)
y2 = int(xmlbox.find('ymax').text)
if first:
content += str(class_num) + " " + \
str((x1 + x2) / 2 / w) + " " + str((y1 + y2) / 2 / h) + " " + \
str((x2 - x1) / w) + " " + str((y2 - y1) / h)
first = False
else:
content += "\n" + \
str(class_num) + " " + \
str((x1 + x2) / 2 / w) + " " + str((y1 + y2) / 2 / h) + " " + \
str((x2 - x1) / w) + " " + str((y2 - y1) / h)
# print(str(i / (number - 1) * 100) + "%\n")
print(content)
f_txt.write(content)
f_txt.close()
xml_file.close()
i += 1
print("done!")
二、斜矩形框
<object>
<coordinate>pixel</coordinate>
<type>rectangle</type>
<description>None</description>
<possibleresult>
<name>A220</name>
</possibleresult>
<points>
<point>316.000000,544.000000</point>
<point>281.000000,575.000000</point>
<point>250.000000,540.000000</point>
<point>285.000000,509.000000</point>
<point>316.000000,544.000000</point>
</points>
</object>
1、先将xml格式转化为dota数据集格式
import os
import xml.etree.ElementTree as ET
# 定义一个函数来处理单个 XML 文件
def convert_xml_to_txt(xml_file, txt_file):
# 解析 XML 文件
tree = ET.parse(xml_file)
root = tree.getroot()
# 打开一个 txt 文件用于存储结果
with open(txt_file, 'w') as f:
# 遍历 XML 文件中的每个对象
for obj in root.findall('.//object'):
# 获取对象的名称
name = obj.find('possibleresult/name').text
# 获取对象的前四个点坐标,并将它们转换为整数
points = obj.find('points')
point_values = []
for i in range(4):
point = points.findall('point')[i].text.split(',')
point_values.extend([int(float(p)) for p in point])
# 将整数坐标值、名称和0写入 txt 文件
f.write(f'{" ".join(map(str, point_values))} {name} 0\n')
# 定义要处理的文件夹路径和目标文件夹路径
source_folder_path = './xml_labels' # 替换为你的源文件夹路径
target_folder_path = './txt_labels' # 替换为你的目标文件夹路径
# 遍历源文件夹中的所有 XML 文件,并将每个文件转换为 TXT 文件保存到目标文件夹中
for filename in os.listdir(source_folder_path):
if filename.endswith('.xml'):
# 获取 XML 文件的完整路径
xml_file = os.path.join(source_folder_path, filename)
# 获取目标 TXT 文件的完整路径,并确保文件名一致
txt_file = os.path.join(target_folder_path, filename.replace('.xml', '.txt'))
# 调用函数将 XML 文件转换为 TXT 文件
convert_xml_to_txt(xml_file, txt_file)
2、再通过以下文章中代码修改
全网首发!Yolov8_obb旋转框训练、测试、推理手把手教学(DOTA1.0数据集map50已达80%)