yolov5图像识别voc转yolo代码解析

https://github.com/ultralytics/JSON2YOLO
https://blog.csdn.net/qq_51831335/article/details/127237772
目标检测数据集标签转换COCO2VOC、YOLO2VOC、JSON2YOLO

<annotation>
	<folder>VOC2007</folder>
	<filename>000001.jpg</filename>
	<source>
		<database>The VOC2007 Database</database>
		<annotation>PASCAL VOC2007</annotation>
		<image>flickr</image>
		<flickrid>341012865</flickrid>
	</source>
	<owner>
		<flickrid>Fried Camels</flickrid>
		<name>Jinky the Fruit Bat</name>
	</owner>
	<size>
		<width>353</width>
		<height>500</height>
		<depth>3</depth>
	</size>
	<segmented>0</segmented>
	<object>
		<name>dog</name>
		<pose>Left</pose>
		<truncated>1</truncated>
		<difficult>0</difficult>
		<bndbox>
			<xmin>48</xmin>
			<ymin>240</ymin>
			<xmax>195</xmax>
			<ymax>371</ymax>
		</bndbox>
	</object>
	<object>
		<name>person</name>
		<pose>Left</pose>
		<truncated>1</truncated>
		<difficult>0</difficult>
		<bndbox>
			<xmin>8</xmin>
			<ymin>12</ymin>
			<xmax>352</xmax>
			<ymax>498</ymax>
		</bndbox>
	</object>
</annotation>

#定义一个voc到yolo数据格式的转换代码
#voc的数据标定的时候给定的是：左上角，右下角的坐标（xmin/ymin/xmax/ymax）
#yolo数据标定的时候给定的是：中心点坐标，宽度，高度（x/y/w/h）,并且是和原始图像width和height的百分比

import os
from xml.etree import ElementTree as ET
import tqdm

if __name__ == '__main__':
	label_2_id = {
    
    
	'plate':0
	}
	img_dir= “”
	voc_label_dir = ""
	yolo_label_dir = ""
	#读取所有voc_xml文件名称列表
	if not os.path.exists(yolo_label_dir):
		os.makedirs(yolo_label_dir):
			os.makedirs(yolo_label_dir)
	#遍历处理每个xml文件
	for voc_label_name in tqdm(voc_label_names):
		#1.解析xml文件
		tree = ET.parse(os.path.join(voc_label_dir,voc_label_name))
		#2.获取根节点的信息
		root = tree.getroot()
		#3.获取文件名称
		filename = root.find('filename').text
		#4.判断图像文件是是否存在
		img_file = os.path.join(img_dir,filename)
		if not os.path.exists(img_file)
			continue
		#5.将标注的box信息转换成yolo格式，然后保存到txt文件中
		#6.提取图像的大小
		size_obj = root.find('size')
		width = float(size_obj.find('width').text)
		height = float(size_obj.find('height').text)
		yolo_label_file = os.path.join(yolo_label_dir,f"{
      
      os.path.splitext(filename)[0]}.txt")
		with open(yolo_label_file,'w',encoding = 'UTF-8') as writer:
			for obj in root.findall('object'):
			  #提取标签名称
				label = obj.find('name').text
				#提取左上角，右下角坐标
				bbox = obj.find('bndbox')
				xmin = float(bbox.find('xmin').text)
				xmax = float(bbox.find('xmin').text)
				ymin = float(bbox.find('xmin').text)
				ymax = float(bbox.find('xmin').text)
				#转换成中心点坐标，宽度，高度
				w = xmax-xmin
				h = ymax -ymin
				x = xmin + w/2.0
				y = ymin + h/2.0
				#需要相对宽度和高度
				w = w/width
				h = h/height
				x = x/width
				y = y/height
				#输出
				writer.writelines(f"{
      
      label_2_id{
      
      label} {
      
      x} {
      
      y} {
      
      w} {
      
      h}\n}")
		#img_name = voc_label_name[:-3]
		#print(voc_label_names)