voc数据集提取某类或某几类的图片和xml

网上大多数都是下面方法:

https://blog.csdn.net/weixin_39881922/article/details/85070808

 但是遇到某些不规则图片名时会报错:

 使用如下代码便不会有任何问题:

import xml.etree.ElementTree as ET
import os
import shutil
 
def copy(xml_path,jpg_path,aim_path,aimxml_path):
    filelist = os.listdir(xml_path)
    # 打开xml文档
    for xmlfile in filelist:
        jpgname = xmlfile.replace("xml","jpg")
        doc = ET.parse(xml_path + xmlfile)
        root = doc.getroot()
        names = [element.find('name').text for element in root.findall('object')]
        if "person" in names:
        #if "person" in names or "dog" in names:
            shutil.copy(xml_path+xmlfile,aimxml_path)
            shutil.copy(jpg_path+jpgname,aim_path)
 
if __name__ == '__main__':
    xml_path = "./Annotations/"
    jpg_path = "./JPEGImages/"
    aim_path = "./JPEGImages1/"
    aimxml_path = "./Annotations1/"
    copy(xml_path,jpg_path,aim_path,aimxml_path)