VOC提取出指定的目标类别xml文件和jpg文件

#!/usr/bin/python
# -*- coding: UTF-8 -*-
# get annotation object bndbox location## 
import shutil, os
import cv2
from pathlib import Path  # 比os.walk 和递归 更牛的遍历文件夹文件方式

try:
    import xml.etree.cElementTree as ET  # 解析xml的c语言版的模块
except ImportError:
    import xml.etree.ElementTree as ET


class ExtractFile:
    def __init__(self, in_xml_path, in_jpg_path):
        self.in_xml_path = in_xml_path
        self.in_jpg_path = in_jpg_path

    def getAnn(self, out_xml_path):  # AnotPath VOC标注文件路径
        for parent, dirnames, filenames in os.walk(self.in_xml_path):
            for filename in filenames:
                filename_qianzhui = filename.split(".")[0]
                xml_file_path = os.path.join(self.in_xml_path, filename_qianzhui) + ".xml"
                copy_file = os.path.join(self.in_xml_path, xml_file_path)
                tree = ET.ElementTree(file=copy_file)  # 打开文件，解析成一棵树型结构
                root = tree.getroot()  # 获取树型结构的根
                ObjectSet = root.findall('object')  # 找到文件中所有含有object关键字的地方，这些地方含有标注目标
                ObjNames = ["person"]  # 以目标类别为关键字，目标框为值组成的字典结构
                for Object in ObjectSet:
                    ObjName = Object.find('name').text
                    if ObjName in ObjNames:
                        self.copy_xml = shutil.copy(xml_file_path, out_xml_path)
                        self.out_xml_toList = []
                        self.out_xml_toList.append(self.copy_xml)
                        return self.out_xml_toList

    print("提取xml完成")

    def getImg(self, out_jpg_path):
        for xml_file_name in extractFile.getAnn(out_xml_path):
            pic_file_name = os.path.split(xml_file_name)[1]
            ##########方法1
            # filename_qianzhui = pic_file_name.split(".")[0]
            # pic_file_path = os.path.join(self.in_jpg_path, filename_qianzhui) + ".jpg"
            ##########方法2
            filename_qianzhui = pic_file_name[:-4] + ".jpg"
            pic_file_path = os.path.join(self.in_jpg_path, filename_qianzhui)
            # 读取存放所有图片路径
            image = cv2.imread(pic_file_path)
            # 提取到的对应图片写入指定路径
            cv2.imwrite(out_jpg_path + filename_qianzhui, image)

    print("根据xml找到对应的jpg，并写入指定路径完成")


if __name__ == '__main__':
    in_xml_path = r"tiqu_voc_xml_input"
    in_jpg_path = r"tiqu_voc_jpg_input"
    extractFile = ExtractFile(in_xml_path, in_jpg_path)

    out_xml_path = r"tiqu_voc_xml_out"
    out_jpg_path = r"tiqu_voc_jpg_out\\" #\\
    extractFile.getImg(out_jpg_path)