【素材处理】统计xml标注文件中指定类别的标注框数量

xml文件解析

导出xml的解析包

import xml.etree.ElementTree as ET

读取xml内容代码

def read_xml(xml_path):
    '''
    统计 xml_path 文件属于列表useful_class中的目标
    '''
    etree = ET.parse(xml_path)
    eroot = etree.getroot()

    for child in eroot.findall('object'):
        name = child.find('name').text                      # 获取当前子节点类别
        if name in useful_class:
            obj_dict[name] += 1

文件结构

--- /home/shares/myproj/dir1
	--- 1.xml
	--- 2.xml
	...
--- /home/shares/myproj/dir2
	--- 1.xml
	--- 2.xml
	...
...

需要统计dir1与dir2下所有xml文件中属于useful_class类别的数量框

完整代码

'''
hqauto: 
Description: 
version: 
Author: YYQQ.
Date: 2023-03-06 14:50:45
LastEditors: Please set LastEditors
LastEditTime: 2023-03-06 15:22:27
'''
import os
import xml.etree.ElementTree as ET

# 需要统计的类别列表
useful_class = ['normalface', 'callface', 'smokeface', 'cigaretteface']

# 创建空字典
obj_dict ={}
for cls in useful_class:
    obj_dict[cls]= 0

def read_xml(xml_path):
    '''
    统计 xml_path 文件属于列表useful_class中的目标
    '''
    etree = ET.parse(xml_path)
    eroot = etree.getroot()

    for child in eroot.findall('object'):
        name = child.find('name').text                      # 获取当前子节点类别
        if name in useful_class:
            obj_dict[name] += 1

def main():
    # 需要统计的目录列表
    xml_dir_list = ['/home/shares/myproj/dir1',
    				'/home/shares/myproj/dir2',
                     ]    # 需要统计的文件路径

    # 遍历目录列表
    for xml_dir in xml_dir_list:
        # 路径不存在则报错
        assert os.path.exists(xml_dir), print('input xml_dir dose not exist, please input again.')
        print(xml_dir + ':')

        xml_files = [x for x in os.listdir(xml_dir) if x.endswith('.xml')]  # 获取所有xml文件
        xml_nums = len(xml_files)          # 当前目录下xml的数量
        xml_index = 1                      # 计数
        for xml_file in xml_files:
            xml_path = os.path.join(xml_dir, xml_file)
            read_xml(xml_path)
            print('\rprocessing xml : [{} / {}]'.format(xml_index, xml_nums), end = "")
            xml_index += 1
        
if __name__ == "__main__":
    main()
    print('\n', obj_dict)