【脚本】统计指定目录下所有xml的name（即统计标注的类别）

2022.8.18记录：

目前只标注了87张，经过下面脚本统计共有7类：

[‘Bolt_hole’, ‘Falling_block’, ‘Pipe’, ‘Cable’, ‘Support’, ‘Segment_joint’, ‘Grouting_hole’]

import os
import xml.etree.ElementTree as ET

def get_xml_class(rootdir):
    filelist = os.listdir(rootdir)
    labelname = []
    for file in filelist:
        filepath = os.path.join(rootdir, file)
        tree = ET.parse(filepath)
        root = tree.getroot()

        for obj in root.iter('object'):
            name = obj.find('name').text
            if name not in labelname:
                labelname.append(name)
    return labelname, len(labelname)

if __name__ == '__main__':
    labelname, len = get_xml_class(r'D:\A_dataset\liao_Upan\orig_clip45_2\labels_xml')
    print(labelname)
    print(len)