python txt 和 xml 格式互转

背景：

图片检测中标注文件 txt 格式 和 xml 格式 互相转化

label.txt

Car 1701 915 1920 1039
Car 625 765 808 839
Car 1827 783 1919 859
Cyclist 1252 767 1317 835
Cyclist 1395 816 1487 894
Car 24 636 135 683
Car 1027 673 1075 717
Car 900 685 950 730
Car 912 651 949 684

label.xml



xml --> txt

import xml.etree.ElementTree as ET
import os


# 数据标签
classes = ['Pedestrian','Cyclist','Car','Bus','Tram','Truck','Dump_Truck','Cement_truck','Fule_Tank','Trailer','Misc','TrafficLight_Dig','TrafficLight_Black','TrafficLight_Yellow','yup','yright','ycircle','TrafficLight_Red','rup','rright','rleft','rperson','rcircle','TrafficLight_Green','gup','gdown','gright','gleft','gcircle','gperson','gbike']



def convert_annotation(xmlpath,txtpath,xmlname):
    name = xmlname.split('.')[0]

    if not os.path.exists(txtpath):
        os.makedirs(txtpath)
    txtfile = os.path.join(txtpath, name +".txt")

    xmlfile = os.path.join(xmlpath,xmlname)
    with open(xmlfile, "r", encoding='UTF-8') as in_file:
        with open(txtfile, "w+" ,encoding='UTF-8') as out_file:
            tree=ET.parse(in_file)

            root = tree.getroot()
            size = root.find('size')
            w = int(size.find('width').text)
            h = int(size.find('height').text)
            d = int(size.find('depth').text)
            out_file.truncate()
            for obj in root.find('outputs').find('object'):
                cls = obj.find('name').text
                if cls not in classes:
                    continue
                cls_id = classes.index(cls)
                xmlbox = obj.find('bndbox')
                b = (int(xmlbox.find('xmin').text), int(xmlbox.find('ymin').text), int(xmlbox.find('xmax').text),
                     int(xmlbox.find('ymax').text))


                out_file.write(cls + " " + " ".join([str(a) for a in b]) + '\n')
        print(txtfile + " was written ！！")


if __name__ == "__main__":

    rootpath = '/media/wxf/Elements/data/video'
    xmlpath = rootpath + os.sep + 'label_xml'
    txtpath = rootpath + os.sep + 'label_xml2txt'

    xml_list = os.listdir(xmlpath)

    for i in range(0, len(xml_list)):
        path = os.path.join(xmlpath, xml_list[i])
        if ('.xml' in path) or ('.XML' in path):
            convert_annotation(xmlpath, txtpath, xml_list[i])
            print('done', i)
        else:
            print('not xml file', i)

txt --> xml

import os
import numpy as np
import xml.etree.ElementTree as ET
from xml.etree.ElementTree import Element, SubElement, tostring
from xml.dom.minidom import parseString

import cv2
import time


def MakeTxt2Xml(txt_path,xml_path,png_path):
    data = np.loadtxt(txt_path,dtype=np.str_)
    if data.size == 0:
        output = "Begin process, " + txt_path + " is null!!!"
        print(output)
    if data.ndim == 1:
        data = np.array([data])
    doc_root = ET.Element('doc')
    tree = ET.ElementTree(doc_root)

    img_path_element = ET.Element('path')
    img_path_element.text = png_path
    doc_root.append(img_path_element)

    outputs_element = ET.Element('outputs')
    object_element = SubElement(outputs_element,'object')
    # element.append(SubElement) 与  subelement = SubElement(main_element,'subelement') 是相同的意思
    for i in range(data.shape[0]):

        item_element = SubElement(object_element, 'item')


        name_element = SubElement(item_element, 'name')
        name_element.text = data[i][0]

        bndbox_element = SubElement(item_element, 'bndbox')

        xmin_element = SubElement(bndbox_element, 'xmin')
        xmin_element.text = str(data[i][1])

        ymin_element = SubElement(bndbox_element, 'ymin')
        ymin_element.text = str(data[i][2])

        xmax_element = SubElement(bndbox_element, 'xmax')
        xmax_element.text = str(data[i][3])

        ymax_element = SubElement(bndbox_element, 'ymax')
        ymax_element.text = str(data[i][4])


    doc_root.append(outputs_element)

    time_element = ET.Element('time_labeled')
    # time_element.text = time.strftime("%Y-%m-%d_%H:%M:%S", time.localtime())
    time_element.text = str(int(round(time.time() * 1000)))
    doc_root.append(time_element)

    labeled_element = ET.Element('labeled')
    if (os.path.exists(txt_path)):
        labeled_element.text = str("true")
    else:
        labeled_element.text = str("false")
    doc_root.append(labeled_element)

    size_element = ET.Element('size')
    width_element = SubElement(size_element, 'width')
    img = cv2.imread(png_path)
    # width_element.text = str(img.shape[1])
    # height_element = SubElement(size_element, 'height')
    # height_element.text = str(img.shape[0])
    # depth_element = SubElement(size_element, 'depth')
    # depth_element.text = str(img.shape[2])
    width_element.text = str(1920)
    height_element = SubElement(size_element, 'height')
    height_element.text = str(1080)
    depth_element = SubElement(size_element, 'depth')
    depth_element.text = str(3)

    doc_root.append(size_element)

    xml = tostring(doc_root)
    dom = parseString(xml)

    # xml_name = pic_name.replace(".jpg", "")
    # xml_name = os.path.join(save_xml_path, xml_name + '.xml')
    with open(xml_path, 'wb') as f:
        f.write(dom.toprettyxml(indent='\t', encoding='utf-8'))
    print(xml_path+" is writed!")



if __name__ == "__main__":
    labeltxt = "/media/wxf/Elements/data/video/label_txt"
    labelxml = "/media/wxf/Elements/data/video/label_xml"
    img_path = "."


    for txt in sorted(os.listdir(labeltxt)):
        txt_path = os.path.join(labeltxt,txt)
        name = txt.split(".")[-2]
        xml_path = labelxml + os.sep + name + ".xml"
        png_path = img_path + os.sep + name + ".png"
        if os.path.exists(txt_path):
            # print(txt_path)
            MakeTxt2Xml(txt_path,xml_path,png_path)