pytorch实战02-处理帧和将xml转为txt文件

1、简介

前一节是对HMDB数据视频处理为帧

此章节：将使用labelIM软件对图片进行标注生成的xml文件转成对应的txt文件

2、代码解讲

  xml转为txt文件：

在"E:/dataSet/meng_data/DATA/xml"文件夹下皆为xml文件,此处改为自己的文件夹即可

import os
import numpy as np
from xml.etree.ElementTree import ElementTree


# 将labelIMG标注的图片转换为txt文件
class Xml2txt:
    def __init__(self, path):
        self.path = path

    def convert2Txt(self):
        file_path = self.path
        tree = ElementTree()
        tree.parse(file_path)
        root = tree.getroot()
        for obj in root.iter('object'):
            xmlbox = obj.find('bndbox')
            # print(xmlbox[0])
            b = [int(xmlbox.find('xmin').text), int(xmlbox.find('ymin').text), int(xmlbox.find('xmax').text),
                 int(xmlbox.find('ymax').text)]
        return b


if __name__ == "__main__":

    folder = "E:/dataSet/meng_data/DATA/xml"
    # 遍历文件夹中的所有xml文件
    for root, dirs, files in os.walk(folder):
        for f in files:
            file_name, _ = f.split('.')
            file_path = os.path.join(folder, f)
            # print(file_path)
            xml = Xml2txt(file_path)
            box = xml.convert2Txt()
            # 此处可以自己设置加一些属性，例如类别之类的
            box.insert(0, 1)
            box = [box]
            np.savetxt(folder + "/" + "{}.txt".format(file_name), box, fmt="%d", delimiter=" ")

 结果如下：保存的是[class， xmin，ymin,  xmax,  ymax]

3、完整代码

import os
import numpy as np
import shutil
from xml.etree.ElementTree import ElementTree


# 将labelIMG标注的图片转换为txt文件
class Xml2txt:
    def __init__(self, path):
        self.path = path

    def convert2Txt(self):
        file_path = self.path
        tree = ElementTree()
        tree.parse(file_path)
        root = tree.getroot()
        for obj in root.iter('object'):
            xmlbox = obj.find('bndbox')
            # print(xmlbox[0])
            b = [int(xmlbox.find('xmin').text), int(xmlbox.find('xmax').text), int(xmlbox.find('ymin').text),
                 int(xmlbox.find('ymax').text)]
        return b


if __name__ == "__main__":

    # folder = "E:/dataSet/meng_data/DATA/xml"
    # # 遍历文件夹中的所有xml文件
    # for root, dirs, files in os.walk(folder):
    #     for f in files:
    #         file_name, _ = f.split('.')
    #         file_path = os.path.join(folder, f)
    #         # print(file_path)
    #         xml = Xml2txt(file_path)
    #         box = xml.convert2Txt()
    #         box.insert(0, 1)
    #         box = [box]
    #         np.savetxt(folder + "/" + "{}.txt".format(file_name), box, fmt="%d", delimiter=" ")

    # 定义原始图片和XML文件的路径
    origin_root_path = "E:/dataSet/meng_data/YOWO-DATA"
    origin_img_path = "train_imgs"
    origin_xml_path = "train_labels"

    # 定义图片和标签将要保存的根路径
    dst_root_path = "E:/dataSet/meng_data/DATA"
    # 定义图片保存的txt文件路径
    dst_img_path = "imgs"
    # 定义标签保存的路径
    dst_label_path = "labels"
    # 定义标签索引保存的路径
    train_list = ""
    test_list = ""

    label_name = os.listdir(os.path.join(origin_root_path, origin_img_path))
    # print(label_name)
    index = 1
    for i in label_name:
        if i.startswith('.'):
            continue
        sub_folders = os.path.join(origin_root_path, origin_img_path, i)

        # 创建动作类型文件
        dst_sub_img = os.path.join(dst_root_path, dst_img_path, i)
        dst_sub_label = os.path.join(dst_root_path, dst_label_path, i)
        if not os.path.exists(dst_sub_img):
            os.mkdir(dst_sub_img)
        if not os.path.exists(dst_sub_label):
            os.mkdir(dst_sub_label)


        # print(sub_folder)
        list_sub_folder = os.listdir(sub_folders)
        # print(list_sub_folder)
        for sub_folder in list_sub_folder:
            folder = os.path.join(sub_folders, sub_folder)

            # 创建目标动作类型子文件
            dst_sub_img_folder = os.path.join(dst_sub_img, sub_folder)
            dst_sub_label_folder = os.path.join(dst_sub_label, sub_folder)
            if not os.path.exists(dst_sub_img_folder):
                os.mkdir(dst_sub_img_folder)
            if not os.path.exists(dst_sub_label_folder):
                os.mkdir(dst_sub_label_folder)

            # print(folder)
            # 遍历文件夹中的所有文件
            for root, dirs, files in os.walk(folder):
                for f in files:
                    file_path = os.path.join(folder, f)
                    if f.endswith('.jpg'):
                        # cv2.imwrite(dst_sub_img_folder + "/" + f, f)
                        shutil.copyfile(file_path, dst_sub_img_folder+"/"+f)
                    if f.endswith('.xml'):
                        file_name, _ = f.split('.')
                        xml = Xml2txt(file_path)
                        box = xml.convert2Txt()
                        box.insert(0, index)
                        box = [box]
                        np.savetxt(dst_sub_label_folder + "/" + "{}.txt".format(file_name), box, fmt="%d", delimiter=" ")

        index += 1

完成之后的样子：全部生成txt格式的文件