Yolo format data set processing (XML to txt)

---

Preface

YOLO The data set of the network is txt Text , When we want to train some models , The data found on the Internet are xml Format , At this time, we need to process the data , Get the data format we want .

---

One 、 Data processing flow

1. Read xml file , analysis xml Get the width of the picture , high , Coordinate information of calibration frame
2. Data normalization
3. write in txt file

Two 、xml File data format

The picture above , I intercepted it xml Some data of the file , We just need to get size Medium width,height and bndbox Coordinate information in .

3、 ... and 、 Code

import os
import glob
import xml.etree.ElementTree as ET

xml_file=r'E:\ desktop \ Information \cv4\ Data sets \voc Data sets \Annotations'

l=['aeroplane', 'bicycle', 'bird', 'boat', 'bottle', 'bus', 'car', 'cat', 'chair', 'cow', 'diningtable', 'dog',
        'horse', 'motorbike', 'person', 'pottedplant', 'sheep', 'sofa', 'train', 'tvmonitor']

def convert(box,dw,dh):
    x=(box[0]+box[2])/2.0
    y=(box[1]+box[3])/2.0
    w=box[2]-box[0]
    h=box[3]-box[1]

    x=x/dw
    y=y/dh
    w=w/dw
    h=h/dh

    return x,y,w,h

def f(name_id):
    xml_o=open(r'E:\ desktop \ Information \cv4\ Data sets \voc Data sets \Annotations\%s.xml'%name_id)
    txt_o=open(r'E:\ desktop \ Information \cv4\ Data sets \voc Data sets \labels1\%s.txt'%name_id,'w')

    pares=ET.parse(xml_o)
    root=pares.getroot()
    objects=root.findall('object')
    size=root.find('size')
    dw=int(size.find('width').text)
    dh=int(size.find('height').text)

    for obj in objects :
        c=l.index(obj.find('name').text)
        bnd=obj.find('bndbox')

        b=(float(bnd.find('xmin').text),float(bnd.find('ymin').text),
           float(bnd.find('xmax').text),float(bnd.find('ymax').text))

        x,y,w,h=convert(b,dw,dh)

        write_t="{} {:.5f} {:.5f} {:.5f} {:.5f}\n".format(c,x,y,w,h)
        txt_o.write(write_t)

    xml_o.close()
    txt_o.close()

name=glob.glob(os.path.join(xml_file,"*.xml"))
for i in name :
    name_id=os.path.basename(i)[:-4]
    f(name_id)

summary

That's all xml turn txt The whole content of the document , What problems occur during use , Leave a comment in the comments section .