train_road相关脚本文件

最新推荐文章于 2024-09-17 23:15:58 发布

unbekannten

最新推荐文章于 2024-09-17 23:15:58 发布

阅读量132

点赞数

文章标签： python 计算机视觉 opencv

本文链接：https://blog.csdn.net/unbekannten/article/details/131988051

版权

这些Python脚本主要用于视频处理，包括将MP4视频转换为特定分辨率和帧率的MJPG编码格式，以及将XML标注文件转换为VOC格式的数据集，用于机器学习训练。此外，还有一个脚本用于将XML文件可视化到图像上，另一个脚本用于分割训练集和测试集。

摘要由CSDN通过智能技术生成

convert_mp4_size.py


#功能：把/home/swjtu/zhouyanfei目录下的所有mp4格式的视频，统一转换成1024*720，15帧，MJPG编码格式的视频
import cv2
import os
dirs = os.listdir('/home/swjtu/yanfei')
for dir in dirs:
    if dir.split('.')[1] == "mp4":
        cap = cv2.VideoCapture(dir)
        videowriter = cv2.VideoWriter("c"+dir.split('.')[0]+".mp4", cv2.VideoWriter_fourcc('M', 'J', 'P', 'G'), 15, (1920,1080))

        success, _ = cap.read()

        while success:
            success, img1 = cap.read()
            try:
                img = cv2.resize(img1, (1920, 1080), interpolation=cv2.INTER_LINEAR)
                videowriter.write(img)
            except:
                break

2.convertr_xml2.py

# id_list=[1,63,131 ,196 ,261 ,322 ,388 ,448 ,511 ,574 ,637 ,701 ,770 ,847 ,908 ,973 ,1035,1096,1158,1219,1280,1342,1405,1469,1530,1592,1700,1772,1835,1923,1987,2050,2110,2170,2230,2290,2350,2414,2475,2535,2600,2660,2720,2784,2844,2904,2945,3005,3005,3085,3145,3205,3223,3287,3347,3409,3485,3545,3614,3674,3734,3800,3862,3922,3992,4091,4173,4257,4326,4396,4496,4563,4637,4700,4760,4828,4891,4958,5032,5095,5158,5229]
# path_list=['/home/vagrant/vagrant_data/departing/day/IMG_7872_batch.mov','/home/vagrant/vagrant_data/departing/day/IMG_7839_batch.mov','/home/vagrant/vagrant_data/departing/day/IMG_7836_batch.mov','/home/vagrant/vagrant_data/departing/day/IMG_7871_batch.mov','/home/vagrant/vagrant_data/departing/day/IMG_7874_batch.mov','/home/vagrant/vagrant_data/departing/day/IMG_7838_batch.mov','/home/vagrant/vagrant_data/departing/day/IMG_7834_batch.mov','/home/vagrant/vagrant_data/departing/day/IMG_7870_batch.mov','/home/vagrant/vagrant_data/departing/day/IMG_7835_batch.mov','/home/vagrant/vagrant_data/departing/day/IMG_7873_batch.mov','/home/vagrant/vagrant_data/departing/night/IMG_7966_batch.mov','/home/vagrant/vagrant_data/departing/night/IMG_7964_batch.mov','/home/vagrant/vagrant_data/departing/night/IMG_7965_batch.mov','/home/vagrant/vagrant_data/departing/night/IMG_7962_batch.mov','/home/vagrant/vagrant_data/departing/night/IMG_7963_batch.mov','/home/vagrant/vagrant_data/oncoming/day/IMG_7832_batch.mov','/home/vagrant/vagrant_data/oncoming/day/IMG_7878_batch.mov','/home/vagrant/vagrant_data/oncoming/day/IMG_7829_batch.mov','/home/vagrant/vagrant_data/oncoming/day/IMG_7875_batch.mov','/home/vagrant/vagrant_data/oncoming/day/IMG_7877_batch.mov','/home/vagrant/vagrant_data/oncoming/day/IMG_7833_batch.mov','/home/vagrant/vagrant_data/oncoming/day/IMG_7831_batch.mov','/home/vagrant/vagrant_data/oncoming/day/IMG_7879_batch.mov','/home/vagrant/vagrant_data/oncoming/day/IMG_7876_batch.mov','/home/vagrant/vagrant_data/oncoming/day/IMG_7830_batch.mov','/home/vagrant/vagrant_data/oncoming/night/IMG_7961_batch.mov','/home/vagrant/vagrant_data/oncoming/night/IMG_7957_batch.mov','/home/vagrant/vagrant_data/oncoming/night/IMG_7958_batch.mov','/home/vagrant/vagrant_data/oncoming/night/IMG_7959_batch.mov','/home/vagrant/vagrant_data/oncoming/night/IMG_7960_batch.mov','/home/vagrant/vagrant_data/departing/day/VID_20210519_111820.mp4','/home/vagrant/vagrant_data/departing/day/VID_20210519_105630.mp4','/home/vagrant/vagrant_data/departing/day/VID_20210518_152255.mp4','/home/vagrant/vagrant_data/departing/day/VID_20210518_144605.mp4','/home/vagrant/vagrant_data/departing/day/VID_20210519_104611.mp4','/home/vagrant/vagrant_data/departing/day/VID_20210518_143528.mp4','/home/vagrant/vagrant_data/departing/day/VID_20210519_113018.mp4','/home/vagrant/vagrant_data/departing/day/VID_20210519_110729.mp4','/home/vagrant/vagrant_data/departing/day/VID_20210518_151004.mp4','/home/vagrant/vagrant_data/departing/night/video_20210519_220754.mp4','/home/vagrant/vagrant_data/departing/night/VID_20210520_211700.mp4','/home/vagrant/vagrant_data/departing/night/VID_20210520_210042.mp4','/home/vagrant/vagrant_data/departing/night/VID_20210519_224219.mp4','/home/vagrant/vagrant_data/departing/night/VID_20210517_204930.mp4','/home/vagrant/vagrant_data/departing/night/VID_20210520_203226.mp4','/home/vagrant/vagrant_data/departing/night/video_20210519_221915/','/home/vagrant/vagrant_data/departing/night/VID_20210520_210145.mp4','/home/vagrant/vagrant_data/departing/night/video_20210519_215357.mp4','/home/vagrant/vagrant_data/departing/night/video_20210519_215357.mp4','/home/vagrant/vagrant_data/departing/night/VID_20210517_210311.mp4','/home/vagrant/vagrant_data/departing/night/VID_20210517_203714.mp4','/home/vagrant/vagrant_data/departing/night/VID_20210519_223115.mp4','/home/vagrant/vagrant_data/departing/night/VID_20210520_205845.mp4','/home/vagrant/vagrant_data/departing/night/VID_20210517_211524.mp4','/home/vagrant/vagrant_data/oncoming/day/VID_20210519_175034.mp4','/home/vagrant/vagrant_data/oncoming/day/VID_20210518_150300.mp4','/home/vagrant/vagrant_data/oncoming/day/VID_20210519_094855.mp4','/home/vagrant/vagrant_data/oncoming/day/VID_20210519_171122.mp4','/home/vagrant/vagrant_data/oncoming/day/VID_20210519_101011.mp4','/home/vagrant/vagrant_data/oncoming/day/VID_20210519_095939.mp4','/home/vagrant/vagrant_data/oncoming/day/VID_20210518_151436.mp4','/home/vagrant/vagrant_data/oncoming/day/VID_20210519_102038.mp4','/home/vagrant/vagrant_data/oncoming/day/VID_20210519_093508.mp4','/home/vagrant/vagrant_data/oncoming/day/video_20210518_152924.mp4','/home/vagrant/vagrant_data/oncoming/day/VID_20210518_145013.mp4','/home/vagrant/vagrant_data/oncoming/day/VID_20210519_172527.mp4','/home/vagrant/vagrant_data/oncoming/day/VID_20210519_173953.mp4','/home/vagrant/vagrant_data/oncoming/night/VID_20210517_204633.mp4','/home/vagrant/vagrant_data/oncoming/night/video_20210517_210030.mp4','/home/vagrant/vagrant_data/oncoming/night/VID_20210519_210202.mp4','/home/vagrant/vagrant_data/oncoming/night/VID_20210517_203448.mp4','/home/vagrant/vagrant_data/oncoming/night/VID_20210520_220338.mp4','/home/vagrant/vagrant_data/oncoming/night/VID_20210519_211225.mp4','/home/vagrant/vagrant_data/oncoming/night/VID_20210520_221422.mp4','/home/vagrant/vagrant_data/oncoming/night/VID_20210519_212354.mp4','/home/vagrant/vagrant_data/oncoming/night/VID_20210519_213427.mp4','/home/vagrant/vagrant_data/oncoming/night/VID_20210520_215048.mp4','/home/vagrant/vagrant_data/oncoming/night/VID_20210520_213903.mp4','/home/vagrant/vagrant_data/oncoming/night/VID_20210520_222512.mp4','/home/vagrant/vagrant_data/oncoming/night/VID_20210519_214505.mp4','/home/vagrant/vagrant_data/oncoming/night/VID_20210517_202236.mp4','/home/vagrant/vagrant_data/oncoming/night/video_20210517_211336.mp4']
#
# file=open('t.txt','w+')
# for i in range(len(id_list)):
#     file.write('turkic dump '+str(id_list[i])+' -o ../vagrant_data/txt_output/'+path_list[i].split('.')[0].split('/')[-1]+'.txt\n')
#     file.write('cp ' + path_list[i] + ' ../vagrant_data/txt_output/\n')
# file.close()
import codecs

file_list = ['1.mp4']

# file_list=['IMG_7835_batch.mov']
import cv2
import os
import numpy as np
import pandas as pd
from tqdm import trange

if not os.path.exists('./VOCdevkit/'):
    os.mkdir('./VOCdevkit/')
if not os.path.exists('./VOCdevkit/VOC2012/'):
    os.mkdir('./VOCdevkit/VOC2012')
if not os.path.exists('./VOCdevkit/VOC2012/Annotations/'):
    os.mkdir('./VOCdevkit/VOC2012/Annotations')
if not os.path.exists('./VOCdevkit/VOC2012/JPEGImages/'):
    os.mkdir('./VOCdevkit/VOC2012/JPEGImages')
if not os.path.exists('./VOCdevkit/VOC2012/ImageSets/'):
    os.mkdir('./VOCdevkit/VOC2012/ImageSets')

if not os.path.exists('./VOCdevkit/VOC2012/ImageSets/Main/'):
    os.mkdir('./VOCdevkit/VOC2012/ImageSets/Main/')


for f in file_list:
    print(f.split('.')[0] + '.txt')
    try:
        file = pd.read_csv(f.split('.')[0] + '.txt', sep=' ')
    except:
        print('empty file')
        continue

    cap = cv2.VideoCapture(f)
    # file = open(f.split('.')[0] + '.txt', 'r')
    # file=np.loadtxt(f.split('.')[0] + '.txt', delimiter=' ')
    file=np.array(file)
    if file.shape[0]==0:
        continue

    max_frame=np.max(file[:,5])
    for i in trange(max_frame):
        metrix=file[file[:,5]==i]
        if metrix.shape[0]==0:
            continue

        xml = codecs.open('./VOCdevkit/VOC2012/Annotations/' + f.split('.')[0] + '_' + str(i) + '.xml', 'w')
        cap.set(cv2.CAP_PROP_POS_FRAMES, i)
        flag, fra = cap.read()


        try:
            cv2.imwrite('./VOCdevkit/VOC2012/JPEGImages/' + f.split('.')[0] + '_' + str(i) + '.jpg', fra)
        except:
            ''
        xml.write('<?xml version="1.0" encoding="UTF-8"?>\n'+'<annotation>\n'+'\t<folder>' + 'VOC2012' + '</folder>\n'+'\t<filename>' + f.split('.')[0] + '_' + str(i) + '</filename>\n'
                  '\t<source>\n'+'\t\t<database> vatic-version1 </database>\n'+'\t</source>\n'+'\t<size>\n'
                  +'\t\t<width>' + str(fra.shape[1]) + '</width>\n'+'\t\t<height>' + str(fra.shape[0]) + '</height>\n'
                  +'\t\t<depth>' + str(3) + '</depth>\n'+'\t</size>\n'+'\t\t<segmented>0</segmented>\n')

        for j in range(metrix.shape[0]):
            Track, xmin, ymin, xmax, ymax, new_frame, lost, occluded, generated, label = metrix[j]
            if lost==1 or occluded ==1:
                continue
            if label=='car1' or label=='car2':
                xml.write('\t<object>\n'+'\t\t<name>'+'car'+'</name>\n'+'\t\t<pose>Unspecified</pose>\n'+'\t\t<truncated>0</truncated>\n'
                      +'\t\t<difficult>0</difficult>\n'+'\t\t<bndbox>\n'+'\t\t\t<xmin>' + str(xmin) + '</xmin>\n'+'\t\t\t<ymin>' + str(ymin) + '</ymin>\n'
                      +'\t\t\t<xmax>' + str(xmax) + '</xmax>\n'+'\t\t\t<ymax>' + str(ymax) + '</ymax>\n'+'\t\t</bndbox>\n'+'\t</object>\n')
            elif label=='vehicle1' or label=='vehicle2':
                xml.write('\t<object>\n'+'\t\t<name>'+'vehicle'+'</name>\n'+'\t\t<pose>Unspecified</pose>\n'+'\t\t<truncated>0</truncated>\n'
                      +'\t\t<difficult>0</difficult>\n'+'\t\t<bndbox>\n'+'\t\t\t<xmin>' + str(xmin) + '</xmin>\n'+'\t\t\t<ymin>' + str(ymin) + '</ymin>\n'
                      +'\t\t\t<xmax>' + str(xmax) + '</xmax>\n'+'\t\t\t<ymax>' + str(ymax) + '</ymax>\n'+'\t\t</bndbox>\n'+'\t</object>\n')
            else:
                xml.write('\t<object>\n'+'\t\t<name>'+label+'</name>\n'+'\t\t<pose>Unspecified</pose>\n'+'\t\t<truncated>0</truncated>\n'
                      +'\t\t<difficult>0</difficult>\n'+'\t\t<bndbox>\n'+'\t\t\t<xmin>' + str(xmin) + '</xmin>\n'+'\t\t\t<ymin>' + str(ymin) + '</ymin>\n'
                      +'\t\t\t<xmax>' + str(xmax) + '</xmax>\n'+'\t\t\t<ymax>' + str(ymax) + '</ymax>\n'+'\t\t</bndbox>\n'+'\t</object>\n')



        xml.write('</annotation>')
        xml.close()

3.label_visualize.py

import os
import cv2
import re
 
pattens = ['name','xmin','ymin','xmax','ymax']
 
def get_annotations(xml_path):
    bbox = []
    with open(xml_path,'r') as f:
        text = f.read().replace('\n','return')
        p1 = re.compile(r'(?<=<object>)(.*?)(?=</object>)')
        result = p1.findall(text)
        for obj in result:
            tmp = []
            for patten in pattens:
                p = re.compile(r'(?<=<{}>)(.*?)(?=</{}>)'.format(patten,patten))
                if patten == 'name':
                    tmp.append(p.findall(obj)[0])
                else:
                    tmp.append(int(float(p.findall(obj)[0])))
            bbox.append(tmp)
    return bbox
 
def save_viz_image(image_path,xml_path,save_path):
    bbox = get_annotations(xml_path)
    image = cv2.imread(image_path)
    for info in bbox:
        cv2.rectangle(image,(info[1],info[2]),(info[3],info[4]),(255,255,255),thickness=2)
        cv2.putText(image,info[0],(info[1],info[2]),cv2.FONT_HERSHEY_PLAIN,1.2,(255,255,255),2)
    if not os.path.exists(save_path):
        os.mkdir(save_path)
    name = os.path.join(save_path,image_path.split('/')[-1])
    cv2.imwrite(name,image)
 
if __name__ == '__main__':
    image_dir = '/home/yms/video1/VOCdevkit/VOC2012/JPEGImages/'
    xml_dir = '/home/yms/video1/VOCdevkit/VOC2012/Annotations/'
    save_dir = '/home/yms/video1/VOCdevkit/VOC2012/val_visualize/'
    image_list = os.listdir(image_dir)
    for i in  image_list:
        image_path = os.path.join(image_dir,i)
        xml_path = os.path.join(xml_dir,i.replace('.jpg','.xml'))
        save_viz_image(image_path,xml_path,save_dir)

4.set_ytrain_test.py

import os
import random

trainval_percent = 0.9      # 训练集和验证集 占 数据集 的比例
train_percent = 0.8        # 训练集 占 训练集和验证集 的比例
# 把xml路径修改为自己的Annotations文件夹路径
xmlfilepath = './VOCdevkit/VOC2012/Annotations'
# 把保存路径修改为自己的Main文件夹路径
savepath = './VOCdevkit/VOC2012/ImageSets/Main'
total_xml = os.listdir(xmlfilepath)

num = len(total_xml)
list = range(num)
tv = int(num * trainval_percent)
tr = int(tv * train_percent)
trainval = random.sample(list, tv)
train = random.sample(trainval, tr)

ftrainval = open(savepath + '/trainval.txt', 'w')
ftest = open(savepath + '/test.txt', 'w')
ftrain = open(savepath + '/train.txt', 'w')
fval = open(savepath + '/val.txt', 'w')

for i in list:
    name = total_xml[i][:-4] + '\n'
    if i in trainval:
        ftrainval.write(name)
        if i in train:
            ftrain.write(name)
        else:
            fval.write(name)
    else:
        ftest.write(name)

ftrainval.close()
ftrain.close()
fval.close()
ftest.close()

5.train.py

import os
import sys
import pickle
import random
import shutil
import getopt
from os.path import join
from os import listdir, getcwd
import xml.etree.ElementTree as xmlet

train_percent = 0.8
trainval_percent = 0.9

pwd = getcwd()
trainPath = pwd+ "/temp2Train"
trainLablePath = pwd+ "/temp2Train/labels"

train_classes = []
mask_classes = ['dotted line','w_solid_line','far_line','double_yellow_line','truck','y_solid_line','double_white_line','dotted_line','y_dotted_line','y_tail','y_head','solid_lines','solid_line','blank','dotted_lines','line','taill','tail ']
def creatTrainPath():
    if not os.path.isdir(trainPath):
        os.mkdir(trainPath)
        os.mkdir(trainPath + '/labels')
        os.mkdir(trainPath + '/JPEGImages')
        os.mkdir(trainPath + '/backup')
    else:
        shutil.rmtree(trainPath)
        os.mkdir(trainPath)
        os.mkdir(trainPath + '/labels')
        os.mkdir(trainPath + '/JPEGImages')
        os.mkdir(trainPath + '/backup')
 
def convert(size, box):
    dw = 1./(size[0])
    dh = 1./(size[1])
    x = (box[0] + box[1])/2.0 - 1
    y = (box[2] + box[3])/2.0 - 1
    w = box[1] - box[0]
    h = box[3] - box[2]
    x = x*dw
    w = w*dw
    y = y*dh
    h = h*dh
    return (x,y,w,h)

def convert_annotation(xmlfilepath,image_label):
    in_file = open(xmlfilepath + '/%s.xml'%(image_label))
    out_file = open(trainLablePath+ '/%s.txt'%(image_label), 'w')
    tree=xmlet.parse(in_file)
    root = tree.getroot()
    size = root.find('size')
    w = int(size.find('width').text)
    h = int(size.find('height').text)
    if w == 0 or h ==0:
        return 1

    for obj in root.iter('object'):
        difficult = obj.find('difficult').text
        cls = obj.find('name').text
        if int(difficult)==1:
            print("{}:{}.int(difficult)==1".format(in_file,cls))
            continue
        if cls in mask_classes:
	        continue     
        if cls not in train_classes:
            train_classes.append(cls)
            
        cls_id = train_classes.index(cls)
        xmlbox = obj.find('bndbox')
        b = (float(xmlbox.find('xmin').text), float(xmlbox.find('xmax').text), float(xmlbox.find('ymin').text), float(xmlbox.find('ymax').text))
        bb = convert((w,h), b)
        out_file.write(str(cls_id) + " " + " ".join([str(a) for a in bb]) + '\n')
        
    return 0
    
if __name__ == "__main__":
    
    yoloPath = ""
    dataSetPath = ""
    
    opts,args=getopt.getopt(sys.argv[1:],"i:ho:",["yolo=","dataset="])
    if len(opts) < 2:
        print("eg: python3 train.py --yolo=../darknet --dataset=./xxx")

    for opt,arg in opts:
        if opt=="--dataset" :
            dataSetPath = arg
        if opt=="--yolo" :   
            yoloPath = arg 

    if dataSetPath != "":
        creatTrainPath()
        
        #dataset done
        #JPEGImages and labels
        xmlfilepath = dataSetPath + '/Annotations'           
        total_xml = os.listdir(xmlfilepath)
        num = len(total_xml)
        list = range(num)
        tv = int(num * trainval_percent)
        tr = int(tv * train_percent)
        trainval = random.sample(list, tv)
        train = random.sample(trainval, tr)
        
        test_file = open(trainPath + '/test.txt', 'w')
        train_list = []
        val_list = []  
        for i in list:
            xmlName = total_xml[i][:-4]
            
            if not os.path.exists('%s/JPEGImages/%s.jpg'%(dataSetPath,xmlName)):
                continue
            
            if convert_annotation(xmlfilepath,xmlName):
                continue
                
            shutil.copy('%s/JPEGImages/%s.jpg'%(dataSetPath,xmlName),'%s/JPEGImages/%s.jpg'%(trainPath,xmlName))
            
            if i in trainval:
                if i in train:
                    train_list.append('%s/JPEGImages/%s.jpg\n'%(trainPath,xmlName))
                else:
                    val_list.append('%s/JPEGImages/%s.jpg\n'%(trainPath,xmlName))        
            else:
                test_file.write('%s/JPEGImages/%s.jpg\n'%(trainPath,xmlName))
                
        test_file.close()
        
        #train.txt val.txt
        val_file = open(trainPath + '/val.txt', 'w')
        train_file = open(trainPath + '/train.txt', 'w') 
        for item in val_list:
            train_file.write(item)
            val_file.write(item)    
        for item in train_list:
            train_file.write(item)               
        val_file.close()
        train_file.close()
        
        #train.names
        _file = open(trainPath + '/train.names', 'w')
        for item in  train_classes:
	        _file.write("{}\n".format(item))
        _file.close()  
        print(train_classes)

        #train.data
        _file = open(trainPath + '/train.data', 'w')
        _file.write("classes = {}\n".format(len(train_classes)))
        _file.write("train = {}\n".format(trainPath + '/train.txt'))
        _file.write("valid = {}\n".format(trainPath + '/val.txt'))
        _file.write("names = {}\n".format(trainPath + '/train.names'))
        _file.write("backup = {}\n".format(trainPath + '/backup/'))
        _file.close()

6.voc_label_new.py

import xml.etree.ElementTree as ET
import pickle
import os
from os import listdir, getcwd
from os.path import join

sets = [('2012', 'train'),('2012', 'val'),('2012', 'test')]

classes = ["bicycle","worker","person"]# 修改为自己的类别


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

def convert_annotation(year, image_id):
    in_file = open('VOCdevkit/VOC%s/Annotations/%s.xml'%(year, image_id))
    out_file = open('VOCdevkit/VOC%s/labels/%s.txt'%(year, image_id), 'w')
    tree=ET.parse(in_file)
    root = tree.getroot()
    size = root.find('size')
    w = int(size.find('width').text)
    h = int(size.find('height').text)

    for obj in root.iter('object'):
        difficult = obj.find('difficult').text
        cls = obj.find('name').text
        if cls not in classes or int(difficult)==1:
            continue
        cls_id = classes.index(cls)
        xmlbox = obj.find('bndbox')
        b = (float(xmlbox.find('xmin').text), float(xmlbox.find('xmax').text), float(xmlbox.find('ymin').text), float(xmlbox.find('ymax').text))
        bb = convert((w,h), b)
        out_file.write(str(cls_id) + " " + " ".join([str(a) for a in bb]) + '\n')

wd = getcwd()

for year, image_set in sets:
    if not os.path.exists('VOCdevkit/VOC%s/labels/'%(year)):
        os.makedirs('VOCdevkit/VOC%s/labels/'%(year))
    image_ids = open('VOCdevkit/VOC%s/ImageSets/Main/%s.txt'%(year, image_set)).read().strip().split()
    list_file = open('%s_%s.txt'%(year, image_set), 'w')
    for image_id in image_ids:
        #list_file.write('./VOCdevkit/VOC%s/JPEGImages/%s.jpg\n'%(year, image_id))
        try:
            convert_annotation(year, image_id)
            list_file.write('%s/VOCdevkit/VOC%s/JPEGImages/%s.jpg\n'%(wd, year, image_id))
        except:
            print("xml size == 0")
    list_file.close()

os.system("cat 2012_train.txt 2012_val.txt > train.txt")
os.system("cat 2012_train.txt 2012_val.txt 2012_test.txt > train.all.txt")