YOLO数据集划分脚本整合

import os
import cv2
import sys
import random
import xml.etree.ElementTree as ET


def mkdir(path):
    if not os.path.exists(path):
        os.makedirs(path)


def divide(filename, trainval_percent, train_percent):

    xmlfilepath = '%s/Annotations' % filename
    txtsavepath = '%s/ImageSets' % filename
    mkdir(txtsavepath)
    try:
        total_xml = os.listdir(xmlfilepath)
    except:
        mkdir(xmlfilepath)
        print('--------Annotations is empty--------')
        sys.exit()
        
    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('%s/ImageSets/trainval.txt' % filename, 'w')
    ftest = open('%s/ImageSets/test.txt' % filename, 'w')
    ftrain = open('%s/ImageSets/train.txt' % filename, 'w')
    fval = open('%s/ImageSets/val.txt' % filename, '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()


def convert(size, box):

    dw = 1. / size[0]
    dh = 1. / size[1]
    x = (box[0] + box[1]) / 2.0  
    y = (box[2] + box[3]) / 2.0 
    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(image_id, filename, classes):

    in_file = open((filename + '/Annotations/%s.xml') % (image_id), encoding='utf-8')
    out_file = open((filename + '/labels/%s.txt') % (image_id), 'w', encoding='utf-8')
    tree = ET.parse(in_file)
    root = tree.getroot()
    size = root.find('size')
    if size is not None:
        w = int(size.find('width').text)
        h = int(size.find('height').text)

        for obj in root.iter('object'):
            cls = obj.find('name').text
            if cls not in classes:
                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')


def make_txt(filename, sets, classes):

    for image_set in sets:
        mkdir(filename + '/labels/')
        image_ids = open('%s/ImageSets/%s.txt' % (filename, image_set)).read().strip().split()
        list_file = open('%s/%s.txt' % (filename, image_set), 'w')
        for image_id in image_ids:
            list_file.write('%s/images/%s.jpg\n' % (filename, image_id))
            convert_annotation(image_id, filename, classes)
        list_file.close()


def read_pic(filename):

    image_ids = open('%s/ImageSets/test.txt' % filename).read().strip().split()
    list_file = open('%s/test.txt' % filename).read().strip().split()
    for image in range(len(list_file)):
        img = cv2.imread(list_file[image])
        cv2.imwrite('%s/inference/%s.jpg' % (filename, image_ids[image]), img)


if __name__ == '__main__':

	# (trainval = train + val) + test = 1
    filename = '../yolov5/data'
    trainval_percent = 0.8
    train_percent = 0.75

    sets = ['train', 'val', 'test']
    classes = ["Car", "Bus"]

    # 比例划分
    divide(filename, trainval_percent, train_percent)
    # 导出对应图片路径的txt
    make_txt(filename, sets, classes)
    # 根据txt提取测试集
    read_pic(filename)