物体检测数据集处理总结

1. 标注工具Labelme的安装。

​​​​​​​

安装pyqt

pip install PyQt5

 

安装 PIL包

pip install Pillow

安装labelme

1、官方的安装命令。

pip install labelme

 

2、官方的labelme不支持大图像的打开，比如遥感图像。如果遇到图片打不开的情况，可以使用我修改的labelme。

labelme-master（修改后支持大图像）.zip

https://download.csdn.net/download/hhhhhhhhhhwwwwwwwwww/12047245

1. 标注工具的使用
2. VOC格式的数据逆向转为Labelme标注的数据集

import sys
import os.path as osp
import io
from labelme.logger import logger
from labelme import PY2
from labelme import QT4
import PIL.Image
import base64
from labelme import utils
import os
import cv2
import xml.etree.ElementTree as ET
module_path = os.path.abspath(os.path.join('..'))
if module_path not in sys.path:
    sys.path.append(module_path)
import json
from PIL import Image
Image.MAX_IMAGE_PIXELS = None
imageroot = 'RSOD/'
listDir = ['aircraft', 'oiltank']

def load_image_file(filename):
    try:
        image_pil = PIL.Image.open(filename)
    except IOError:
        logger.error('Failed opening image file: {}'.format(filename))
        return
    # apply orientation to image according to exif
    image_pil = utils.apply_exif_orientation(image_pil)
    with io.BytesIO() as f:
        ext = osp.splitext(filename)[1].lower()
        if PY2 and QT4:
            format = 'PNG'
        elif ext in ['.jpg', '.jpeg']:
            format = 'JPEG'
        else:
            format = 'PNG'
        image_pil.save(f, format=format)
        f.seek(0)
        return f.read()
def dict_json(flags, imageData, shapes, imagePath, fillColor=None, lineColor=None, imageHeight=100, imageWidth=100):
    '''
    :param imageData: str
    :param shapes: list
    :param imagePath: str
    :param fillColor: list
    :param lineColor: list
    :return: dict
    '''
    return {"version": "3.16.4", "flags": flags, "shapes": shapes, 'lineColor': lineColor, "fillColor": fillColor,
            'imagePath': imagePath.split('/')[1], "imageData": imageData, 'imageHeight': imageHeight,
            'imageWidth': imageWidth}
data = json.load(open('1.json'))
for subPath in listDir:
    xmlpathName = imageroot + subPath + '/Annotation/xml'
    imagepath = imageroot + subPath + '/JPEGImages'
    resultFile = os.listdir(xmlpathName)
    for file in resultFile:
        print(file)
        imagePH = imagepath + '/' + file.split('.')[0] + '.jpg'
        print(imagePH)
        tree = ET.parse(xmlpathName + '/' + file)
        image = cv2.imread(imagePH)
        shapes = data["shapes"]
        version = data["version"]
        flags = data["flags"]
        lineColor = data["lineColor"]
        fillColor = data['fillColor']
        newshapes = []
        for elem in tree.iter():
            if 'object' in elem.tag:
                name = ''
                xminNode = 0
                yminNode = 0
                xmaxNode = 0
                ymaxNode = 0
                for attr in list(elem):
                    if 'name' in attr.tag:
                        name = attr.text
                    if 'bndbox' in attr.tag:
                        for dim in list(attr):
                            if 'xmin' in dim.tag:
                                xminNode = int(round(float(dim.text)))
                            if 'ymin' in dim.tag:
                                yminNode = int(round(float(dim.text)))
                            if 'xmax' in dim.tag:
                                xmaxNode = int(round(float(dim.text)))
                            if 'ymax' in dim.tag:
                                ymaxNode = int(round(float(dim.text)))
                line_color = None
                fill_color = None
                newPoints = [[float(xminNode), float(yminNode)], [float(xmaxNode), float(ymaxNode)]]
                shape_type = 'rectangle'
                flags = flags
                newshapes.append(
                    {"label": name, "line_color": line_color, "fill_color": fill_color, "points": newPoints,
                     "shape_type": shape_type, "flags": flags})
        imageData_90 = load_image_file(imagePH)
        imageData_90 = base64.b64encode(imageData_90).decode('utf-8')
        imageHeight = image.shape[0]
        imageWidth = image.shape[1]
        data_90 = dict_json(flags, imageData_90, newshapes, imagePH, fillColor, lineColor, imageHeight, imageWidth)
        json_file = imagePH[:-4] + '.json'
        json.dump(data_90, open(json_file, 'w'))

1. Labelme标注的数据集转VOC2007格式的数据集。

VOC2007数据文件夹说明

1)JPEGImages文件夹

文件夹里包含了训练图片和测试图片，混放在一起

2)Annatations文件夹

文件夹存放的是xml格式的标签文件，每个xml文件都对应于JPEGImages文件夹的一张图片

3)ImageSets文件夹

Action存放的是人的动作，我们暂时不用

Layout存放的人体部位的数据。我们暂时不用

Main存放的是图像物体识别的数据，Main里面有test.txt, train.txt, val.txt,trainval.txt.这四个文件我们后面会生成

XML说明

        

                   optic rs image

                   Lmars RSDS2016

                   0

                   Lmars Detection Dataset of RS

        

        

         aircraft_773.jpg

        

         0

        

                   Lmars, Wuhan University

                   I do not know

        

         RSDS2016

        

                   1044

                   3

                   915

        

完整代码：

import os

from typing import List, Any



import numpy as np

import codecs

import json

from glob import glob

import cv2

import shutil

from sklearn.model_selection import train_test_split

# 1.标签路径

labelme_path = "LabelmeData/"  # 原始labelme标注数据路径

saved_path = "VOC2007/"  # 保存路径

isUseTest=True#是否创建test集

# 2.创建要求文件夹

if not os.path.exists(saved_path + "Annotations"):

    os.makedirs(saved_path + "Annotations")

if not os.path.exists(saved_path + "JPEGImages/"):

    os.makedirs(saved_path + "JPEGImages/")

if not os.path.exists(saved_path + "ImageSets/Main/"):

    os.makedirs(saved_path + "ImageSets/Main/")

# 3.获取待处理文件

files = glob(labelme_path + "*.json")

files = [i.replace("\\","/").split("/")[-1].split(".json")[0] for i in files]

print(files)

# 4.读取标注信息并写入 xml

for json_file_ in files:

    json_filename = labelme_path + json_file_ + ".json"

    json_file = json.load(open(json_filename, "r", encoding="utf-8"))

    height, width, channels = cv2.imread(labelme_path + json_file_ + ".jpg").shape

    with codecs.open(saved_path + "Annotations/" + json_file_ + ".xml", "w", "utf-8") as xml:

        xml.write('\n')

        xml.write('\t' + 'WH_data' + '\n')

        xml.write('\t' + json_file_ + ".jpg" + '\n')

        xml.write('\t\n')

        xml.write('\t\tWH Data\n')

        xml.write('\t\tWH\n')

        xml.write('\t\tflickr\n')

        xml.write('\t\tNULL\n')

        xml.write('\t\n')

        xml.write('\t\n')

        xml.write('\t\tNULL\n')

        xml.write('\t\tWH\n')

        xml.write('\t\n')

        xml.write('\t\n')

        xml.write('\t\t' + str(width) + '\n')

        xml.write('\t\t' + str(height) + '\n')

        xml.write('\t\t' + str(channels) + '\n')

        xml.write('\t\n')

        xml.write('\t\t0\n')

        for multi in json_file["shapes"]:

            points = np.array(multi["points"])

            labelName=multi["label"]

            xmin = min(points[:, 0])

            xmax = max(points[:, 0])

            ymin = min(points[:, 1])

            ymax = max(points[:, 1])

            label = multi["label"]

            if xmax <= xmin:

                pass

            elif ymax <= ymin:

                pass

            else:

                xml.write('\t\n')

                print(json_filename, xmin, ymin, xmax, ymax, label)

        xml.write('')

# 5.复制图片到 VOC2007/JPEGImages/下

image_files = glob(labelme_path + "*.jpg")

print("copy image files to VOC007/JPEGImages/")

for image in image_files:

    shutil.copy(image, saved_path + "JPEGImages/")

# 6.split files for txt

txtsavepath = saved_path + "ImageSets/Main/"

ftrainval = open(txtsavepath + '/trainval.txt', 'w')

ftest = open(txtsavepath + '/test.txt', 'w')

ftrain = open(txtsavepath + '/train.txt', 'w')

fval = open(txtsavepath + '/val.txt', 'w')

total_files = glob("./VOC2007/Annotations/*.xml")

total_files = [i.replace("\\","/").split("/")[-1].split(".xml")[0] for i in total_files]

trainval_files=[]

test_files=[]

if isUseTest:

    trainval_files, test_files = train_test_split(total_files, test_size=0.15, random_state=55)

else:

    trainval_files=total_files

for file in trainval_files:

    ftrainval.write(file + "\n")

# split

train_files, val_files = train_test_split(trainval_files, test_size=0.15, random_state=55)

# train

for file in train_files:

    ftrain.write(file + "\n")

# val

for file in val_files:

    fval.write(file + "\n")

for file in test_files:

    print(file)

    ftest.write(file + "\n")

ftrainval.close()

ftrain.close()

fval.close()

ftest.close()

注：训练集和验证集的划分方法是采用 sklearn.model_selection.train_test_split 进行分割的。

1. Labelme转VOC，将没有标注的数据生成测试集，并统计每个类别的个数。

对标注数据的统计。

 

将没有标注的数据放在test文件夹下面。

 

import os
import numpy as np
import codecs
import json
from glob import glob
import cv2
import shutil
from sklearn.model_selection import train_test_split
dicImg={}
imgLis=[]
# 1.标签路径
labelme_path = "USA/"  # 原始labelme标注数据路径
saved_path = "VOC2007/"  # 保存路径
# 2.创建要求文件夹
if not os.path.exists(saved_path + "Annotations"):
    os.makedirs(saved_path + "Annotations")
if not os.path.exists(saved_path + "JPEGImages/"):
    os.makedirs(saved_path + "JPEGImages/")
if not os.path.exists(saved_path + "test/"):
    os.makedirs(saved_path + "test/")
if not os.path.exists(saved_path + "ImageSets/Main/"):
    os.makedirs(saved_path + "ImageSets/Main/")
# 3.获取待处理文件
files = glob(labelme_path + "*.json")
files = [i.replace("\\","/").split("/")[-1].split(".json")[0] for i in files]
print(files)
# 4.读取标注信息并写入 xml
for json_file_ in files:
    json_filename = labelme_path + json_file_ + ".json"
    json_file = json.load(open(json_filename, "r", encoding="utf-8"))
    height, width, channels = cv2.imread(labelme_path + json_file_ + ".jpg").shape
    with codecs.open(saved_path + "Annotations/" + json_file_ + ".xml", "w", "utf-8") as xml:
        xml.write('\n')
        xml.write('\t' + 'UAV_data' + '\n')
        xml.write('\t' + json_file_ + ".jpg" + '\n')
        xml.write('\t\n')
        xml.write('\t\tThe UAV autolanding\n')
        xml.write('\t\tUAV AutoLanding\n')
        xml.write('\t\tflickr\n')
        xml.write('\t\tNULL\n')
        xml.write('\t\n')
        xml.write('\t\n')
        xml.write('\t\tNULL\n')
        xml.write('\t\twanghao\n')
        xml.write('\t\n')
        xml.write('\t\n')
        xml.write('\t\t' + str(width) + '\n')
        xml.write('\t\t' + str(height) + '\n')
        xml.write('\t\t' + str(channels) + '\n')
        xml.write('\t\n')
        xml.write('\t\t0\n')
        for multi in json_file["shapes"]:
            points = np.array(multi["points"])
            labelName=multi["label"].lower()
            if labelName in dicImg:
                count=dicImg[labelName]
                count=count+1;
                dicImg[labelName]=count
            else:
                dicImg[labelName]=1
            xmin = min(points[:, 0])
            xmax = max(points[:, 0])
            ymin = min(points[:, 1])
            ymax = max(points[:, 1])
            label = multi["label"]
            if xmax <= xmin:
                pass
            elif ymax <= ymin:
                pass
            else:
                xml.write('\t\n')
                print(json_filename, xmin, ymin, xmax, ymax, label)
        xml.write('')
    imgLis.append(json_file_+'.jpg')
    shutil.copy(labelme_path+json_file_+'.jpg', saved_path + "JPEGImages/")
print(imgLis)
# 5.复制图片到 VOC2007/JPEGImages/下
image_files = glob(labelme_path + "*.jpg")
txtsavepath = saved_path + "ImageSets/Main/"
print("copy image files to VOC007/JPEGImages/")
ftest = open(txtsavepath + '/test.txt', 'w')
for image in image_files:
    if image.split('\\')[1] not in imgLis:
        print(image)
        shutil.copy(image, saved_path + "test/")
        ftest.write(image.replace("\\","/").split("/")[-1].split(".jpg")[0] + "\n")
# 6.split files for txt
ftrainval = open(txtsavepath + '/trainval.txt', 'w')
ftrain = open(txtsavepath + '/train.txt', 'w')
fval = open(txtsavepath + '/val.txt', 'w')
total_files = glob("./VOC2007/Annotations/*.xml")
total_files = [i.replace("\\","/").split("/")[-1].split(".xml")[0] for i in total_files]
for file in total_files:
    ftrainval.write(file + "\n")
# split
train_files, val_files = train_test_split(total_files, test_size=0.15, random_state=42)
# train
for file in train_files:
    ftrain.write(file + "\n")
# val
for file in val_files:
    fval.write(file + "\n")
ftrainval.close()
ftrain.close()
fval.close()
print(dicImg)
ftest.close()

 

1. 将Labelme标注的数据转为txt格式的数据集。

​​​​​​​

 

一张图像对应一个txt，txt中每行对应一个标记物体。

格式：类别 xmin ymin xmax ymax

import json
import os
from glob import glob
import shutil
# convert labelme json to DOTA txt format
def custombasename(fullname):
    return os.path.basename(os.path.splitext(fullname)[0])
IN_PATH = 'USA'
OUT_PATH = 'labeltxt'
if not os.path.exists(OUT_PATH):
    os.makedirs(OUT_PATH)
file_list = glob(IN_PATH + '/*.json')
for i in range(len(file_list)):
    with open(file_list[i]) as f:
        label_str = f.read()
        label_dict = json.loads(label_str)  # json文件读入dict
        imgepath=file_list[i].split('.')[0]+'.jpg'
        # 输出 txt 文件的路径
        out_file = OUT_PATH + '/' + custombasename(file_list[i]) + '.txt'
        shutil.copy(imgepath, OUT_PATH)
        # 写入 poly 四点坐标 和 label
        fout = open(out_file, 'w')
        out_str = ''
        for shape_dict in label_dict['shapes']:
            out_str += shape_dict['label'] + ' '
            points = shape_dict['points']
            for p in points:
                out_str += (str(p[0]) + ' ' + str(p[1]) + ' ')
            out_str +='\n'
        fout.write(out_str)
        fout.close()
    print('%d/%d' % (i + 1, len(file_list)))

 

1. 对Labelme标注图像，进行90、180、270的旋转，实现标注数据的扩充。

在制作做遥感图像物体检测数据集的时候，遥感图像的物体都是平面的，有角度的问题，

可以对被检测物体实现不同角度的旋转，丰富数据集同时减少标注的工作量。

    

 

比如上图中的飞机，机头的朝向是斜向下的，现实中的飞机可能有各种的朝向，如果不做旋转，就会降低模型的检测能力。下图是旋转90度的效果。

  

 

需要安装的包：

labelme

scipy1.0.0版本

pyqt5

 

旋转最大的难点在于旋转后，需要对标注的点重新计算，保证标注的坐标不出现错乱。

 旋转90度后，坐标转化：

  points=shapelabel['points']#获取初始的坐标。
  newPoints = [[float(points[0][1]), w-float(points[1][0])],
                 [float(points[1][1]), w-float(points[0][0])]]#旋转90度，重新对应坐标。w表示原始图像的宽度。
选旋转180度后，坐标转化：

points = shapelabel['points']
newPoints = [[w-float(points[1][0]), h - float(points[1][1])],
                     [w-float(points[0][0]), h - float(points[0][1])]] #旋转180度，重新对应坐标。h表示原始图像的高度。
旋转270度，坐标转化：

points = shapelabel['points']

newPoints = [[h - float(points[1][1]), float(points[0][0])],
                     [h - float(points[0][1]),  float(points[1][0])]]

完整代码如下：

#scipy的版本为1.0.0

import scipy

from scipy import misc

import os

import glob

import PIL.Image

from labelme.logger import logger

from labelme import PY2

from labelme import QT4

import io

import json

import os.path as osp

import PIL.Image

from scipy import ndimage

import base64

from labelme import utils

def load_image_file(filename):
    try:
        image_pil = PIL.Image.open(filename)
    except IOError:
        logger.error('Failed opening image file: {}'.format(filename))
        return
    # apply orientation to image according to exif
    image_pil = utils.apply_exif_orientation(image_pil)
    with io.BytesIO() as f:
        ext = osp.splitext(filename)[1].lower()
        if PY2 and QT4:
            format = 'PNG'
        elif ext in ['.jpg', '.jpeg']:
            format = 'JPEG'
        else:
            format = 'PNG'
        image_pil.save(f, format=format)
        f.seek(0)
        return f.read()
def dict_json(flags,imageData,shapes,imagePath,fillColor=None,lineColor=None,imageHeight=100,imageWidth=100):
    '''
    :param imageData: str
    :param shapes: list
    :param imagePath: str
    :param fillColor: list
    :param lineColor: list
    :return: dict
    '''
    return {"version":"3.16.4","flags":flags,"shapes":shapes,'lineColor':lineColor,"fillColor":fillColor,'imagePath':imagePath.split('\\')[1],"imageData":imageData,'imageHeight':imageHeight,'imageWidth':imageWidth}
def get_image_paths(folder):
    return glob.glob(os.path.join(folder, '*.jpg'))
def create_read_img(filename):
    data = json.load(open(filename.split('.')[0]+'.json'))
    shapes = data["shapes"]
    version = data["version"]
    flags = data["flags"]
    lineColor = data["lineColor"]
    fillColor = data['fillColor']
    newshapes = []
    im = misc.imread(filename)
    h,w,d=im.shape
    img_rote_90 = ndimage.rotate(im, 90)
    img_path_90=filename[:-4]+'_90.jpg'
    scipy.misc.imsave(img_path_90,img_rote_90)
    imageData_90 = load_image_file(img_path_90)
    imageData_90 = base64.b64encode(imageData_90).decode('utf-8')
    imageHeight =w
    imageWidth = h
    for shapelabel in shapes:
        newLabel=shapelabel['label']
        newline_color=shapelabel['line_color']
        newfill_color=shapelabel['fill_color']
        points=shapelabel['points']
        newPoints = [[float(points[0][1]), w-float(points[1][0])],
                 [float(points[1][1]), w-float(points[0][0])]]
        newshape_type=shapelabel['shape_type']
        newflags=shapelabel['flags']      newshapes.append({'label':newLabel,'line_color':newline_color,'fill_color':newfill_color,'points':newPoints,'shape_type':newshape_type,'flags':newflags})
    data_90 = dict_json(flags, imageData_90, newshapes, img_path_90, fillColor, lineColor, imageHeight, imageWidth)
    json_file = img_path_90[:-4] + '.json'
    json.dump(data_90, open(json_file, 'w'))
    img_rote_180 = ndimage.rotate(im, 180)
    img_path_180=filename[:-4]+'_180.jpg'
    scipy.misc.imsave(img_path_180,img_rote_180)
    imageData_180 = load_image_file(img_path_180)
    imageData_180 = base64.b64encode(imageData_180).decode('utf-8')
    imageHeight = h
    imageWidth = w
    newshapes = []
    for shapelabel in shapes:
        newLabel = shapelabel['label']
        newline_color = shapelabel['line_color']
        newfill_color = shapelabel['fill_color']
        points = shapelabel['points']
        newPoints = [[w-float(points[1][0]), h - float(points[1][1])],
                     [w-float(points[0][0]), h - float(points[0][1])]]
        newshape_type = shapelabel['shape_type']
        newflags = shapelabel['flags']
        newshapes.append(
            {'label': newLabel, 'line_color': newline_color, 'fill_color': newfill_color, 'points': newPoints,
             'shape_type': newshape_type, 'flags': newflags})
    data_180 = dict_json(flags, imageData_180, newshapes, img_path_180, fillColor, lineColor, imageHeight, imageWidth)
    json_file = img_path_180[:-4] + '.json'
    json.dump(data_180, open(json_file, 'w'))
    img_rote_270 = ndimage.rotate(im, 270)
    img_path_270=filename[:-4]+'_270.jpg'
    scipy.misc.imsave(img_path_270,img_rote_270)
    imageData_270 = load_image_file(img_path_270)
    imageData_270 = base64.b64encode(imageData_270).decode('utf-8')
    imageHeight = w
    imageWidth = h
    newshapes = []
    for shapelabel in shapes:
        newLabel = shapelabel['label']
        newline_color = shapelabel['line_color']
        newfill_color = shapelabel['fill_color']
        points = shapelabel['points']
        newPoints = [[h - float(points[1][1]), float(points[0][0])],
                     [h - float(points[0][1]),  float(points[1][0])]]
        newshape_type = shapelabel['shape_type']
        newflags = shapelabel['flags']
        newshapes.append(
            {'label': newLabel, 'line_color': newline_color, 'fill_color': newfill_color, 'points': newPoints,
             'shape_type': newshape_type, 'flags': newflags})
    data_270 = dict_json(flags, imageData_270, newshapes, img_path_270, fillColor, lineColor, imageHeight, imageWidth)
    json_file = img_path_270[:-4] + '.json'
    json.dump(data_270, open(json_file, 'w'))
    print(filename)
img_path = 'USA'  #这个路径是所有图片在的位置
imgs = get_image_paths(img_path)
print (imgs)
for i in imgs:
    create_read_img(i)

1. 对标注格式为txt的数据集，实现90、180、270度的旋转

#scipy的版本为1.0.0

import scipy

from scipy import misc

import os

import glob

from scipy import ndimage



def get_image_paths(folder):

    return glob.glob(os.path.join(folder, '*.jpg'))

def create_read_img(filename):

    objectList = []

    with open(filename.split('.')[0] + ".txt") as f:

        for line in f.readlines():

            for aa in line.split(' '):

                if aa!='\n':

                    objectList.append(aa)

    im = misc.imread(filename)

    h,w,d=im.shape

    img_rote_90 = ndimage.rotate(im, 90)

    img_path_90=filename[:-4]+'_90.jpg'

    scipy.misc.imsave(img_path_90,img_rote_90)

    img_path_90_txt=img_path_90[:-4]+'.txt'

    outLable = ''

    for i in range(int(len(objectList)/5)):

        object_label = objectList[i * 5]

        outLable+=object_label+' '

        object_x1 = objectList[i * 5 + 1]

        object_y1 = objectList[i * 5 + 2]

        object_x2 = objectList[i * 5 + 3]

        object_y2 = objectList[i * 5 + 4]

        outLable += object_y1 + ' '

        outLable += str(w-float(object_x2)) + ' '

        outLable += object_y2 + ' '

        outLable += str(w-float(object_x1)) + '\n'

    fout = open(img_path_90_txt, 'w')

    fout.write(outLable)

    fout.close()

    img_rote_180 = ndimage.rotate(im, 180)

    img_path_180=filename[:-4]+'_180.jpg'

    scipy.misc.imsave(img_path_180,img_rote_180)

    img_path_180_txt = img_path_180[:-4] + '.txt'

    outLable = ''

    for i in range(int(len(objectList) / 5)):

        object_label = objectList[i * 5]

        outLable += object_label + ' '

        object_x1 = objectList[i * 5 + 1]

        object_y1 = objectList[i * 5 + 2]

        object_x2 = objectList[i * 5 + 3]

        object_y2 = objectList[i * 5 + 4]

        outLable += str(w-float(object_x2)) + ' '

        outLable += str(h-float(object_y2) )+ ' '

        outLable += str(w-float(object_x1)) + ' '

        outLable += str(h - float(object_y1)) + '\n'

    fout = open(img_path_180_txt, 'w')

    fout.write(outLable)

    fout.close()

    img_rote_270 = ndimage.rotate(im, 270)

    img_path_270=filename[:-4]+'_270.jpg'

    scipy.misc.imsave(img_path_270,img_rote_270)

    img_path_270_txt = img_path_270[:-4] + '.txt'

    outLable = ''

    for i in range(int(len(objectList) / 5)):

        object_label = objectList[i * 5]

        outLable += object_label + ' '

        object_x1 = objectList[i * 5 + 1]

        object_y1 = objectList[i * 5 + 2]

        object_x2 = objectList[i * 5 + 3]

        object_y2 = objectList[i * 5 + 4]

        outLable += str(h-float(object_y2)) + ' '

        outLable += (object_x1) + ' '

        outLable +=str (h-float(object_y1)) + ' '

        outLable += (object_x2) + '\n'

    fout = open(img_path_270_txt, 'w')

    fout.write(outLable)

    fout.close()

    print(filename)

img_path = 'CutResult'  #这个路径是所有图片在的位置

imgs = get_image_paths(img_path)

print (imgs)

for i in imgs:

    create_read_img(i)