需求

由于之前切分的图像有一些存在没有缺陷，需要再次筛选

将可视化的图像更改后缀

更改为xml的

可视化代码

可视化后只有7000多个图像

原本的图像有1W多张

代码

# 按照xml文件删除对应的图片
# coding: utf-8
from PIL import Image, ImageDraw, ImageFont
import os
import shutil
import cv2 as cv
import numpy as np
import json

#删除重名图象文件

def read_json(file_name):
    with open(file_name, 'rb') as f:
        data = json.load(f)
    return data

#读取源配置文件的路径
def delete1(image_root_path,suffix):
    delImg = []
    for root, dirs, files in os.walk(image_root_path):
        # 使用os模块获取文件夹中所有文件的路径
        all_files = os.listdir(root)
        filtered_files = [file for file in all_files if file.endswith(suffix)]
        if (len(filtered_files)):
            # print(root)  #当前工作文件夹
            for i in range(len(filtered_files)):
                bmp_path = root + "\\" + filtered_files[i]
                delImg.append(filtered_files[i])
    return delImg

#在目标文件中删除图象
def delete2(image_root_path,suffix,delImg):
    del_delImg = []
    for root, dirs, files in os.walk(image_root_path):
        # 使用os模块获取文件夹中所有文件的路径
        all_files = os.listdir(root)
        filtered_files = [file for file in all_files if file.endswith(suffix)]
        if (len(filtered_files)):
            print(root)  #当前工作文件夹
            for i in range(len(filtered_files)):
                bmp_path = root + "\\" + filtered_files[i]
                for j in range(len(delImg)):
                    if(delImg[j]==filtered_files[i]):
                        print(bmp_path)
                        del_delImg.append(bmp_path)
    return del_delImg


def myCopyImg(del_delImg,desPath):
    if(len(del_delImg)>0):
        for i in range(len(del_delImg)):
            bmp_path = del_delImg[i]
            #os.remove(bmp_path)
            shutil.copy(bmp_path, desPath)  # shutil.copy函数放入原文件的路径文件全名  然后放入目标文件夹
    else:
        print("无文件")

#替换列表中文件的后缀
def tihuanhouozhui(delImg):

    old_suffix1 = '.txt'
    old_suffix2 = '.xml'
    new_suffix = '.bmp'
    delImg = [file.replace(old_suffix1, new_suffix) for file in delImg]
    delImg = [file.replace(old_suffix2, new_suffix) for file in delImg]
    return delImg


if __name__ == "__main__":
    #my_copy("./1/kuaisu.json","D:\\code\\select\\1\\","D:\\code\\select\\final\\")
    delImg1 = delete1("G:\\datasets\\coco\\vis",".xml")
    #delImg2 = delete1("E:\\黄花标注\\glass\\testtt\\yesann", ".xml")
    #delImg = delImg1+delImg2
    #delImg = delImg1
    #new_delImg = tihuanhouozhui(delImg1)

    # for i in range(len(delImg)):
    #     os.remove(delImg[i])
    del_delImg = delete2("G:\\datasets\\coco\\split\\xml",".xml",delImg1)
    desPath = "G:\\datasets\\coco\\final"
    myCopyImg(del_delImg,desPath)

将xml复制过来了

同理复制jpg

也是先改名
（将vis中的文件改为jpg）

总图像文件：

然后使用代码：

# 按照xml文件删除对应的图片
# coding: utf-8
from PIL import Image, ImageDraw, ImageFont
import os
import shutil
import cv2 as cv
import numpy as np
import json

#删除重名图象文件

def read_json(file_name):
    with open(file_name, 'rb') as f:
        data = json.load(f)
    return data

#读取源配置文件的路径
def delete1(image_root_path,suffix):
    delImg = []
    for root, dirs, files in os.walk(image_root_path):
        # 使用os模块获取文件夹中所有文件的路径
        all_files = os.listdir(root)
        filtered_files = [file for file in all_files if file.endswith(suffix)]
        if (len(filtered_files)):
            # print(root)  #当前工作文件夹
            for i in range(len(filtered_files)):
                bmp_path = root + "\\" + filtered_files[i]
                delImg.append(filtered_files[i])
    return delImg

#在目标文件中删除图象
def delete2(image_root_path,suffix,delImg):
    del_delImg = []
    for root, dirs, files in os.walk(image_root_path):
        # 使用os模块获取文件夹中所有文件的路径
        all_files = os.listdir(root)
        filtered_files = [file for file in all_files if file.endswith(suffix)]
        if (len(filtered_files)):
            print(root)  #当前工作文件夹
            for i in range(len(filtered_files)):
                bmp_path = root + "\\" + filtered_files[i]
                for j in range(len(delImg)):
                    if(delImg[j]==filtered_files[i]):
                        print(bmp_path)
                        del_delImg.append(bmp_path)
    return del_delImg


def myCopyImg(del_delImg,desPath):
    if(len(del_delImg)>0):
        for i in range(len(del_delImg)):
            bmp_path = del_delImg[i]
            #os.remove(bmp_path)
            shutil.copy(bmp_path, desPath)  # shutil.copy函数放入原文件的路径文件全名  然后放入目标文件夹
    else:
        print("无文件")

#替换列表中文件的后缀
def tihuanhouozhui(delImg):

    old_suffix1 = '.txt'
    old_suffix2 = '.xml'
    new_suffix = '.bmp'
    delImg = [file.replace(old_suffix1, new_suffix) for file in delImg]
    delImg = [file.replace(old_suffix2, new_suffix) for file in delImg]
    return delImg


if __name__ == "__main__":
    #my_copy("./1/kuaisu.json","D:\\code\\select\\1\\","D:\\code\\select\\final\\")
    delImg1 = delete1("G:\\datasets\\coco\\vis",".jpg")
    #delImg2 = delete1("E:\\黄花标注\\glass\\testtt\\yesann", ".xml")
    #delImg = delImg1+delImg2
    #delImg = delImg1
    #new_delImg = tihuanhouozhui(delImg1)

    # for i in range(len(delImg)):
    #     os.remove(delImg[i])
    del_delImg = delete2("G:\\datasets\\coco\\split\\yes",".jpg",delImg1)
    desPath = "G:\\datasets\\coco\\final2"
    myCopyImg(del_delImg,desPath)