41、labelimg数据集转dota以及rolabelimg、labelimg数据集旋转水平镜像 垂直镜像 和水平垂直镜像

基本思想：为了给BBAVectors-Oriented-Object-Detection 提供数据，所以记录一下操作，搞了旋转，但是旋转影响图片大小，所以先写个镜像方法，进行训练和使扩充数据集

一、labelimg转dota数据集 ，并做了顺时针排序处理 

import json
import os
from glob import glob
import argparse
import numpy as np
import shutil
from PIL import Image


# convert labelme json to DOTA txt format
# convert DOTA json to lableme txt format
def custombasename(fullname):
    return os.path.basename(os.path.splitext(fullname)[0])


def order_points_new(pts): # clock -https://zhuanlan.zhihu.com/p/10643062
    # sort the points based on their x-coordinates
    xSorted = pts[np.argsort(pts[:, 0]), :]

    # grab the left-most and right-most points from the sorted
    # x-roodinate points
    leftMost = xSorted[:2, :]
    rightMost = xSorted[2:, :]
    if leftMost[0,1]!=leftMost[1,1]:
        leftMost=leftMost[np.argsort(leftMost[:,1]),:]
    else:
        leftMost=leftMost[np.argsort(leftMost[:,0])[::-1],:]
    (tl, bl) = leftMost
    if rightMost[0,1]!=rightMost[1,1]:
        rightMost=rightMost[np.argsort(rightMost[:,1]),:]
    else:
        rightMost=rightMost[np.argsort(rightMost[:,0])[::-1],:]
    (tr,br)=rightMost
    # return the coordinates in top-left, top-right,
    # bottom-right, and bottom-left order
    return np.array([tl, tr, br, bl], dtype="float32")


parser = argparse.ArgumentParser(
    formatter_class=argparse.ArgumentDefaultsHelpFormatter
)
parser.add_argument('--input_dir', default=r'C:\Users\PHILIPS\Desktop\source',type=str,help='input annotated directory')
parser.add_argument('--output_dir', default=r'C:\Users\PHILIPS\Desktop\dest', type=str, help='output dataset directory')
parser.add_argument('--verify_dir', default=r'C:\Users\PHILIPS\Desktop\verify', type=str, help='verify dataset directory')
parser.add_argument('--verify', default=True,type=bool,help='verify')
parser.add_argument('--labels', default=r'labels.txt',type=str, help='labels annotated directory')
args = parser.parse_args()

if not os.path.exists(args.output_dir):
    os.makedirs(args.output_dir)

print('Creating dataset:', args.output_dir)


file_list = glob(os.path.join(args.input_dir , ".".join(["*","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

        # 输出 txt 文件的路径
        out_file = os.path.join(args.output_dir ,".".join([ custombasename(file_list[i]), 'txt']))
        # 写入 poly 四点坐标 和 label
        fout = open(out_file, 'w')
        out_str = ''
        #np.array(box, dtype="int")
        for shape_dict in label_dict['shapes']:
            points = shape_dict['points']
            item_points = []
            for p in points:
                item_points.append([p[0],p[1]])
            item_points=order_points_new(np.array(item_points, dtype="float"))
            for p in item_points.tolist():
                out_str += (str(p[0]) + ' ' + str(p[1]) + ' ')
            out_str += shape_dict['label'] + ' 0\n'
        fout.write(out_str)
        fout.close()
    print('%d/%d' % (i + 1, len(file_list)))
    print("labelme2dota...")
if args.verify:
    if not os.path.exists(args.verify_dir):
        os.makedirs(args.verify_dir)
    txt_list = glob(os.path.join(args.output_dir, ".".join(["*", "txt"])))
    for i in range(len(txt_list)):
        (filepath, tempfilename) = os.path.split(txt_list[i])
        (filename, extension) = os.path.splitext(tempfilename)
        sourcePath=None
        image_filename=None
        if os.path.exists(os.path.join(args.input_dir, ".".join([filename, "jpg"]))):
            sourcePath = os.path.join(args.input_dir, ".".join([filename, "jpg"]))
            image_filename=".".join([filename, "jpg"])
        elif os.path.exists(os.path.join(args.input_dir, ".".join([filename, "png"]))):
            sourcePath = os.path.join(args.input_dir, ".".join([filename, "png"]))
            image_filename = ".".join([filename, "png"])
        if sourcePath is None:
            print("check photo type")
            continue
        targetPath = args.verify_dir
        shutil.copy(sourcePath, targetPath)
        img= Image.open(sourcePath)
        imgSize = img.size  # 大小/尺寸
        w = img.width  # 图片的宽
        h = img.height  # 图片的高

        data = {}
        data['imagePath'] = image_filename
        data['flags'] = {}
        data['imageWidth'] = w
        data['imageHeight'] = h
        data['imageData'] = None
        data['version'] = "5.0.1"
        data["shapes"] = []

        with open(txt_list[i]) as f:
            label_str = f.readlines()
            for label_item in label_str:
                line_char = label_item.split("\n")[0].split(' ')
                points = [[eval（line_char[0]), eval（line_char[1])], [eval（line_char[2]), eval（line_char[3])],[eval（line_char[4]), eval（line_char[5])],[eval（line_char[6]), eval（line_char[7])]]
                itemData = {'points': []}
                itemData['points'].extend(points)
                itemData["flag"] = {}
                itemData["group_id"] = None
                itemData["shape_type"] = "polygon"
                itemData["label"] = line_char[-2]
                data["shapes"].append(itemData)

            jsonName = ".".join([filename, "json"])
            jsonPath = os.path.join(targetPath, jsonName)
            with open(jsonPath, "w") as f:
                json.dump(data, f)
            print(jsonName)
            print("dota2labelme...")

robelimg数据集扩充

import numpy as np
import os
import cv2
import math
from xml.dom.minidom import Document
from xml.etree.ElementTree import parse, Element

dir_path = r'G:\1\high_black_svg'
sv_path = r"G:\1\black_total"  # 修改后的xml文件存放路径
if not os.path.exists(sv_path):
    os.mkdir(sv_path)
list_mirror=[1,0,-1] #水平镜像 垂直镜像  水平垂直翻转   0
file_ext="jpg"
for mirror in list_mirror:
    for File in os.listdir(dir_path):
        if File.endswith('.xml'):
            print(File)
            xml_path = os.path.join(dir_path, File)
            dom = parse(xml_path)  ###最核心的部分,路径拼接,输入的是具体路径
            root = dom.getroot()
            folder_text=root.find("folder").text
            filename_text = root.find("filename").text
            filename_text,_=os.path.splitext(filename_text)
            xml_name_=".".join([filename_text+str(mirror),"xml"])
            save_path=os.path.join(sv_path,xml_name_)
            img_name_=".".join([filename_text+str(mirror) , file_ext])
            save_image = os.path.join(sv_path,img_name_ )
            path_text = root.find("path").text
            source_text = root.find("source").find("database").text
            width_text = root.find("size").find("width").text
            height_text = root.find("size").find("height").text
            depth_text = root.find("size").find("depth").text
            segmented_text=root.find("segmented").text
            (filename, extension) = os.path.splitext(File)
            img_jpg = ".".join([filename,file_ext])
            img_path = os.path.join(dir_path, img_jpg)
            image=cv2.imread(img_path)
            doc = Document()  # 创建DOM文档对象
            DOCUMENT = doc.createElement('annotation')  # 创建根元素

            folder = doc.createElement('folder')  ##建立自己的开头
            folder_text = doc.createTextNode(folder_text)  ##建立自己的文本信息
            folder.appendChild(folder_text)  ##自己的内容
            DOCUMENT.appendChild(folder)
            doc.appendChild(DOCUMENT)

            filename = doc.createElement('filename')
            filename_text = doc.createTextNode(filename_text+str(mirror))
            filename.appendChild(filename_text)
            DOCUMENT.appendChild(filename)
            doc.appendChild(DOCUMENT)

            path_ = doc.createElement('path')
            path_text = doc.createTextNode(img_name_)
            path_.appendChild(path_text)
            DOCUMENT.appendChild(path_)
            doc.appendChild(DOCUMENT)

            source = doc.createElement('source')
            database = doc.createElement('database')
            database_text = doc.createTextNode(source_text)  # 元素内容写入
            database.appendChild(database_text)
            source.appendChild(database)
            DOCUMENT.appendChild(source)
            doc.appendChild(DOCUMENT)

            size = doc.createElement('size')
            width = doc.createElement('width')
            width_text_ = doc.createTextNode(width_text)  # 元素内容写入
            width.appendChild(width_text_)
            size.appendChild(width)

            height = doc.createElement('height')
            height_text_ = doc.createTextNode(height_text)
            height.appendChild(height_text_)
            size.appendChild(height)

            depth = doc.createElement('depth')
            depth_text = doc.createTextNode(depth_text)
            depth.appendChild(depth_text)
            size.appendChild(depth)

            DOCUMENT.appendChild(size)

            segmented = doc.createElement('segmented')
            segmented_text = doc.createTextNode(segmented_text)
            segmented.appendChild(segmented_text)
            DOCUMENT.appendChild(segmented)
            doc.appendChild(DOCUMENT)


            for ano in root.iter('annotation'):  # 获取object节点中的name子节点

                for obj in ano.iter("object"):
                    type_text=obj.find('type').text
                    name_text = obj.find('name').text
                    pose_text = obj.find('pose').text
                    truncated_text = obj.find('truncated').text
                    difficult_text = obj.find('difficult').text

                    object = doc.createElement('object')
                    type = doc.createElement('type')
                    type_text = doc.createTextNode(type_text)
                    type.appendChild(type_text)
                    object.appendChild(type)

                    name = doc.createElement('name')
                    name_text = doc.createTextNode(name_text)
                    name.appendChild(name_text)
                    object.appendChild(name)

                    pose = doc.createElement('pose')
                    pose_text = doc.createTextNode(pose_text)
                    pose.appendChild(pose_text)
                    object.appendChild(pose)

                    truncated = doc.createElement('truncated')
                    truncated_text = doc.createTextNode(truncated_text)
                    truncated.appendChild(truncated_text)
                    object.appendChild(truncated)

                    difficult = doc.createElement('difficult')
                    difficult_text = doc.createTextNode(difficult_text)
                    difficult.appendChild(difficult_text)
                    object.appendChild(difficult)
                    robndbox = doc.createElement('robndbox')
                    cx_text = obj.find('robndbox').find('cx').text
                    cy_text = obj.find('robndbox').find('cy').text
                    w_text = obj.find('robndbox').find('w').text
                    h_text = obj.find('robndbox').find('h').text
                    angle_text = obj.find('robndbox').find('angle').text

                    image = cv2.flip(image, mirror)  # 水平翻转
                    #image2 = cv2.flip(image, 0)  # 垂直翻转
                    #image3 = cv2.flip(image, -1)  # 水平垂直翻转

                    #https://github.com/ChenCongGit/RoLabelImg_Transform/blob/master/txt_to_xml.py
                    if mirror==1 or  mirror==-1:
                        cx_ = doc.createElement('cx')
                        cx_text = doc.createTextNode(str(eval（width_text)-eval（cx_text)))
                        cx_.appendChild(cx_text)
                        robndbox.appendChild(cx_)
                    else:
                        cx_ = doc.createElement('cx')
                        cx_text = doc.createTextNode(str(cx_text))
                        cx_.appendChild(cx_text)
                        robndbox.appendChild(cx_)
                    if mirror == 0 or  mirror==-1:
                        cy_ = doc.createElement('cy')
                        cy_text = doc.createTextNode(str(eval（height_text)-eval（cy_text)))
                        cy_.appendChild(cy_text)
                        robndbox.appendChild(cy_)
                    else:
                        cy_ = doc.createElement('cy')
                        cy_text = doc.createTextNode(str(cy_text))
                        cy_.appendChild(cy_text)
                        robndbox.appendChild(cy_)

                    w_ = doc.createElement('w')
                    w_text = doc.createTextNode(str(w_text))
                    w_.appendChild(w_text)
                    robndbox.appendChild(w_)

                    h_ = doc.createElement('h')
                    h_text = doc.createTextNode(str(h_text))
                    h_.appendChild(h_text)
                    robndbox.appendChild(h_)

                    if mirror == 0 or mirror == 1:
                        angle_ = doc.createElement('angle')
                        angle_text = doc.createTextNode(str(np.pi-eval（angle_text)))
                        angle_.appendChild(angle_text)
                        robndbox.appendChild(angle_)
                        object.appendChild(robndbox)
                    else:
                        angle_ = doc.createElement('angle')
                        angle_text = doc.createTextNode(str(np.pi + eval（angle_text)))
                        angle_.appendChild(angle_text)
                        robndbox.appendChild(angle_)
                        object.appendChild(robndbox)


                    DOCUMENT.appendChild(object)


            ########### 将DOM对象doc写入文件
            with open(save_path, 'w') as f:
                 doc.writexml(f, indent='\t', newl='\n', addindent='\t', encoding='utf-8')
            cv2.imwrite(save_image,image)

labelimg数据的水平、垂直、水平垂直镜像翻转

import numpy as np
import os
import cv2
import math
from xml.dom.minidom import Document
from xml.etree.ElementTree import parse, Element

dir_path = r'G:\k\test'
sv_path = r"G:\k\mirror"  # 修改后的xml文件存放路径
if not os.path.exists(sv_path):
    os.mkdir(sv_path)
list_mirror = [1, 0, -1]  # 水平镜像 垂直镜像  水平垂直翻转   0
file_ext = "jpg"
for mirror in list_mirror:
    for File in os.listdir(dir_path):
        if File.endswith('.xml'):
            print(File)
            xml_path = os.path.join(dir_path, File)
            dom = parse(xml_path)  ###最核心的部分,路径拼接,输入的是具体路径
            root = dom.getroot()
            folder_text = root.find("folder").text
            filename_text = root.find("filename").text
            filename_text, _ = os.path.splitext(filename_text)
            if mirror==1:
               label_ext='mirror1'
            elif mirror==0:
                label_ext = 'mirror0'
            else:
                label_ext = 'mirror_1'
            xml_name_ = ".".join([filename_text + label_ext, "xml"])
            save_path = os.path.join(sv_path, xml_name_)
            img_name_ = ".".join([filename_text + label_ext, file_ext])
            save_image = os.path.join(sv_path, img_name_)
            path_text = root.find("path").text
            source_text = root.find("source").find("database").text
            width_text = root.find("size").find("width").text
            height_text = root.find("size").find("height").text
            depth_text = root.find("size").find("depth").text
            segmented_text = root.find("segmented").text
            (filename, extension) = os.path.splitext(File)
            img_jpg = ".".join([filename, file_ext])
            img_path = os.path.join(dir_path, img_jpg)
            image = cv2.imread(img_path)
            doc = Document()  # 创建DOM文档对象
            DOCUMENT = doc.createElement('annotation')  # 创建根元素

            folder = doc.createElement('folder')  ##建立自己的开头
            folder_text = doc.createTextNode(folder_text)  ##建立自己的文本信息
            folder.appendChild(folder_text)  ##自己的内容
            DOCUMENT.appendChild(folder)
            doc.appendChild(DOCUMENT)

            filename = doc.createElement('filename')
            filename_text = doc.createTextNode(filename_text +label_ext)
            filename.appendChild(filename_text)
            DOCUMENT.appendChild(filename)
            doc.appendChild(DOCUMENT)

            path_ = doc.createElement('path')
            path_text = doc.createTextNode(img_name_)
            path_.appendChild(path_text)
            DOCUMENT.appendChild(path_)
            doc.appendChild(DOCUMENT)

            source = doc.createElement('source')
            database = doc.createElement('database')
            database_text = doc.createTextNode(source_text)  # 元素内容写入
            database.appendChild(database_text)
            source.appendChild(database)
            DOCUMENT.appendChild(source)
            doc.appendChild(DOCUMENT)

            size = doc.createElement('size')
            width = doc.createElement('width')
            width_text_ = doc.createTextNode(width_text)  # 元素内容写入
            width.appendChild(width_text_)
            size.appendChild(width)

            height = doc.createElement('height')
            height_text_ = doc.createTextNode(height_text)
            height.appendChild(height_text_)
            size.appendChild(height)

            depth = doc.createElement('depth')
            depth_text = doc.createTextNode(depth_text)
            depth.appendChild(depth_text)
            size.appendChild(depth)

            DOCUMENT.appendChild(size)

            segmented = doc.createElement('segmented')
            segmented_text = doc.createTextNode(segmented_text)
            segmented.appendChild(segmented_text)
            DOCUMENT.appendChild(segmented)
            doc.appendChild(DOCUMENT)

            for ano in root.iter('annotation'):  # 获取object节点中的name子节点

                for obj in ano.iter("object"):
                    name_text = obj.find('name').text
                    pose_text = obj.find('pose').text
                    truncated_text = obj.find('truncated').text
                    difficult_text = obj.find('difficult').text

                    object = doc.createElement('object')
                    name = doc.createElement('name')
                    name_text = doc.createTextNode(name_text)
                    name.appendChild(name_text)
                    object.appendChild(name)

                    pose = doc.createElement('pose')
                    pose_text = doc.createTextNode(pose_text)
                    pose.appendChild(pose_text)
                    object.appendChild(pose)

                    truncated = doc.createElement('truncated')
                    truncated_text = doc.createTextNode(truncated_text)
                    truncated.appendChild(truncated_text)
                    object.appendChild(truncated)

                    difficult = doc.createElement('difficult')
                    difficult_text = doc.createTextNode(difficult_text)
                    difficult.appendChild(difficult_text)
                    object.appendChild(difficult)
                    bndbox = doc.createElement('bndbox')
                    xmin_text = obj.find('bndbox').find('xmin').text
                    ymin_text = obj.find('bndbox').find('ymin').text
                    xmax_text = obj.find('bndbox').find('xmax').text
                    ymax_text = obj.find('bndbox').find('ymax').text

                    image_flip = cv2.flip(image, mirror)
                    # image2 = cv2.flip(image, 1)  # 水平翻转
                    # image2 = cv2.flip(image, 0)  # 垂直翻转
                    # image3 = cv2.flip(image, -1)  # 水平垂直翻转

                    # https://github.com/ChenCongGit/RoLabelImg_Transform/blob/master/txt_to_xml.py
                    if mirror == 1 or mirror == -1:
                        xmin_ = doc.createElement('xmin')
                        xmin_text_ = doc.createTextNode(str(min(eval（width_text) - eval（xmin_text),eval（width_text) - eval（xmax_text))))
                        xmin_.appendChild(xmin_text_)
                        bndbox.appendChild(xmin_)
                        xmax_ = doc.createElement('xmax')
                        xmax_text_ = doc.createTextNode(str(max(eval（width_text) - eval（xmin_text),eval（width_text) - eval（xmax_text))))
                        xmax_.appendChild(xmax_text_)
                        bndbox.appendChild(xmax_)
                    else:
                        xmin_ = doc.createElement('xmin')
                        xmin_text_ = doc.createTextNode(str(min(eval（xmin_text),eval（xmax_text))))
                        xmin_.appendChild(xmin_text_)
                        bndbox.appendChild(xmin_)
                        xmax_ = doc.createElement('xmax')
                        xmax_text_ = doc.createTextNode(str(max(eval（xmin_text),eval（xmax_text))))
                        xmax_.appendChild(xmax_text_)
                        bndbox.appendChild(xmax_)

                    if mirror == 0 or mirror == -1:
                        ymin_ = doc.createElement('ymin')
                        ymin_text_ = doc.createTextNode(str(min(eval（height_text) - eval（ymin_text),eval（height_text) - eval（ymax_text))))
                        ymin_.appendChild(ymin_text_)
                        bndbox.appendChild(ymin_)
                        ymax_ = doc.createElement('ymax')
                        ymax_text_ = doc.createTextNode(str(max(eval（height_text) - eval（ymin_text),eval（height_text) - eval（ymax_text))))
                        ymax_.appendChild(ymax_text_)
                        bndbox.appendChild(ymax_)
                    else:
                        ymin_ = doc.createElement('ymin')
                        ymin_text_ = doc.createTextNode(str(min(eval（ymin_text),eval（ymax_text))))
                        ymin_.appendChild(ymin_text_)
                        bndbox.appendChild(ymin_)
                        ymax_ = doc.createElement('ymax')
                        ymax_text_ = doc.createTextNode(str(max(eval（ymin_text),eval（ymax_text))))
                        ymax_.appendChild(ymax_text_)
                        bndbox.appendChild(ymax_)
                    object.appendChild(bndbox)

                    DOCUMENT.appendChild(object)

            ########### 将DOM对象doc写入文件
            with open(save_path, 'w') as f:
                doc.writexml(f, indent='\t', newl='\n', addindent='\t', encoding='utf-8')
            cv2.imwrite(save_image, image_flip)