voc生成xml 代码

目录

python生成voc xml文件

lxml windows安装

读取示例

可视化

voc转换为json格式：

生成示例

上面是代码，下面有调用示例

api调用代码，其实只有几行：

---

python生成voc xml文件

python生成voc xml文件_jacke121的专栏-CSDN博客

lxml windows安装

pip install lxml可能失败，

一种可行方法，在这里下载对应版本whl，然后安装，亲测OK。

lxml · PyPI

读取示例

import os
import cv2
import numpy as np
import xml.etree.ElementTree as ET
CLASSES = ('m',)
_class_to_ind=dict(list(zip(CLASSES, list(range(len(CLASSES))))))
def get_label(label_path,shape, flip=False):
    h, w, c = shape
    annotations = np.zeros((0, 5),dtype=np.float32)
    if os.path.exists(label_path):
        tree = ET.parse(label_path)
        objs = tree.findall('object')
        for ix, obj in enumerate(objs):

            if obj.find('name').text.lower().strip() not in _class_to_ind:
                # logging.info("label not select: {}".format(obj.find('name').text.lower().strip()))
                continue
            bbox = obj.find('bndbox')
            x1 = max(float(bbox.find('xmin').text), 1)  # - 1
            y1 = max(float(bbox.find('ymin').text), 1)  # - 1
            x2 = min(float(bbox.find('xmax').text), 10000)  # - 1
            y2 = min(float(bbox.find('ymax').text), 10000)  # - 1

            if obj.find("difficult") is not None and obj.find("difficult").text == '1':
                if (x2-x1) * (y2-y1)< 10*10:
                    continue

可视化

import os
import cv2
import re

import xml.etree.ElementTree as ET

CLASSES = ('car',)
_class_to_ind=dict(list(zip(CLASSES, list(range(len(CLASSES))))))

if __name__ == '__main__':
    image_dir = r'G:\img'
    xml_dir = r'G:\xml'
    image_list = os.listdir(image_dir)
    cnt = 0
    for i in  image_list:
        image_path = os.path.join(image_dir ,i)
        xml_path = os.path.join(xml_dir ,os.path.splitext(i)[0 ] +'.xml')

        tree = ET.parse(xml_path)
        objs = tree.findall('object')
        image = cv2.imread(image_path)
        for ix, obj in enumerate(objs):

            if obj.find('name').text.lower().strip() not in _class_to_ind:
                # logging.info("label not select: {}".format(obj.find('name').text.lower().strip()))
                continue
            bbox = obj.find('bndbox')
            x1 = max(float(bbox.find('xmin').text), 1)  # - 1
            y1 = max(float(bbox.find('ymin').text), 1)  # - 1
            x2 = min(float(bbox.find('xmax').text), 10000)  # - 1
            y2 = min(float(bbox.find('ymax').text), 10000)  # - 1
            cv2.rectangle(image, (int(x1), int(y1)), (int(x2), int(y2)), (255, 0, 0), thickness=2)
            # cv2.putText(image, info[0], (info[1], info[2]), cv2.FONT_HERSHEY_PLAIN, 0.5, (0, 0, 255), 1)
        cv2.imshow("image",image)
        cv2.waitKey()

voc转换为json格式：

python生成coco格式_AI视觉网奇的博客-CSDN博客_coco python

生成示例

上面是代码，下面有调用示例

api调用代码，其实只有几行：

        anno = GEN_Annotations(pic_path)
        anno.set_size(new_width, new_height, 3)
        for (x, y, w, h,name) in bbox_list:
            anno.add_pic_attr(name, x, y, w, h)

        save_xml_path=xml_path.replace(base_path,save_pic_path)
        os.makedirs(os.path.dirname(save_xml_path), exist_ok=True)
        anno.savefile(save_xml_path)

# -*- coding: utf-8 -*-

import os


from lxml import etree

class GEN_Annotations:
    def __init__(self, filename):
        self.root = etree.Element("annotation")

        child1 = etree.SubElement(self.root, "folder")
        child1.text = "VOC2007"

        child2 = etree.SubElement(self.root, "filename")
        child2.text = filename

        child3 = etree.SubElement(self.root, "source")
        # child2.set("database", "The VOC2007 Database")
        child4 = etree.SubElement(child3, "annotation")
        child4.text = "PASCAL VOC2007"
        child5 = etree.SubElement(child3, "database")

        child6 = etree.SubElement(child3, "image")
        child6.text = "flickr"
        child7 = etree.SubElement(child3, "flickrid")
        child7.text = "35435"

        # root.append( etree.Element("child1") )
        # root.append( etree.Element("child1", interesting="totally"))
        # child2 = etree.SubElement(root, "child2")

        # child3 = etree.SubElement(root, "child3")
        # root.insert(0, etree.Element("child0"))

    def set_size(self,witdh,height,channel):
        size = etree.SubElement(self.root, "size")
        widthn = etree.SubElement(size, "width")
        widthn.text = str(witdh)
        heightn = etree.SubElement(size, "height")
        heightn.text = str(height)
        channeln = etree.SubElement(size, "channel")
        channeln.text = str(channel)
    def savefile(self,filename):
        tree = etree.ElementTree(self.root)
        tree.write(filename, pretty_print=True, xml_declaration=False, encoding='utf-8')
    def add_pic_attr(self,label,x,y,w,h,diffi_=0):
        object = etree.SubElement(self.root, "object")
        namen = etree.SubElement(object, "name")
        namen.text = label
        diffi = etree.SubElement(object, "difficult")
        diffi.text = str(diffi_)

        truncated_ = etree.SubElement(object, "truncated")
        truncated_.text = str(0)

        bndbox = etree.SubElement(object, "bndbox")
        xminn = etree.SubElement(bndbox, "xmin")
        xminn.text = str(x)
        yminn = etree.SubElement(bndbox, "ymin")
        yminn.text = str(y)
        xmaxn = etree.SubElement(bndbox, "xmax")
        xmaxn.text = str(x+w)
        ymaxn = etree.SubElement(bndbox, "ymax")
        ymaxn.text = str(y+h)

调用：

import cv2
import xml.etree.ElementTree as ET

def change_size(pic_path, xml_path,base_path,save_pic_path):

    img = cv2.imread(pic_path)
    if img is None:
        print('img is null', pic_path)
        return
    if img.shape[1] < 1 or img.shape[0] < 1:
        print('img is null', pic_path)
        return
    height,width=img.shape[:2]


    new_width = width
    new_height = height
    if height>700 and width>1200:
        new_width = int(width * 0.8)
        new_height = int(height * 0.8)
        img=cv2.resize(img,(new_width,new_height))
    if new_height!=576 or new_width!=1024:
        print(new_height,new_width,pic_path)
    pic_path = pic_path.replace(base_path, save_pic_path)
    os.makedirs(os.path.dirname(pic_path), exist_ok=True)
    cv2.imwrite(pic_path, img)
    if os.path.exists(xml_path):
        tree = ET.parse(xml_path)
        # file_objs = tree.findall('filename')

        objs = tree.findall('object')

        objs_diffi = tree.findall('object/difficult')
        if len(objs_diffi)==0:
            print(xml_path, objs_diffi, len(objs_diffi))

        bbox_list = []
        for ix, obj in enumerate(objs):
            name = obj.find('name').text
            bbox = obj.find('bndbox')
            x1 = int(int(bbox.find('xmin').text)*0.8)
            y1 = int(int(bbox.find('ymin').text)*0.8)
            x2 = int(int(bbox.find('xmax').text)*0.8)
            y2 = int(int(bbox.find('ymax').text)*0.8)
            bbox_list.append((x1,y1,(x2-x1),(y2-y1),name))


        anno = GEN_Annotations(pic_path)
        anno.set_size(new_width, new_height, 3)
        for (x, y, w, h,name) in bbox_list:
            anno.add_pic_attr(name, x, y, w, h)

        save_xml_path=xml_path.replace(base_path,save_pic_path)
        os.makedirs(os.path.dirname(save_xml_path), exist_ok=True)
        anno.savefile(save_xml_path)



if __name__ == '__main__':
    # save file path

    source_base_path = r'D:\Team-CV\dataset\chumao_train/'
    save_base_path = r'D:\Team-CV\dataset\chumao_train_1024/'

    os.makedirs(save_base_path, exist_ok=True)
    # file path

    g = os.walk(source_base_path)
    pic_files = ['%s\\%s' % (i[0], j) for i in g if i[0].endswith('JPEGImages') for j in i[-1] if j.endswith('jpg')]

    for i,pic_file in enumerate(pic_files):
        if 'JPEGImages' in pic_file:
            xml_file = pic_file.replace('JPEGImages','Annotations').replace('.jpg','.xml')
            change_size(pic_file, xml_file,source_base_path,save_base_path)
            print(i,pic_file)