VOC数据集验证

annotation_overlay_voc_xml.py

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

import os
import xml.etree.ElementTree as ET
from skimage import io, draw

xml_folder = "/home/haoyu/Annotations"
image_dir = "/home/haoyu/JPEGImages"
for parent, dirnames, filenames in os.walk(xml_folder):
    for filename in filenames:
        tree = ET.parse(os.path.join(xml_folder, filename))
        root = tree.getroot()
        img_filename = root.find(".//filename").text
        img = io.imread(os.path.join(image_dir, img_filename))
        print(img_filename)
        for box in root.findall('.//bndbox'):
            x_min = int(box.find('.//xmin').text)
            y_min = int(box.find('.//ymin').text)
            x_max = int(box.find('.//xmax').text)
            y_max = int(box.find('.//ymax').text)
            rr, cc = draw.polygon_perimeter(
                [y_min, y_max, y_max, y_min],
                [x_min, x_min, x_max, x_max]
            )
            draw.set_color(img, (rr, cc), (0, 0, 255))
        io.imshow(img)
        io.show()

pascal_voc_io.py

# -*- coding: utf8 -*-
from xml.etree import ElementTree
from xml.etree.ElementTree import Element, SubElement
from lxml import etree
import codecs

XML_EXT = '.xml'


class PascalVocWriter:
    def __init__(self, foldername, filename, imgSize, databaseSrc='Unknown', localImgPath=None):
        self.foldername = foldername
        self.filename = filename
        self.databaseSrc = databaseSrc
        self.imgSize = imgSize
        self.boxlist = []
        self.localImgPath = localImgPath
        self.verified = False

    def prettify(self, elem):
        """
            Return a pretty-printed XML string for the Element.
        """
        rough_string = ElementTree.tostring(elem, 'utf8')
        root = etree.fromstring(rough_string)
        return etree.tostring(root, pretty_print=True)

    def genXML(self):
        """
            Return XML root
        """
        # Check conditions
        if self.filename is None or \
                        self.foldername is None or \
                        self.imgSize is None:
            return None

        top = Element('annotation')
        # top.set('verified', 'yes' if self.verified else 'no')

        folder = SubElement(top, 'folder')
        folder.text = self.foldername

        filename = SubElement(top, 'filename')
        filename.text = self.filename

        localImgPath = SubElement(top, 'path')
        localImgPath.text = self.localImgPath

        source = SubElement(top, 'source')
        database = SubElement(source, 'database')
        database.text = self.databaseSrc

        size_part = SubElement(top, 'size')
        width = SubElement(size_part, 'width')
        height = SubElement(size_part, 'height')
        depth = SubElement(size_part, 'depth')
        width.text = str(self.imgSize[1])
        height.text = str(self.imgSize[0])
        if len(self.imgSize) == 3:
            depth.text = str(self.imgSize[2])
        else:
            depth.text = '1'

        segmented = SubElement(top, 'segmented')
        segmented.text = '0'
        return top

    def addBndBox(self, xmin, ymin, xmax, ymax, name):
        bndbox = {'xmin': xmin, 'ymin': ymin, 'xmax': xmax, 'ymax': ymax}
        bndbox['name'] = name
        self.boxlist.append(bndbox)

    def appendObjects(self, top):
        for each_object in self.boxlist:
            object_item = SubElement(top, 'object')
            name = SubElement(object_item, 'name')
            try:
                name.text = unicode(each_object['name'])
            except NameError:
                # Py3: NameError: name 'unicode' is not defined
                name.text = each_object['name']
            pose = SubElement(object_item, 'pose')
            pose.text = "Unspecified"
            truncated = SubElement(object_item, 'truncated')
            truncated.text = "0"
            difficult = SubElement(object_item, 'difficult')
            difficult.text = "0"
            bndbox = SubElement(object_item, 'bndbox')
            xmin = SubElement(bndbox, 'xmin')
            xmin.text = str(each_object['xmin'])
            ymin = SubElement(bndbox, 'ymin')
            ymin.text = str(each_object['ymin'])
            xmax = SubElement(bndbox, 'xmax')
            xmax.text = str(each_object['xmax'])
            ymax = SubElement(bndbox, 'ymax')
            ymax.text = str(each_object['ymax'])

    def save(self, targetFile=None):
        root = self.genXML()
        self.appendObjects(root)
        out_file = None
        if targetFile is None:
            out_file = codecs.open(
                self.filename + XML_EXT, 'w', encoding='utf-8')
        else:
            out_file = codecs.open(targetFile, 'w', encoding='utf-8')

        prettifyResult = self.prettify(root)
        out_file.write(prettifyResult.decode('utf8'))
        out_file.close()


class PascalVocReader:
    def __init__(self, filepath):
        # shapes type:
        # [labbel, [(x1,y1), (x2,y2), (x3,y3), (x4,y4)], color, color]
        self.shapes = []
        self.filepath = filepath
        self.verified = False
        self.parseXML()

    def getShapes(self):
        return self.shapes

    def addShape(self, label, bndbox):
        xmin = int(bndbox.find('xmin').text)
        ymin = int(bndbox.find('ymin').text)
        xmax = int(bndbox.find('xmax').text)
        ymax = int(bndbox.find('ymax').text)
        points = [(xmin, ymin), (xmax, ymin), (xmax, ymax), (xmin, ymax)]
        self.shapes.append((label, points, None, None))

    def parseXML(self):
        assert self.filepath.endswith('.xml'), "Unsupport file format"
        parser = etree.XMLParser(encoding='utf-8')
        xmltree = ElementTree.parse(self.filepath, parser=parser).getroot()
        filename = xmltree.find('filename').text
        # try:
        #     verified = xmltree.attrib['verified']
        #     if verified == 'yes':
        #         self.verified = True
        # except KeyError:
        #     self.verified = False

        for object_iter in xmltree.findall('object'):
            bndbox = object_iter.find("bndbox")
            label = object_iter.find('name').text
            self.addShape(label, bndbox)
        return True


# tempParseReader = PascalVocReader('test.xml')
# print tempParseReader.getShapes()
"""
# Test
tmp = PascalVocWriter('temp','test', (10,20,3))
tmp.addBndBox(10,10,20,30,'chair')
tmp.addBndBox(1,1,600,600,'car')
tmp.save()
"""

range_check.py

# -*- coding: utf-8 -*-
import os
import xml.etree.ElementTree as ET


def get_filenames(file_dir):
    filenames = os.listdir(file_dir)
    filenames.sort()
    return filenames


if __name__ == "__main__":
    file_dir = "/home/haoyu/Annotations"
    for fnm in get_filenames(file_dir):
        tree = ET.parse(os.path.join(file_dir, fnm))
        root = tree.getroot()
        for bndbox in root.findall('./object/bndbox'):
            xmin = int(bndbox.find('xmin').text)
            xmax = int(bndbox.find('xmax').text)
            ymin = int(bndbox.find('ymin').text)
            ymax = int(bndbox.find('ymax').text)
            try:
                assert (xmin > 0)
                assert (ymin > 0)
                assert (xmin < xmax)
                assert (ymin < ymax)
                assert (xmax <= 640)
                assert (ymax <= 480)
            except:
                print(xmin, xmax, ymin, ymax)
                print(os.path.join(file_dir, fnm))

voc_annotation_correct.py

import os
from pascal_voc_io import PascalVocReader, PascalVocWriter

if '__main__' == __name__:
    base_dir = './Annotations'
    for parent, dirnames, filenames in os.walk(base_dir):
        for filename in filenames:
            fnm, fex = os.path.splitext(filename)
            if '.xml' == fex:
                reader = PascalVocReader(os.path.join(base_dir, filename))
                writer = PascalVocWriter('VOC2007', fnm, [480, 640, 3],
                                         localImgPath=os.path.join('JPEGImages', fnm + '.png'))
                print(reader.shapes)
                for label, pos, _, __ in reader.shapes:
                    # (xmin, ymin), (xmax, ymin), (xmax, ymax), (xmin, ymax)
                    # xmin, ymin, xmax, ymax
                    writer.addBndBox(pos[0][0], pos[0][1], pos[2][0], pos[2][1], label)
                writer.save(targetFile=os.path.join("Anno2", filename))