Python讀取VOC中的xml目標(biāo)框?qū)嵗?/h1>

發(fā)布時(shí)間：2020-10-16 18:27:19 來源：腳本之家 閱讀：190 作者：Peanut_范 欄目：開發(fā)技術(shù)

代碼：

#!/usr/bin/python
# -*- coding: UTF-8 -*-
# get annotation object bndbox location
import os
import cv2
try:
  import xml.etree.cElementTree as ET #解析xml的c語言版的模塊
except ImportError:
  import xml.etree.ElementTree as ET
											     
##get object annotation bndbox loc start 
def GetAnnotBoxLoc(AnotPath):#AnotPath VOC標(biāo)注文件路徑
  tree = ET.ElementTree(file=AnotPath) #打開文件，解析成一棵樹型結(jié)構(gòu)
  root = tree.getroot()#獲取樹型結(jié)構(gòu)的根
  ObjectSet=root.findall('object')#找到文件中所有含有object關(guān)鍵字的地方，這些地方含有標(biāo)注目標(biāo)
  ObjBndBoxSet={} #以目標(biāo)類別為關(guān)鍵字，目標(biāo)框?yàn)橹到M成的字典結(jié)構(gòu)
  for Object in ObjectSet:
    ObjName=Object.find('name').text
    BndBox=Object.find('bndbox')
    x1 = int(BndBox.find('xmin').text)#-1 #-1是因?yàn)槌绦蚴前?作為起始位置的
    y1 = int(BndBox.find('ymin').text)#-1
    x2 = int(BndBox.find('xmax').text)#-1
    y2 = int(BndBox.find('ymax').text)#-1
    BndBoxLoc=[x1,y1,x2,y2]
    if ObjName in ObjBndBoxSet:
    	ObjBndBoxSet[ObjName].append(BndBoxLoc)#如果字典結(jié)構(gòu)中含有這個(gè)類別了，那么這個(gè)目標(biāo)框要追加到其值的末尾
    else:
    	ObjBndBoxSet[ObjName]=[BndBoxLoc]#如果字典結(jié)構(gòu)中沒有這個(gè)類別，那么這個(gè)目標(biāo)框就直接賦值給其值吧
  return ObjBndBoxSet
##get object annotation bndbox loc end

def display(objBox,pic):
  img = cv2.imread(pic)
  
  for key in objBox.keys():
    for i in range(len(objBox[key])):
      cv2.rectangle(img, (objBox[key][i][0],objBox[key][i][1]), (objBox[key][i][2], objBox[key][i][3]), (0, 0, 255), 2)    
      cv2.putText(img, key, (objBox[key][i][0],objBox[key][i][1]), cv2.FONT_HERSHEY_COMPLEX, 1, (255,0,0), 1)
  cv2.imshow('img',img)
  cv2.imwrite('display.jpg',img)
  cv2.waitKey(0)


if __name__== '__main__':
  pic = r"./VOCdevkit/VOC2007/JPEGImages/000282.jpg"
  ObjBndBoxSet=GetAnnotBoxLoc(r"./VOCdevkit/VOC2007/Annotations/000282.xml")
  print(ObjBndBoxSet)
  display(ObjBndBoxSet,pic)

輸出結(jié)果：

{'chair': [[335, 263, 484, 373]], 'person': [[327, 104, 476, 300], [232, 57, 357, 374], [3, 32, 199, 374], [58, 139, 296, 374]]}

圖示：

import xml.etree.cElementTree as ET
import cv2
import os
import glob

def GetAnnotBoxLoc(AnotPath):
  tree = ET.ElementTree(file=AnotPath)
  root = tree.getroot()
  ObjectSet=root.findall('object')
  ObjBndBoxSet={} 
  for Object in ObjectSet:
    ObjName=Object.find('name').text
    BndBox=Object.find('bndbox')
    x1 = int(BndBox.find('xmin').text)
    y1 = int(BndBox.find('ymin').text)
    x2 = int(BndBox.find('xmax').text)
    y2 = int(BndBox.find('ymax').text)
    BndBoxLoc=[x1,y1,x2,y2]
    if ObjName in ObjBndBoxSet:
    	ObjBndBoxSet[ObjName].append(BndBoxLoc)
    else:
    	ObjBndBoxSet[ObjName]=[BndBoxLoc]
  return ObjBndBoxSet

def GetAnnotName(AnotPath):
  tree = ET.ElementTree(file=AnotPath) 
  root = tree.getroot()
  path=root.find('path').text
  return path

def Drawpic(xml_path,result_path):
  n = 0
  xmls = glob.glob(os.path.join(xml_path, '*.xml'))
  for xml in xmls:
    n = n + 1
    box=GetAnnotBoxLoc(xml)
    path=GetAnnotName(xml)
    img = cv2.imread(path)
    for classes in list(box.keys()):
      for boxes in box[classes]:
        if classes == "bad1":
          cv2.rectangle(img,(int(boxes[0]),int(boxes[1])),(int(boxes[2]),int(boxes[3])),(255,0,0),3) #blue
        if classes == "bad2":
          cv2.rectangle(img,(int(boxes[0]),int(boxes[1])),(int(boxes[2]),int(boxes[3])),(0,255,0),3) #green
        if classes == "bad3":
          cv2.rectangle(img,(int(boxes[0]),int(boxes[1])),(int(boxes[2]),int(boxes[3])),(0,0,255),3) #red
    cv2.imwrite(result_path+"/"+str(n)+"_result.jpg", img)
    print(path,"還原成功")

Drawpic("/home/wxy/Dashboard/dataset/VOCdevkit/VOC2012/Annotations","/home/wxy/Dashboard/dataset/VOCdevkit/VOC2012/test")