python目標(biāo)檢測(cè)給圖畫(huà)框,bbox畫(huà)到圖上并保存案例

更新時(shí)間：2020年03月10日 10:56:35 作者：別說(shuō)話寫(xiě)代碼

這篇文章主要介紹了python目標(biāo)檢測(cè)給圖畫(huà)框,bbox畫(huà)到圖上并保存案例，具有很好的參考價(jià)值，希望對(duì)大家有所幫助。一起跟隨小編過(guò)來(lái)看看吧

我就廢話不多說(shuō)了，還是直接上代碼吧！

import os
import xml.dom.minidom
import cv2 as cv
 
ImgPath = 'C:/Users/49691/Desktop/gangjin/gangjin_test/JPEGImages/'
AnnoPath = 'C:/Users/49691/Desktop/gangjin/gangjin_test/Annotations/' #xml文件地址
save_path = ''
def draw_anchor(ImgPath,AnnoPath,save_path):
  imagelist = os.listdir(ImgPath)
  for image in imagelist:
 
    image_pre, ext = os.path.splitext(image)
    imgfile = ImgPath + image
    xmlfile = AnnoPath + image_pre + '.xml'
    # print(image)
    # 打開(kāi)xml文檔
    DOMTree = xml.dom.minidom.parse(xmlfile)
    # 得到文檔元素對(duì)象
    collection = DOMTree.documentElement
    # 讀取圖片
    img = cv.imread(imgfile)
 
    filenamelist = collection.getElementsByTagName("filename")
    filename = filenamelist[0].childNodes[0].data
    print(filename)
    # 得到標(biāo)簽名為object的信息
    objectlist = collection.getElementsByTagName("object")
 
    for objects in objectlist:
      # 每個(gè)object中得到子標(biāo)簽名為name的信息
      namelist = objects.getElementsByTagName('name')
      # 通過(guò)此語(yǔ)句得到具體的某個(gè)name的值
      objectname = namelist[0].childNodes[0].data
 
      bndbox = objects.getElementsByTagName('bndbox')
      # print(bndbox)
      for box in bndbox:
        x1_list = box.getElementsByTagName('xmin')
        x1 = int(x1_list[0].childNodes[0].data)
        y1_list = box.getElementsByTagName('ymin')
        y1 = int(y1_list[0].childNodes[0].data)
        x2_list = box.getElementsByTagName('xmax')  #注意坐標(biāo)，看是否需要轉(zhuǎn)換
        x2 = int(x2_list[0].childNodes[0].data)
        y2_list = box.getElementsByTagName('ymax')
        y2 = int(y2_list[0].childNodes[0].data)
        cv.rectangle(img, (x1, y1), (x2, y2), (255, 255, 255), thickness=2)
        cv.putText(img, objectname, (x1, y1), cv.FONT_HERSHEY_COMPLEX, 0.7, (0, 255, 0),
              thickness=2)
        # cv.imshow('head', img)
        cv.imwrite(save_path+'/'+filename, img)  #save picture

補(bǔ)充知識(shí)：深度學(xué)習(xí)python之用Faster-rcnn 檢測(cè)結(jié)果（txt文件）在原圖畫(huà)出box

使用Faster-rcnn 的test_net.py 檢測(cè)網(wǎng)絡(luò)的mAP等精度會(huì)生成一個(gè)檢測(cè)結(jié)果（txt文件），格式如下：

000004 0.972 302.8 94.5 512.0 150.0
000004 0.950 348.1 166.1 512.0 242.9
000004 0.875 1.0 25.7 292.6 126.3
000004 0.730 1.0 138.5 488.3 230.0
000004 0.699 1.0 120.9 145.5 139.9
000004 0.592 54.4 227.4 431.9 343.4
000004 0.588 1.0 159.8 18.8 231.6
000004 0.126 1.0 247.1 342.3 270.0
000004 0.120 1.0 225.4 185.7 309.3

每行分別為名稱檢測(cè)概率 xmin ymin xmax ymax

問(wèn)題在于每一行只顯示一個(gè)box數(shù)據(jù)，每幅圖像可能包括多個(gè)box，需要判斷提取的多行數(shù)據(jù)是不是屬于同一圖片

下面使用python提取這些數(shù)據(jù)，在原圖上畫(huà)出box并且保存起來(lái)

import os
import os.path
import numpy as np
import xml.etree.ElementTree as xmlET
from PIL import Image, ImageDraw
import cPickle as pickle 

txt_name = 'comp4_8a226fd7-753d-40fc-8013-f68d2a465579_det_test_ship.txt'
file_path_img = '/home/JPEGImages'
save_file_path = '/home/detect_results'


source_file = open(txt_name)

img_names = []
for line in source_file:
  staff = line.split()
  img_name = staff[0]
  img_names.append(img_name)

name_dict = {}
for i in img_names:
  if img_names.count(i)>0:
    name_dict[i] = img_names.count(i) 

source_file.close()

source_file = open(txt_name)
for idx in name_dict:
  img = Image.open(os.path.join(file_path_img, idx + '.jpg')) 
  draw = ImageDraw.Draw(img)
  for i in xrange(name_dict[idx]):
    line = source_file.readline()
    staff = line.split()
    score = staff[1]
    box = staff[2:6]
    draw.rectangle([int(np.round(float(box[0]))), int(np.round(float(box[1]))), 
          int(np.round(float(box[2]))), int(np.round(float(box[3])))], outline=(255, 0, 0))
  img.save(os.path.join(save_file_path, idx + '.jpg')) 

source_file.close()

運(yùn)行完即可在保存文件夾中得到效果圖。

以上這篇python目標(biāo)檢測(cè)給圖畫(huà)框,bbox畫(huà)到圖上并保存案例就是小編分享給大家的全部?jī)?nèi)容了，希望能給大家一個(gè)參考，也希望大家多多支持腳本之家。

您可能感興趣的文章: