Python使用OPENCV的目标跟踪算法实现自动视频标注效果

先上效果

1.首先，要使用opencv的目标跟踪算法，必须要有opencv环境

使用：opencv==4.4.0 和 opencv-contrib-python==4.4.0.46，lxml 这三个环境包。

也可以使用以下方法进行下载：

pip install opencv-python==4.4.0
pip install opencv-contrib-python==4.4.0.4

pip install lxml

2.使用方法：

（1）：英文状态下的 “s” 是进行标注

（2）：使用小键盘 1-9 按下对应的标签序号，标签序号和标签可自定义（需要提前定义）

（3）：对目标进行绘制

（4）：按空格键继续

重复进行（1）（2）（3）（4）步骤，可实现多个目标的跟踪绘制

英文状态下的 “r” 是所有清除绘制

英文状态下的 “q” 是退出

当被跟踪目标丢失时，自动清除所有绘制

import cv2
import os
import time
from lxml import etree
 
#视频路径
vs = cv2.videocapture('peaple.avi')
#自定义标签
label = {1:"people",2:"car",3:"camera"}
#图片保存路径 ，一定使用要用绝对路径！！
imgpath = r"c:\users\bgt\desktop\opencv\img"
#xml保存路径 ，一定使用要用绝对路径！！
xmlpath = r"c:\users\bgt\desktop\opencv\xml"
#设置视频缩放
cv2.namedwindow("frame", 0)
#设置视频宽高
cv2.resizewindow("frame", 618, 416)
 
#定义生成xml类
class gen_annotations:
    def __init__(self, json_info):
        self.root = etree.element("annotation")
 
        child1 = etree.subelement(self.root, "folder")
        child1.text = str(json_info["pic_dirname"])
 
        child2 = etree.subelement(self.root, "filename")
        child2.text = str(json_info["filename"])
 
        child3 = etree.subelement(self.root, "path")
        child3.text = str(json_info["pic_path"])
 
        child4 = etree.subelement(self.root, "source")
 
        child5 = etree.subelement(child4, "database")
        child5.text = "my name is bgt"
 
    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, "depth")
        channeln.text = str(channel)
        segmented = etree.subelement(self.root, "segmented")
        segmented.text = "0"
 
    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, x0, y0, x1, y1):
        object = etree.subelement(self.root, "object")
        namen = etree.subelement(object, "name")
        namen.text = label
        pose = etree.subelement(object, "pose")
        pose.text = "unspecified"
        truncated = etree.subelement(object, "truncated")
        truncated.text = "0"
        difficult = etree.subelement(object, "difficult")
        difficult.text = "0"
        bndbox = etree.subelement(object, "bndbox")
        xminn = etree.subelement(bndbox, "xmin")
        xminn.text = str(x0)
        yminn = etree.subelement(bndbox, "ymin")
        yminn.text = str(y0)
        xmaxn = etree.subelement(bndbox, "xmax")
        xmaxn.text = str(x1)
        ymaxn = etree.subelement(bndbox, "ymax")
        ymaxn.text = str(y1)
 
 #定义生成xml的方法
def voc_opencv_xml(a,b,c,d,e,f,boxes,label,label_a,save="1.xml"):
    json_info = {}
    json_info["pic_dirname"] = a
    json_info["pic_path"] = b
    json_info["filename"] = c
    anno = gen_annotations(json_info)
 
    anno.set_size(d, e, f)
 
    for box in range(len(boxes)):
        x,y,w,h = [int(v) for v in boxes[box]]
        anno.add_pic_attr(label[label_a[box]],x,y,x+w,y+h)
    anno.savefile(save)
  
if __name__ == '__main__':
    label_a = []
    contents = os.path.split(imgpath)[1]
    trackers = cv2.multitracker_create()
    while true:
        filename_jpg = str(time.time()).split(".")[0] + "_" + str(time.time()).split(".")[1] + ".jpg"
        filename_xml = str(time.time()).split(".")[0] + "_" + str(time.time()).split(".")[1] + ".xml"
 
        path_filename_jpg = os.path.join(imgpath,filename_jpg)
        path_filename_xml = os.path.join(xmlpath,filename_xml)
 
        ret,frame = vs.read()
        if not ret:
            break
 
        success,boxes = trackers.update(frame)
        if len(boxes)>0:
            cv2.imwrite(path_filename_jpg, frame)
            judge = true
        else:
            judge = false
 
         if success==false:
            print("目标丢失")
            trackers = cv2.multitracker_create()
            label_a = []
            judge = false
        if judge:
            voc_opencv_xml(contents,filename_jpg,path_filename_jpg,frame.shape[1],frame.shape[0],frame.shape[2],boxes,label,label_a,path_filename_xml)
        if judge:
            for box in range(len(boxes)):
                x,y,w,h = [int(v) for v in boxes[box]]
                cv2.puttext(frame, label[label_a[box]], (x, y), cv2.font_hershey_complex, 1, (255, 255, 255), 1)
                cv2.rectangle(frame,(x,y),(x+w,y+h),(255,0,0),2)
  
        cv2.imshow('frame',frame)
 
        var = cv2.waitkey(30)
 
        if var == ord('s'):
            imgzi = cv2.puttext(frame, str(label), (50, 50), cv2.font_hershey_triplex, 1, (0, 255, 0), 2)
            cv2.imshow('frame', frame)
            var = cv2.waitkey(0)
            if var-48<len(label) or var-48<=len(label):
                label_a.append(int(var-48))
            box = cv2.selectroi("frame", frame, fromcenter=false,showcrosshair=true)
            tracker = cv2.trackercsrt_create()
            trackers.add(tracker,frame,box)
        elif var == ord("r"):
            trackers = cv2.multitracker_create()
            label_a = []
        elif var == ord('q'): #退出
            break
 
    vs.release()
    cv2.destroyallwindows()

3.得到xml和img数据是voc格式，img和xml文件以时间戳进行命名。防止同名覆盖。

4.最后使用 labelimg软件对获取到的img和xml进行最后的检查和微调

到此这篇关于python使用opencv的目标跟踪算法进自动视频标注效果的文章就介绍到这了,更多相关opencv目标跟踪自动视频标注内容请搜索www.887551.com以前的文章或继续浏览下面的相关文章希望大家以后多多支持www.887551.com！

黄山市民网：https://www.huangshanshimin.com/