两款IP Camera+YOLOV3进行目标检测（手机摄像头作为电脑摄像头使用）

关注

原创最新推荐文章于 2026-10-19 11:21:15 发布 1.9k 阅读 AI 写同款· GEO 优化 ›

多摄像头/跨境头多目标跟踪的简单实现前提：在多个摄像头分别位于完全不同的位置进行拍摄这个前提下，我们可以采用简单的将不同摄像头画面进行逐帧拼接再进行目标检测，追踪来实现这个要求。因为我们对视频、摄像头进行目标检测、每个对象添加id并不是对视频这个整体一次性地进行处理，而是将视频进行while循环来做到逐帧提取图片，再对单张图片进行画面内的目标检测，在利用多目标跟踪技术进行单个对象的下一帧位置预测等等来保证单个对象id不变。解决方案：所以我们同时可以从输入多个视频源（或开启多个摄像头），在while循环前 cap1 = cv2.Vide 阅读详情

DarkNet的编译及安装的过程（无GPU的情况详解）

YOLO-V3实时检测实现（opencv+python实现）——改进——＞更加的易懂

YOLO-V3实时检测实现（opencv+python实现）

Opencv+YOLO-V3实现目标跟踪

提示：下面在连接IP Camera的时候，需要电脑和手机的网都在同一个网段（同一个热点或者WIFI）。

1.第一款IP Camera（一）

（1）.IP Camera下载

https://www.pc6.com/az/510980.html

（2）.IP Camera使用

（3）.测试程序

提示：如果读者在以下程序中测试可以打开摄像头，则可以进行后序的步骤啦。

"""
@Author : Keep_Trying_Go
@Major  : Computer Science and Technology
@Hobby  : Computer Vision
@Time   : 2023/5/21 10:27
"""

import cv2  # 导入库
import threading

def run_video():
    cv2.namedWindow("camera", 1)  # 定义启动窗口名称
    video = "http://admin:admin@192.168.236.172:8081/video"
    # video = "http://192.168.43.1:8080/video"
    # 此处根据IP摄像头生成的局域网地址
    capture = cv2.VideoCapture(video)
    while True:
        success, img = capture.read()  # 读取视频
        img = cv2.flip(img, 1)
        cv2.imshow("camera", img)
        key = cv2.waitKey(10)
        if key == 27:  # esc键退出
            break

    capture.release()
    cv2.destroyWindow("camera")

if __name__ == '__main__':
    thread = threading.Thread(target=run_video)
    thread.start()

（4）.YOLOV3进行物体检测

提示：关于YOLOV3的权重文件和cfg配置文件下载：https://pjreddie.com/darknet/yolo/

"""
@Author : Keep_Trying_Go
@Major  : Computer Science and Technology
@Hobby  : Computer Vision
@Time   : 2023/5/19 16:33
"""

import os
import time
import cv2
import cvzone
import numpy as np

#读取YOLOV3-tiny-weight和.cfg文件
net = cv2.dnn.readNet(model='yolov3-tiny.weights',config='yolov3-tiny.cfg')
#设置置信度阈值和非极大值抑制的阈值
Confidence_thresh=0.2
Nms_thresh=0.35
#读取coco.names文件中的类别
with open('coco.names','r') as fp:
    classes=fp.read().splitlines()

def yolov_detect(frame):
    """
    :param frame:
    :return:
    """
    #获取网络模型
    model = cv2.dnn_DetectionModel(net)
    #设置网络的输入
    model.setInputParams(scale = 1 / 255, size=(416,416))
    #进行预测
    class_id,score,boxes = model.detect(
        frame,confThreshold=Confidence_thresh,nmsThreshold=Nms_thresh
    )
    return class_id,score,boxes

#进行实时检测
def detect_time():
    video = "http://admin:admin@192.168.236.172:8081/video"
    cap = cv2.VideoCapture(video)

    while cap.isOpened():
        ret,frame = cap.read()
        if ret is False:
            break
        frame = cv2.flip(src = frame,flipCode=2)
        #输出预测的结果
        start_time = time.time()
        class_ids,scores,boxes = yolov_detect(frame)
        #绘制矩形框
        for (class_id,box) in enumerate(boxes):
            (x,y,w,h) = box
            class_name = classes[class_id]
            confidence = scores[class_id]
            confidence = str(round(confidence,2))
            cv2.rectangle(img = frame,pt1=(x,y),pt2=(x + w, y + h),
                          color = (0,255,0),thickness = 2
                          )
            text = class_name + ' '+ confidence
            cvzone.putTextRect(
                img=frame, text=text, pos=(x, y - 10), font=
                cv2.FONT_HERSHEY_SIMPLEX,scale=1.0,colorR=(255,0,255),
                thickness=2
            )
            # cv2.putText(img = frame,text = text,org=(x,y - 10), fontFace=
            #             cv2.FONT_HERSHEY_SIMPLEX,fontScale=1.0,color=(0,255,0),
            #             thickness=2)
        # 计算结束时间
        end_time = time.time()
        FPS = round(1 / (end_time - start_time), 0)
        cv2.putText(img=frame, text='FPS: ' + str(FPS), org=(10, 50),
                    fontFace=cv2.FONT_HERSHEY_SIMPLEX,
                    fontScale=1.0, color=(0, 255, 0), thickness=2)
        cv2.imshow('detect',frame)
        key = cv2.waitKey(1)
        if key == 27:
            break
    cap.release()
    cv2.destroyAllWindows()

#单张图片的检测
def signal_detect(image_path='images/face1.jpg'):
    frame=cv2.imread(image_path)
    frame = cv2.resize(src=frame, dsize=(416, 416))
    # 进行预测
    class_ids, scores, boxes = yolov_detect(frame)
    # 绘制矩形框
    for (class_id, box) in enumerate(boxes):
        (x, y, w, h) = box
        class_name = classes[class_ids[class_id]]
        confidence = scores[class_ids[class_id]]
        confidence = str(round(confidence, 2))
        cv2.rectangle(img=frame, pt1=(x, y), pt2=(x + w, y + h),
                      color=(0, 255, 0), thickness=2)
        text = class_name + ' ' + confidence
        cvzone.putTextRect(
            img=frame, text=text, pos=(x, y - 10), font=
            cv2.FONT_HERSHEY_SIMPLEX, scale=1.0, colorR=(255, 0, 255),
            thickness=2
        )
        # cv2.putText(img=frame, text=text,
        #             org=(x, y - 10), fontFace=cv2.FONT_HERSHEY_SIMPLEX,
        #             fontScale=1.0, color=(0, 255, 0), thickness=2)

    cv2.imshow('detect', frame)
    cv2.waitKey(0)
    cv2.destroyAllWindows()

if __name__ == '__main__':
    detect_time()

2.第二款IP Camera（二）

（1）下载地址

提示：下载专业版，不然好像运行不了（我这个上面是这样的）http://m.itmop.com/downinfo/431802.html（安卓手机上安装）

（2）使用方式

提示：可以使用上面的ip地址和端口即可在浏览器中进行访问：http://192.168.*.*:8080/

访问的界面可以看到如下：

提示：关于里面更多的操作，读者可以自己去摸索。

（3）程序进行测试

"""
@Author : Keep_Trying_Go
@Major  : Computer Science and Technology
@Hobby  : Computer Vision
@Time   : 2023/5/19 16:33
"""

import os
import time
import cv2
import cvzone
import numpy as np

#读取YOLOV3-tiny-weight和.cfg文件
net = cv2.dnn.readNet(model='yolov3-tiny.weights',config='yolov3-tiny.cfg')
#设置置信度阈值和非极大值抑制的阈值
Confidence_thresh=0.2
Nms_thresh=0.35
#读取coco.names文件中的类别
with open('coco.names','r') as fp:
    classes=fp.read().splitlines()

def yolov_detect(frame):
    """
    :param frame:
    :return:
    """
    #获取网络模型
    model = cv2.dnn_DetectionModel(net)
    #设置网络的输入
    model.setInputParams(scale = 1 / 255, size=(416,416))
    #进行预测
    class_id,score,boxes = model.detect(
        frame,confThreshold=Confidence_thresh,nmsThreshold=Nms_thresh
    )
    return class_id,score,boxes

#进行实时检测
def detect_time():
    # video = "http://admin:admin@192.168.236.172:8081/video"
    video = "http://admin:admin@192.168.236.172:8080/video"
    cap = cv2.VideoCapture(video)

    while cap.isOpened():
        ret,frame = cap.read()
        if ret is False:
            break
        frame = cv2.flip(src = frame,flipCode=2)
        #输出预测的结果
        start_time = time.time()
        class_ids,scores,boxes = yolov_detect(frame)
        #绘制矩形框
        for (class_id,box) in enumerate(boxes):
            (x,y,w,h) = box
            class_name = classes[class_id]
            confidence = scores[class_id]
            confidence = str(round(confidence,2))
            cv2.rectangle(img = frame,pt1=(x,y),pt2=(x + w, y + h),
                          color = (0,255,0),thickness = 2
                          )
            text = class_name + ' '+ confidence
            cvzone.putTextRect(
                img=frame, text=text, pos=(x, y - 10), font=
                cv2.FONT_HERSHEY_SIMPLEX,scale=1.0,colorR=(255,0,255),
                thickness=2
            )
            # cv2.putText(img = frame,text = text,org=(x,y - 10), fontFace=
            #             cv2.FONT_HERSHEY_SIMPLEX,fontScale=1.0,color=(0,255,0),
            #             thickness=2)
        # 计算结束时间
        end_time = time.time()
        FPS = round(1 / (end_time - start_time), 0)
        cv2.putText(img=frame, text='FPS: ' + str(FPS), org=(10, 50),
                    fontFace=cv2.FONT_HERSHEY_SIMPLEX,
                    fontScale=1.0, color=(0, 255, 0), thickness=2)
        frame = cv2.resize(src = frame,dsize=(820,520))
        cv2.imshow('detect',frame)
        key = cv2.waitKey(1)
        if key == 27:
            break
    cap.release()
    cv2.destroyAllWindows()

#单张图片的检测
def signal_detect(image_path='images/face1.jpg'):
    frame=cv2.imread(image_path)
    frame = cv2.resize(src=frame, dsize=(416, 416))
    # 进行预测
    class_ids, scores, boxes = yolov_detect(frame)
    # 绘制矩形框
    for (class_id, box) in enumerate(boxes):
        (x, y, w, h) = box
        class_name = classes[class_ids[class_id]]
        confidence = scores[class_ids[class_id]]
        confidence = str(round(confidence, 2))
        cv2.rectangle(img=frame, pt1=(x, y), pt2=(x + w, y + h),
                      color=(0, 255, 0), thickness=2)
        text = class_name + ' ' + confidence
        cvzone.putTextRect(
            img=frame, text=text, pos=(x, y - 10), font=
            cv2.FONT_HERSHEY_SIMPLEX, scale=1.0, colorR=(255, 0, 255),
            thickness=2
        )
        # cv2.putText(img=frame, text=text,
        #             org=(x, y - 10), fontFace=cv2.FONT_HERSHEY_SIMPLEX,
        #             fontScale=1.0, color=(0, 255, 0), thickness=2)

    cv2.imshow('detect', frame)
    cv2.waitKey(0)
    cv2.destroyAllWindows()

if __name__ == '__main__':
    detect_time()