---

title: IP摄像头实现远程目标检测

硬件准备

不用的手机，或者esp32_CAM、电脑
实现思路：
摄像头负责采集图像传送到云端
电脑获取云端视频流，在PC端对视频进行图像处理。
然后将处理结果反馈给摄像头终端（未实现）

软件准备

Python 3.9、Pycharm、OpenCV、pytorch、手机安装IP摄像头软件

代码实现

以下只是使用手机IP摄像头的代码

import time
import torch
import cv2 as cv


class MultipleTarget:

    def __init__(self, url):
        """
        初始化
        """
        # 加载训练模型
        self.model = torch.hub.load('./yolov5', 'custom', path='./weight/yolov5s.pt', source='local')
        # 设置阈值
        self.model.conf = 0.52  # confidence threshold (0-1)
        self.model.iou = 0.45  # NMS IoU threshold (0-1)
        # 加载摄像头
        self.url = url
        self.cap = cv.VideoCapture(self.url)
        if not self.cap.isOpened():
            print("Cannot open camera")
            exit()

    def draw(self, list_temp, image_temp):
        for temp in list_temp:
            name = temp[6]  # 取出标签名
            temp = temp[:4].astype('int')  # 转成int加快计算
            cv.rectangle(image_temp, (temp[0], temp[1]), (temp[2], temp[3]), (0, 0, 255), 3)  # 框出识别物体
            cv.putText(image_temp, name, (int(temp[0] - 10), int(temp[1] - 10)), cv.FONT_ITALIC, 1, (0, 255, 0), 2)

    def detect(self):
        """
        目标检测
        """
        while True:
            ret, frame = self.cap.read()
            # 如果正确读取帧，ret为True
            if not ret:
                print("Can't receive frame (stream end?). Exiting ...")
                break
            frame = cv.flip(frame, 1)

            # FPS计算time.start
            start_time = time.time()

            # Inference
            results = self.model(frame)
            pd = results.pandas().xyxy[0]  # tensor-->pandas的DataFrame
            # 取出对应标签的list
            person_list = pd[pd['name'] == 'person'].to_numpy()
            bus_list = pd[pd['name'] == 'bus'].to_numpy()
            # 框出物体
            self.draw(person_list, frame)
            self.draw(bus_list, frame)
            # end_time
            end_time = time.time()
            fps = 1 / (end_time - start_time)

            # 控制台显示
            # results.print()  # or .show(), .save(), .crop(), .pandas(), etc.
            # print(results.xyxy[0])  # img1 predictions (tensor)
            # print('----------------')
            # print(results.pandas().xyxy[0])  # img1 predictions (pandas)

            # FPS显示
            cv.putText(frame, 'FPS:' + str(int(fps)), (30, 50), cv.FONT_ITALIC, 1, (0, 255, 0), 2)

            cv.imshow('results', frame)
            cv.waitKey(10)
            if cv.waitKey(10) & 0xFF == ord('q'):
                break

        self.cap.release()
        cv.destroyAllWindows()


url = 'rtsp://admin:admin@192.168.43.229:8554/live'
test = MultipleTarget(url)
test.detect()

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80

结果

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-F5O6Atei-1656468386884)(https://i.niupic.com/images/2022/06/20/a0B1.jpg)]
test = MultipleTarget(url)
test.detect()

# 结果
[外链图片转存中...(img-F5O6Atei-1656468386884)]
[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-G3F6pfjF-1656468386888)(https://i.niupic.com/images/2022/06/20/a0B2.jpg)]
1
2
3