pyqt5+yolov5的AI功能实现……

最新推荐文章于 2024-04-22 22:50:20 发布
最新推荐文章于 2024-04-22 22:50:20 发布
阅读量3.3k 收藏 41
点赞数 1
分类专栏： yolov5 文章标签： python pyqt5 opencv pytorch 目标检测
版权声明：本文为博主原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。
本文链接：https://blog.csdn.net/qq_41319718/article/details/119391464
版权

项目场景：

基于yolov5封装在UI上的功能实现。此次项目目标是实现监测手机为label的功能实现UI

配置：

windows 7 64位
Jetson NX （不做处理的话默认达到10帧。勉强能跑）
gtx 1080Ti
yolov5 4.0 + pytorch1.6 + opencv4.5.x +cuda 10.2
Anaconda -------pyqt5
海康网络监控摄像头——rtsp
Jetson Xavier NX ——英伟达家的开发板

实现效果：

1.图档侦测
2.视频侦测
3.USB物体侦测
4.rtsp物体侦测
5.截图、存图、LOG档
6.最大化、最小化、固定窗
7.实时滚动图——侦测Label 显示最近

后续补充优化：
1、GPIO输出语音提示模块
2、工业PLC 网络模块交握
3、侦测计数统计图 total
在这里插入图片描述

# 移植yolov5 detect.py 模型初始化： 
import os, threading
import sys
import cv2
import argparse
import random
import torch
import numpy as np
import torch.backends.cudnn as cudnn
import time

from PyQt5 import QtCore, QtGui, QtWidgets

from utils.torch_utils import select_device
from models.experimental import attempt_load
from utils.general import check_img_size, non_max_suppression, scale_coords
from utils.datasets import letterbox
from utils.plots import plot_one_box
from PyQt5.QtWidgets import *

flag = False
flag2= False

class Ui_MainWindow(QtWidgets.QMainWindow):
    def __init__(self, parent=None):
        super(Ui_MainWindow, self).__init__(parent)
        self.timer_video = QtCore.QTimer()
        self.timer_video_1 = QtCore.QTimer()
        self.setupUi(self)
        self.init_slots()
        self.cap = cv2.VideoCapture()



####---模型初始化---
        parser = argparse.ArgumentParser()
        parser.add_argument('--weights', nargs='+', type=str, default='weights/yolov5s.pt', help='model.pt path(s)')
        parser.add_argument('--source', type=str, default='data/images', help='source')  # file/folder, 0 for webcam
        parser.add_argument('--img-size', type=int, default=640, help='inference size (pixels)')
        parser.add_argument('--conf-thres', type=float, default=0.25, help='object confidence threshold')
        parser.add_argument('--iou-thres', type=float, default=0.45, help='IOU threshold for NMS')
        parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
        parser.add_argument('--view-img', action='store_true', help='display results')
        parser.add_argument('--save-txt', action='store_true', help='save results to *.txt')
        parser.add_argument('--save-conf', action='store_true', help='save confidences in --save-txt labels')
        parser.add_argument('--nosave', action='store_true', help='do not save images/videos')
        parser.add_argument('--classes', nargs='+', type=int, help='filter by class: --class 0, or --class 0 2 3')
        parser.add_argument('--agnostic-nms', action='store_true', help='class-agnostic NMS')
        parser.add_argument('--augment', action='store_true', help='augmented inference')
        parser.add_argument('--update', action='store_true', help='update all models')
        parser.add_argument('--project', default='runs/detect', help='save results to project/name')
        parser.add_argument('--name', default='exp', help='save results to project/name')
        parser.add_argument('--exist-ok', action='store_true', help='existing project/name ok, do not increment')
        self.opt = parser.parse_args()
        print(self.opt)

        source, weights, view_img, save_txt, imgsz = self.opt.source, self.opt.weights, self.opt.view_img, self.opt.save_txt, self.opt.img_size

        self.device = select_device(self.opt.device)
        self.half = self.device.type != 'cpu'  # half precision only supported on CUDA

        cudnn.benchmark = True

        # Load model
        self.model = attempt_load(weights, map_location=self.device)  # load FP32 model
        stride = int(self.model.stride.max())  # model stride
        self.imgsz = check_img_size(imgsz, s=stride)  # check img_size
        if self.half:
            self.model.half()  # to FP16

        # Get names and colors
        self.names = self.model.module.names if hasattr(self.model, 'module') else self.model.names
        self.colors = [[random.randint(0, 255) for _ in range(3)] for _ in self.names]


# 基于Anaconda 底下 designer.exe   配置pyqt5 封装UI

#######---UI界面---
    def setupUi(self, MainWindow):
        MainWindow.setObjectName("MainWindow")
        MainWindow.resize(1270, 855)
        MainWindow.setBaseSize(QtCore.QSize(5, 0))
        MainWindow.setFixedSize(self.width(), self.height())
        MainWindow.setStyleSheet("\n" "#MainWindow {background-image: url(C:/Users/Administrator/Anaconda3/envs/pytorch1.6/yolov5/img/pyqt5 backup.png);}")
        self.centralwidget = QtWidgets.QWidget(MainWindow)
        self.centralwidget.setObjectName("centralwidget")
        self.frame = QtWidgets.QFrame(self.centralwidget)
        self.frame.setGeometry(QtCore.QRect(9, 10, 1251, 741))
        self.frame.setFrameShape(QtWidgets.QFrame.StyledPanel)
        self.frame.setFrameShadow(QtWidgets.QFrame.Sunken)
        self.frame.setObjectName("frame")
        self.label = QtWidgets.QLabel(self.frame)
        self.label.setGeometry(QtCore.QRect(7, 14, 1231, 721))
       # self.label.setStyleSheet("background-color: rgb(15, 15, 15);")
        self.label.setTextFormat(QtCore.Qt.AutoText)
        self.label.setObjectName("label")
        self.pushButton = QtWidgets.QPushButton(self.frame)
        self.pushButton.setGeometry(QtCore.QRect(1220, 717, 21, 21))
        self.pushButton.setStyleSheet("")
        self.pushButton.setText("")
        icon = QtGui.QIcon()
        icon.addPixmap(QtGui.QPixmap("C:/Users/Administrator/Anaconda3/envs/pytorch1.6/yolov5/img/mm.png"), QtGui.QIcon.Normal, QtGui.QIcon.Off)
        self.pushButton.setIcon(icon)
        self.pushButton.setIconSize(QtCore.QSize(18, 18))
        self.pushButton.setObjectName("pushButton")
        self.train = QtWidgets.QPushButton(self.frame)
        self.train.setGeometry(QtCore.QRect(1190, 717, 21, 21))
        self.train.setStyleSheet("")
        self.train.setText("")
        icon1 = QtGui.QIcon()
        icon1.addPixmap(QtGui.QPixmap("C:/Users/Administrator/Anaconda3/envs/pytorch1.6/yolov5/img/ppn.png"), QtGui.QIcon.Normal, QtGui.QIcon.Off)
        self.train.setIcon(icon1)
        self.train.setIconSize(QtCore.QSize(20, 20))
        self.train.setObjectName("train")
        self.pushButton_3 = QtWidgets.QPushButton(self.frame)
        self.pushButton_3.setGeometry(QtCore.QRect(1160, 717, 21, 21))
        self.pushButton_3.setStyleSheet("")
        self.pushButton_3.setText("")
        icon2 = QtGui.QIcon()
        icon2.addPixmap(QtGui.QPixmap("C:/Users/Administrator/Anaconda3/envs/pytorch1.6/yolov5/img/min.png"), QtGui.QIcon.Normal, QtGui.QIcon.Off)
        self.pushButton_3.setIcon(icon2)
        self.pushButton_3.setIconSize(QtCore.QSize(18, 18))
        self.pushButton_3.setObjectName("pushButton_3")
        self.label_2 = QtWidgets.QLabel(self.frame)
        self.label_2.setGeometry(QtCore.QRect(430, 720, 511, 411))
        self.label_2.setObjectName("label_2")
        self.btn_opencam_img = QtWidgets.QPushButton(self.centralwidget)
        self.btn_opencam_img.setGeometry(QtCore.QRect(100, 760, 131, 61))
        self.btn_opencam_img.setStyleSheet("color: rgb(255, 255, 255);\n"
"\n"
"font: 75 9pt \"Aharoni\";\n"
"color: rgb(0, 0, 0);\n"
"font: 75 9pt \"微软雅黑\";")
        self.btn_opencam_img.setObjectName("btn_opencam_img")
        self.btn_opencam_video = QtWidgets.QPushButton(self.centralwidget)
        self.btn_opencam_video.setGeometry(QtCore.QRect(310, 760, 131, 61))
        self.btn_opencam_video.setStyleSheet("color: rgb(255, 255, 255);\n"
"font: 75 9pt \"微软雅黑\";\n"
"color: rgb(2, 2, 2);")
        self.btn_opencam_video.setObjectName("btn_opencam_video")
        self.btn_opencam = QtWidgets.QPushButton(self.centralwidget)
        self.btn_opencam.setGeometry(QtCore.QRect(540, 760, 131, 61))
        self.btn_opencam.setStyleSheet("color: rgb(255, 255, 255);\n"
"font: 75 9pt \"微软雅黑\";\n"
"color: rgb(0, 0, 0);")
        self.btn_opencam.setObjectName("btn_opencam")
        self.frame_2 = QtWidgets.QFrame(self.centralwidget)
        self.frame_2.setGeometry(QtCore.QRect(9, 750, 1251, 81))
        self.frame_2.setFrameShape(QtWidgets.QFrame.StyledPanel)
        self.frame_2.setFrameShadow(QtWidgets.QFrame.Sunken)
        self.frame_2.setObjectName("frame_2")
        self.btn_opencam_2 = QtWidgets.QPushButton(self.frame_2)
        self.btn_opencam_2.setGeometry(QtCore.QRect(760, 10, 131, 61))
        self.btn_opencam_2.setStyleSheet("color: rgb(255, 255, 255);\n"
"font: 75 9pt \"微软雅黑\";\n"
"color: rgb(0, 0, 0);")
        self.btn_opencam_2.setObjectName("btn_opencam_2")
        self.iccon4 = QtWidgets.QLabel(self.frame_2)
        self.iccon4.setGeometry(QtCore.QRect(850, 40, 41, 31))
        self.iccon4.setStyleSheet("\n"
"image: url(C:/Users/Administrator/Anaconda3/envs/pytorch1.6/yolov5/img/0.png);")
        self.iccon4.setText("")
        self.iccon4.setObjectName("iccon4")
        self.btn_opencam_3 = QtWidgets.QPushButton(self.frame_2)
        self.btn_opencam_3.setGeometry(QtCore.QRect(990, 10, 141, 61))
        self.btn_opencam_3.setStyleSheet("color: rgb(255, 255, 255);\n"
"background-color: rgb(127, 127, 127);\n"
"font: 75 14pt \"Aharoni\";\n"
"color: rgb(255, 255, 255);")
        self.btn_opencam_3.setObjectName("btn_opencam_3")
        self.iccon5 = QtWidgets.QLabel(self.frame_2)
        self.iccon5.setGeometry(QtCore.QRect(1090, 40, 41, 31))
       # self.iccon5.setStyleSheet("\n"
#"image: url(C:/Users/Administrator/Anaconda3/envs/pytorch1.6/yolov5/img/0.png);")
        self.iccon5.setText("")
        self.iccon5.setObjectName("iccon5")
        self.down = QtWidgets.QLabel(self.frame_2)
        self.down.setGeometry(QtCore.QRect(1150, -10, 101, 101))
        self.down.setStyleSheet("font: 75 18pt \"Aharoni\";\n"
"border-image: url(C:/Users/Administrator/Anaconda3/envs/pytorch1.6/yolov5/img/down.png);")
        self.down.setText("")
        self.down.setObjectName("down")
        self.iccon = QtWidgets.QLabel(self.centralwidget)
        self.iccon.setGeometry(QtCore.QRect(200, 790, 41, 31))
        self.iccon.setStyleSheet("\n"
"image: url(C:/Users/Administrator/Anaconda3/envs/pytorch1.6/yolov5/img/0.png);")
        self.iccon.setText("")
        self.iccon.setObjectName("iccon")
        self.iccon2 = QtWidgets.QLabel(self.centralwidget)
        self.iccon2.setGeometry(QtCore.QRect(400, 790, 41, 31))
        self.iccon2.setStyleSheet("\n"
"image: url(C:/Users/Administrator/Anaconda3/envs/pytorch1.6/yolov5/img/0.png);")
        self.iccon2.setText("")
        self.iccon2.setObjectName("iccon2")
        self.iccon3 = QtWidgets.QLabel(self.centralwidget)
        self.iccon3.setGeometry(QtCore.QRect(630, 790, 41, 31))
        self.iccon3.setStyleSheet("\n"
"image: url(C:/Users/Administrator/Anaconda3/envs/pytorch1.6/yolov5/img/0.png);")
        self.iccon3.setText("")
        self.iccon3.setObjectName("iccon3")
        self.frame_2.raise_()
        self.frame.raise_()
        self.btn_opencam_img.raise_()
        self.btn_opencam_video.raise_()
        self.btn_opencam.raise_()
        self.iccon.raise_()
        self.iccon2.raise_()
        self.iccon3.raise_()
        MainWindow.setCentralWidget(self.centralwidget)
        self.menubar = QtWidgets.QMenuBar(MainWindow)
        self.menubar.setGeometry(QtCore.QRect(0, 0, 1270, 23))
        self.menubar.setObjectName("menubar")
        MainWindow.setMenuBar(self.menubar)
        self.statusbar = QtWidgets.QStatusBar(MainWindow)
        self.statusbar.setObjectName("statusbar")
        MainWindow.setStatusBar(self.statusbar)

        self.retranslateUi(MainWindow)
        QtCore.QMetaObject.connectSlotsByName(MainWindow)

    def retranslateUi(self, MainWindow):
        _translate = QtCore.QCoreApplication.translate
        MainWindow.setWindowTitle(_translate("MainWindow", "实习工厂_SC"))
        self.label.setText(_translate("MainWindow", "显示区"))
        self.label_2.setText(_translate("MainWindow", "显示区"))
        self.btn_opencam_img.setText(_translate("MainWindow", "Images\n"
"打开图档"))
        self.btn_opencam_video.setText(_translate("MainWindow", "Video\n"
"视频选择"))
        self.btn_opencam.setText(_translate("MainWindow", "Camear\n"
"USB摄像头"))
        self.btn_opencam_2.setText(_translate("MainWindow", "Rtsp\n"
"网络摄像头"))
        self.btn_opencam_3.setText(_translate("MainWindow", "Action\n"
"超级播放器"))



 # 功能实现
   def init_slots(self):
        self.btn_opencam_img.clicked.connect(self.button_image_open)
        self.btn_opencam_video.clicked.connect(self.button_video_open)
        self.timer_video.timeout.connect(self.show_video_frame)
        self.btn_opencam.clicked.connect(self.button_opencameras_click_0)
        self.btn_opencam_2.clicked.connect(self.button_opencameras_click)
        self.timer_video_1.timeout.connect(self.show_video_frame_1)
        self.train.clicked.connect(self.Openvideo_click)
#########---算法描框---
    def show_video_frame(self):
        name_list = []

        flag, img = self.cap.read()
        if img is not None:
            showimg = img
            with torch.no_grad():
                img = letterbox(img, new_shape=self.opt.img_size)[0]
                # Convert
                img = img[:, :, ::-1].transpose(2, 0, 1)  # BGR to RGB, to 3x416x416
                img = np.ascontiguousarray(img)
                img = torch.from_numpy(img).to(self.device)
                img = img.half() if self.half else img.float()  # uint8 to fp16/32
                img /= 255.0  # 0 - 255 to 0.0 - 1.0
                if img.ndimension() == 3:
                    img = img.unsqueeze(0)
                # Inference
                pred = self.model(img, augment=self.opt.augment)[0]

                # Apply NMS
                pred = non_max_suppression(pred, self.opt.conf_thres, self.opt.iou_thres, classes=self.opt.classes,
                                           agnostic=self.opt.agnostic_nms)
                # Process detections
                for i, det in enumerate(pred):  # detections per image
                    if det is not None and len(det):
                        # Rescale boxes from img_size to im0 size
                        det[:, :4] = scale_coords(img.shape[2:], det[:, :4], showimg.shape).round()
                        # Write results
                        for *xyxy, conf, cls in reversed(det):
                            label = '%s %.2f' % (self.names[int(cls)], conf)
                            name_list.append(self.names[int(cls)])
                            print(label)
                            plot_one_box(xyxy, showimg, label=label, color=self.colors[int(cls)], line_thickness=2)

            show = cv2.resize(showimg, (1280, 720))
            self.result = cv2.cvtColor(show, cv2.COLOR_BGR2RGB)
            showImage = QtGui.QImage(self.result.data, self.result.shape[1], self.result.shape[0],
                                     QtGui.QImage.Format_RGB888)
            self.label.setPixmap(QtGui.QPixmap.fromImage(showImage))

        else:
            self.timer_video.stop()
            self.cap.release()
            self.label.clear()
          #  self.pushButton_video.setDisabled(False)
           # self.pushButton_img.setDisabled(False)
            self.init_logo()


    def show_video_frame_1(self):
        name_list = []

        flag, img = self.cap.read()
        if img is not None:
            showimg = img
            with torch.no_grad():
                img = letterbox(img, new_shape=self.opt.img_size)[0]
                # Convert
                img = img[:, :, ::-1].transpose(2, 0, 1)  # BGR to RGB, to 3x416x416
                img = np.ascontiguousarray(img)
                img = torch.from_numpy(img).to(self.device)
                img = img.half() if self.half else img.float()  # uint8 to fp16/32
                img /= 255.0  # 0 - 255 to 0.0 - 1.0
                if img.ndimension() == 3:
                    img = img.unsqueeze(0)
                # Inference
                pred = self.model(img, augment=self.opt.augment)[0]

                # Apply NMS
                pred = non_max_suppression(pred, self.opt.conf_thres, self.opt.iou_thres, classes=self.opt.classes,
                                           agnostic=self.opt.agnostic_nms)
                # Process detections
                for i, det in enumerate(pred):  # detections per image
                    if det is not None and len(det):
                        # Rescale boxes from img_size to im0 size
                        det[:, :4] = scale_coords(img.shape[2:], det[:, :4], showimg.shape).round()
                        # Write results
                        for *xyxy, conf, cls in reversed(det):
                            label = '%s %.2f' % (self.names[int(cls)], conf)
                            name_list.append(self.names[int(cls)])
                            print(label)
                            self.reclabel =  name_list
                            plot_one_box(xyxy, showimg, label=label, color=self.colors[int(cls)], line_thickness=2)

            show = cv2.resize(showimg, (1280, 720))
            self.result = cv2.cvtColor(show, cv2.COLOR_BGR2RGB)
            showImage = QtGui.QImage(self.result.data, self.result.shape[1], self.result.shape[0],
                                     QtGui.QImage.Format_RGB888)
            self.label.setPixmap(QtGui.QPixmap.fromImage(showImage))

        else:
            self.timer_video_1.stop()
            self.cap.release()
            self.label.clear()
          #  self.pushButton_video.setDisabled(False)
           # self.pushButton_img.setDisabled(False)
            self.init_logo()








 #######---图档事件---
    def button_image_open(self):
        print('button_image_open')
        name_list = []

        img_name, _ = QtWidgets.QFileDialog.getOpenFileName(self, "打开图片", "", "*.jpg;;*.png;;All Files(*)")
        flag = self.cap.open(img_name)
        if flag == False:
            QtWidgets.QMessageBox.warning(self, u"Warning", u"打开图片失败", buttons=QtWidgets.QMessageBox.Ok,
                                          defaultButton=QtWidgets.QMessageBox.Ok)

        else:
            img = cv2.imread(img_name)
            print(img_name)
            showimg = img
            with torch.no_grad():
                img = letterbox(img, new_shape=self.opt.img_size)[0]
                # Convert
                img = img[:, :, ::-1].transpose(2, 0, 1)  # BGR to RGB, to 3x416x416
                img = np.ascontiguousarray(img)
                img = torch.from_numpy(img).to(self.device)
                img = img.half() if self.half else img.float()  # uint8 to fp16/32
                img /= 255.0  # 0 - 255 to 0.0 - 1.0
                if img.ndimension() == 3:
                    img = img.unsqueeze(0)
                # Inference
                pred = self.model(img, augment=self.opt.augment)[0]
                # Apply NMS
                pred = non_max_suppression(pred, self.opt.conf_thres, self.opt.iou_thres, classes=self.opt.classes,
                                           agnostic=self.opt.agnostic_nms)
                print(pred)
                # Process detections
                for i, det in enumerate(pred):
                    if det is not None and len(det):
                        # Rescale boxes from img_size to im0 size
                        det[:, :4] = scale_coords(img.shape[2:], det[:, :4], showimg.shape).round()

                        for *xyxy, conf, cls in reversed(det):
                            label = '%s %.2f' % (self.names[int(cls)], conf)
                            name_list.append(self.names[int(cls)])
                            plot_one_box(xyxy, showimg, label=label, color=self.colors[int(cls)], line_thickness=2)

            self.result = cv2.cvtColor(showimg, cv2.COLOR_BGR2BGRA)
            self.result = cv2.resize(self.result, (1280, 720), interpolation=cv2.INTER_AREA)
            self.QtImg = QtGui.QImage(self.result.data, self.result.shape[1], self.result.shape[0], QtGui.QImage.Format_RGB32)
            self.label.setPixmap(QtGui.QPixmap.fromImage(self.QtImg))

###---视频事件---
    def button_video_open(self):
        global flag
        if flag == False:
            flag = True
            video_name, _ = QtWidgets.QFileDialog.getOpenFileName(self, "打开视频", "", "*.mp4;;*.avi;;All Files(*)")
            flag = self.cap.open(video_name)
            if flag == False:
                QtWidgets.QMessageBox.warning(self, u"Warning", u"打开视频失败", buttons=QtWidgets.QMessageBox.Ok,
                                              defaultButton=QtWidgets.QMessageBox.Ok)
            else:
                self.timer_video.start(30)
                # 空间按钮 Enable
                self.btn_opencam_video.setText(u'关闭识别')


        else:
            flag = False
            self.timer_video.stop()
            self.cap.release()
            self.label.clear()
            self.btn_opencam_video.setText(u'打开视频')
            # self.pushButton_2.setDisabled(False)
            # self.pushButton.setDisabled(False)

###---Webcam---
    def button_opencameras_click(self):
        global flag2
        self.timer_video_1.stop()
        self.cap.release()
        if flag2 == False:
            flag2=True
            # 默认是打开usb本地摄像头，如果想打开rtsp码流摄像头修改self.CAM_NUM为你的rtsp地址，例如："rtsp://admin:test123456@192.168.10.65:554/MPEG-4/ch1/main/av_stream";
            # flag = self.cap.open("rtsp://192.168.1.56:554")
            flag = self.cap.open("rtsp://admin:yian1234@192.168.1.56:554/h264/1/main")
            if flag == False:
                msg = QtWidgets.QMessageBox.warning(self, u"Warning", u"请检测相机与电脑是否连接正确",
                                                    buttons=QtWidgets.QMessageBox.Ok,
                                                    defaultButton=QtWidgets.QMessageBox.Ok)
                self.btn_opencam.setDisabled(True)
                self.btn_opencam_2.setText(u'重新启动')
            else:
                self.timer_video_1.start(0)
                self.btn_opencam_2.setText(u'关闭识别')
                self.btn_opencam.setDisabled(True)
        else:
            flag2=False
            self.timer_video_1.stop()
            self.cap.release()
            self.label.clear()
            self.btn_opencam_2.setText(u'Open Webcam')
            self.btn_opencam.setDisabled(False)

###---USB Camear
    def button_opencameras_click_0(self):
        global flag2
        self.timer_video_1.stop()
        self.cap.release()
        if flag2 == False:
            flag2=True
            # 默认是打开usb本地摄像头，如果想打开rtsp码流摄像头修改self.CAM_NUM为你的rtsp地址，例如："rtsp://admin:test123456@192.168.10.65:554/MPEG-4/ch1/main/av_stream";
            # flag = self.cap.open("rtsp://192.168.1.56:554")
            flag = self.cap.open(0)
            if flag == False:
                msg = QtWidgets.QMessageBox.warning(self, u"Warning", u"请检测相机与电脑是否连接正确",
                                                    buttons=QtWidgets.QMessageBox.Ok,
                                                    defaultButton=QtWidgets.QMessageBox.Ok)
                self.btn_opencam_2.setDisabled(True)
                self.btn_opencam.setText(u'重新启动')
            else:
                self.timer_video_1.start(0)
                self.btn_opencam.setText(u'关闭识别')
                self.btn_opencam_2.setDisabled(True)
        else:
            flag2=False
            self.timer_video_1.stop()
            self.cap.release()
            self.label.clear()
            self.btn_opencam.setText(u'Open Camera')
            self.btn_opencam_2.setDisabled(False)


###---定格
    def Openvideo_click(self):
        prev_time = time.time()
        result = cv2.cvtColor(self.result, cv2.COLOR_BGR2RGB)
        cv2.imwrite('save/' + str(prev_time) + '.jpg', result)
        print(self.reclabel)
        fw = open('save/' + str(prev_time) + '.txt', 'a')
        for i in range(len(self.reclabel)):
            fw.write(self.reclabel[i] + '\n')
        fw.close()







###---大屏显示
class B(QtWidgets.QMainWindow):
    def __init__(self, parent=None):
        super(B, self).__init__(parent)
        self.setupUi(self)
        self.timer_video = QtCore.QTimer()

        self.init_slots()
        self.cap = cv2.VideoCapture()


        ####---模型初始化---
        parser = argparse.ArgumentParser()
        parser.add_argument('--weights', nargs='+', type=str, default='weights/yolov5s.pt', help='model.pt path(s)')
        parser.add_argument('--source', type=str, default='data/images', help='source')  # file/folder, 0 for webcam
        parser.add_argument('--img-size', type=int, default=640, help='inference size (pixels)')
        parser.add_argument('--conf-thres', type=float, default=0.25, help='object confidence threshold')
        parser.add_argument('--iou-thres', type=float, default=0.45, help='IOU threshold for NMS')
        parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
        parser.add_argument('--view-img', action='store_true', help='display results')
        parser.add_argument('--save-txt', action='store_true', help='save results to *.txt')
        parser.add_argument('--save-conf', action='store_true', help='save confidences in --save-txt labels')
        parser.add_argument('--nosave', action='store_true', help='do not save images/videos')
        parser.add_argument('--classes', nargs='+', type=int, help='filter by class: --class 0, or --class 0 2 3')
        parser.add_argument('--agnostic-nms', action='store_true', help='class-agnostic NMS')
        parser.add_argument('--augment', action='store_true', help='augmented inference')
        parser.add_argument('--update', action='store_true', help='update all models')
        parser.add_argument('--project', default='runs/detect', help='save results to project/name')
        parser.add_argument('--name', default='exp', help='save results to project/name')
        parser.add_argument('--exist-ok', action='store_true', help='existing project/name ok, do not increment')
        self.opt = parser.parse_args()
        print(self.opt)

        source, weights, view_img, save_txt, imgsz = self.opt.source, self.opt.weights, self.opt.view_img, self.opt.save_txt, self.opt.img_size

        self.device = select_device(self.opt.device)
        self.half = self.device.type != 'cpu'  # half precision only supported on CUDA

        cudnn.benchmark = True

        # Load model
        self.model = attempt_load(weights, map_location=self.device)  # load FP32 model
        stride = int(self.model.stride.max())  # model stride
        self.imgsz = check_img_size(imgsz, s=stride)  # check img_size
        if self.half:
            self.model.half()  # to FP16

        # Get names and colors
        self.names = self.model.module.names if hasattr(self.model, 'module') else self.model.names
        self.colors = [[random.randint(0, 255) for _ in range(3)] for _ in self.names]



    def setupUi(self, MainWindow):
        MainWindow.setObjectName("MainWindow")
        MainWindow.resize(1920, 1080)
        self.centralwidget = QtWidgets.QWidget(MainWindow)
        self.centralwidget.setObjectName("centralwidget")
        self.label = QtWidgets.QLabel(self.centralwidget)
        self.label.setGeometry(QtCore.QRect(9, 9, 1920, 1080))
        self.label.setObjectName("label")
        self.pushButton = QtWidgets.QPushButton(self.centralwidget)
        self.pushButton.setGeometry(QtCore.QRect(1800, 910, 100, 50))
        self.pushButton.setObjectName("pushButton")
        MainWindow.setCentralWidget(self.centralwidget)
        self.menubar = QtWidgets.QMenuBar(MainWindow)
        self.menubar.setGeometry(QtCore.QRect(0, 0, 1491, 23))
        self.menubar.setObjectName("menubar")
        MainWindow.setMenuBar(self.menubar)
        self.statusbar = QtWidgets.QStatusBar(MainWindow)
        self.statusbar.setObjectName("statusbar")
        MainWindow.setStatusBar(self.statusbar)

        self.retranslateUi(MainWindow)
        QtCore.QMetaObject.connectSlotsByName(MainWindow)

    def retranslateUi(self, MainWindow):
        _translate = QtCore.QCoreApplication.translate
        MainWindow.setWindowTitle(_translate("MainWindow", "MainWindow"))
        self.label.setText(_translate("MainWindow", "TextLabel"))
        self.pushButton.setText(_translate("MainWindow", "点击3秒触发"))

    def init_slots(self):

        self.timer_video.timeout.connect(self.show_video_frame)
        self.pushButton.clicked.connect(self.button_opencameras_click)




    def show_video_frame(self):
        name_list = []

        flag, img = self.cap.read()
        if img is not None:
            showimg = img
            with torch.no_grad():
                img = letterbox(img, new_shape=self.opt.img_size)[0]
                # Convert
                img = img[:, :, ::-1].transpose(2, 0, 1)  # BGR to RGB, to 3x416x416
                img = np.ascontiguousarray(img)
                img = torch.from_numpy(img).to(self.device)
                img = img.half() if self.half else img.float()  # uint8 to fp16/32
                img /= 255.0  # 0 - 255 to 0.0 - 1.0
                if img.ndimension() == 3:
                    img = img.unsqueeze(0)
                # Inference
                pred = self.model(img, augment=self.opt.augment)[0]

                # Apply NMS
                pred = non_max_suppression(pred, self.opt.conf_thres, self.opt.iou_thres, classes=self.opt.classes,
                                           agnostic=self.opt.agnostic_nms)
                # Process detections
                for i, det in enumerate(pred):  # detections per image
                    if det is not None and len(det):
                        # Rescale boxes from img_size to im0 size
                        det[:, :4] = scale_coords(img.shape[2:], det[:, :4], showimg.shape).round()
                        # Write results
                        for *xyxy, conf, cls in reversed(det):
                            label = '%s %.2f' % (self.names[int(cls)], conf)
                            name_list.append(self.names[int(cls)])
                            print(label)
                            plot_one_box(xyxy, showimg, label=label, color=self.colors[int(cls)], line_thickness=2)

            show = cv2.resize(showimg, (1920, 1080))
            self.result = cv2.cvtColor(show, cv2.COLOR_BGR2RGB)
            showImage = QtGui.QImage(self.result.data, self.result.shape[1], self.result.shape[0],
                                     QtGui.QImage.Format_RGB888)
            self.label.setPixmap(QtGui.QPixmap.fromImage(showImage))

        else:
            self.timer_video.stop()
            self.cap.release()
            self.label.clear()
          #  self.pushButton_video.setDisabled(False)
           # self.pushButton_img.setDisabled(False)
            self.init_logo()


    def button_opencameras_click(self):
        global flag2
        self.timer_video.stop()
        self.cap.release()
        if flag2 == False:
            flag2 = True
            # 默认是打开usb本地摄像头，如果想打开rtsp码流摄像头修改self.CAM_NUM为你的rtsp地址，例如："rtsp://admin:test123456@192.168.10.65:554/MPEG-4/ch1/main/av_stream";
            # flag = self.cap.open("rtsp://192.168.1.56:554")
            flag = self.cap.open("rtsp://admin:yian1234@192.168.1.56:554/h264/1/main")
            if flag == False:
                msg = QtWidgets.QMessageBox.warning(self, u"Warning", u"请检测相机与电脑是否连接正确",
                                                    buttons=QtWidgets.QMessageBox.Ok,
                                                    defaultButton=QtWidgets.QMessageBox.Ok)
                self.pushButton.setText(u'重新启动')
            else:
                self.timer_video.start(0)
                self.pushButton.setText(u'暂停')
        else:
            flag2 = False
            self.timer_video.stop()
            self.cap.release()
            #self.label.clear()
            self.pushButton.setText(u'Open Webcam')








#滚动窗入口
class ChildWindow(QMainWindow):
    def __init__(self):
        super().__init__()

        # 一张大图
        self.label0 = QLabel(self)
        # 五张小图
        self.label1 = QLabel(self)
        self.label2 = QLabel(self)
        self.label3 = QLabel(self)
        self.label4 = QLabel(self)
        self.label5 = QLabel(self)

        self.initUI(self)

    def initUI(self,MainWindow):
        MainWindow.resize(1269, 888)
        MainWindow.setStyleSheet("background-image: url(C:/Users/Administrator/Anaconda3/envs/pytorch1.6/yolov5/img/pyqt5 backup.png);")
       # self.setGeometry(1500, 100, 1060, 900)
        self.setWindowTitle('AUO-滚动图系列')

        th = threading.Thread(target=self.show_images)
        th.start()

    # 显示图片
    def show_images(self):
        while True:
            path = './images'
            filenames = os.listdir(path)
            filenames = [os.path.join(path, filename) for filename in filenames]
            filenames.sort(key=lambda fp: os.path.getctime(fp), reverse=True)
            filenames_5 = filenames[:5]

            src0 = cv2.imread(filenames_5[0])
            src0 = cv2.resize(src0, (960, 600))
            src0 = cv2.cvtColor(src0, cv2.COLOR_BGR2RGB)
            img = QtGui.QImage(src0.data, src0.shape[1], src0.shape[0], QtGui.QImage.Format_RGB888)
            self.label0.setGeometry(QtCore.QRect(50, 50, 960, 600))
            self.label0.setPixmap(QtGui.QPixmap(img))

            src1 = cv2.imread(filenames_5[0])
            src1 = cv2.resize(src1, (160, 120))
            src1 = cv2.cvtColor(src1, cv2.COLOR_BGR2RGB)
            img = QtGui.QImage(src1.data, src1.shape[1], src1.shape[0], QtGui.QImage.Format_RGB888)
            self.label1.setGeometry(QtCore.QRect(50, 720, 160, 120))
            self.label1.setPixmap(QtGui.QPixmap(img))

            src2 = cv2.imread(filenames_5[1])
            src2 = cv2.resize(src2, (160, 120))
            src2 = cv2.cvtColor(src2, cv2.COLOR_BGR2RGB)
            img = QtGui.QImage(src2.data, src2.shape[1], src2.shape[0], QtGui.QImage.Format_RGB888)
            self.label2.setGeometry(QtCore.QRect(250, 720, 160, 120))
            self.label2.setPixmap(QtGui.QPixmap(img))

            src3 = cv2.imread(filenames_5[2])
            src3 = cv2.resize(src3, (200, 120))
            src3 = cv2.cvtColor(src3, cv2.COLOR_BGR2RGB)
            img = QtGui.QImage(src3.data, src3.shape[1], src3.shape[0], QtGui.QImage.Format_RGB888)
            self.label3.setGeometry(QtCore.QRect(450, 720, 160, 120))
            self.label3.setPixmap(QtGui.QPixmap(img))

            src4 = cv2.imread(filenames_5[3])
            src4 = cv2.resize(src4, (200, 120))
            src4 = cv2.cvtColor(src4, cv2.COLOR_BGR2RGB)
            img = QtGui.QImage(src4.data, src4.shape[1], src4.shape[0], QtGui.QImage.Format_RGB888)
            self.label4.setGeometry(QtCore.QRect(650, 720, 160, 120))
            self.label4.setPixmap(QtGui.QPixmap(img))

            src5 = cv2.imread(filenames_5[4])
            src5 = cv2.resize(src5, (200, 120))
            src5 = cv2.cvtColor(src5, cv2.COLOR_BGR2RGB)
            img = QtGui.QImage(src5.data, src5.shape[1], src5.shape[0], QtGui.QImage.Format_RGB888)
            self.label5.setGeometry(QtCore.QRect(850, 720, 160, 120))
            self.label5.setPixmap(QtGui.QPixmap(img))

            time.sleep(1)












if __name__ == '__main__':
    app = QtWidgets.QApplication(sys.argv)
    ui = Ui_MainWindow()
    b = B()
    c = ChildWindow()
    ui.show()
    ui.pushButton.clicked.connect(b.show)
    ui.btn_opencam_3.clicked.connect(c.show)
    sys.exit(app.exec_())





# 总结
刚入门。很不错的学习例子。做个记录。第一个UI完成  噢耶~~~~
别乱别报错
关注
  • 1
    点赞
  • 踩
  • 41
    收藏
    觉得还不错? 一键收藏
  • 5
    评论
专栏目录
yolo 车辆测距+车辆识别+单目测距+车距检测（单目测距,无需摄像头）
sybh的博客
08-08 1万+
基于yolo目标检测算法实现的车前道路中的车辆和行人检测，并且可以估测出目标与本车之间的距离。
[深度学习]yolov9+bytetrack+pyqt5实现目标追踪
lwx666sl的博客
02-28 230
将YOLOv9、ByteTrack和PyQt5整合起来，可以实现一个功能强大的目标追踪系统。ByteTrack利用目标检测算法（如YOLOv9）提供的边界框信息，通过匈牙利算法等匹配策略，将不同帧中的目标关联起来，从而实现目标追踪。YOLOv9是YOLO系列算法的最新版本，它在保持高检测速度的同时，通过改进网络结构、优化损失函数等方式，提高了检测精度。综上所述，结合YOLOv9的高精度目标检测、ByteTrack的高效目标追踪以及PyQt5的直观用户界面，我们可以实现一个功能强大、易于使用的目标追踪系统。
(超详细流程-适合小白入门-一篇就够)基于YOLOv5和PYQT5进行检测界面的制作
m0_48248371的博客
04-22 1877
以上是项目目录,除了跑GUI文件YOLOv5中需要的文件外:test文件夹中存放:测试YOLOv5_PYQT5_GUI所需要的权重文件和图像,用的都是官方权重和图像.ui文件夹中存放:QT Designer制作的.ui文件和转换生成的.py文件.ui_img文件夹中存放:QT Designer制作GUI界面需要的图像文件.Detect_GUI.py文件主要是直接进入检测界面.Login_GUI.py文件主要是从注册和登录界面进入检测界面.
Visdrone数据集YOLOv5训练权重+pyqt界面
06-09
1、yolov5下俯视场景下车辆行人检测视觉检测，包含YOLOv5s和YOLOv5m两种训练好的visdrone数据集权重，以及PR曲线，loss曲线等等，有pyqt界面，目标类别为车辆行人等 2、pyqt界面可以检测图片、视频、调用摄像头 3、数据集和检测结果参考：https://blog.csdn.net/zhiqingAI/article/details/124230743 4、采用pytrch框架，python代码
YOLOv5烟雾检测+pyqt可视化界面 yolov5-6.0-smoke_detect-qt.zip
05-13
1、YOLOv5烟雾检测训练好的模型，可以直接使用，检测 的目标类别名称为smoke，并可以在可视化界面上直接运行，代码已集成好，直接运行就可展示可视化界面，可以通过界面上的 按钮选择图片进行烟雾检测，并且可以调用摄像头检测以及检测相关视频，使用简单，并附有使用说明 2、烟雾数据集下载：https://download.csdn.net/download/weixin_51154380/85190134 3、其他数据集参考：https://blog.csdn.net/weixin_51154380/article/details/124480876
pyqt5+yolov5+python
05-28
内容概要：多目标检测+pyqt5+yolov5+python+实现yolov5GUI开发。 适用人群：刚刚接触pyqt5开发以及刚刚接触yolo算法的初学者，想要有现成项目练手的人。 从本项目中能学到什么：pyqt5的开发、如何界面设计与后端逻辑...
PyQt5+yolov5+unet+centernet
08-27
集yolov5、centernet、unet算法的pyqt5界面，可实现图片目标检测和语义分割
人工智能+目标识别+yolov8+pyqt5+界面
06-06
本文将详细探讨"人工智能+目标识别+yolov8+pyqt5+界面"这一主题，深入解析如何利用PyQt5构建用户界面，并通过调用Yolov8实现图像和视频的目标检测。 首先，目标识别是人工智能中的一个关键子领域，它涉及计算机视觉...
基于PyQt5+yolov5+RestNet18分类实现的骨龄识别检测源码+项目使用说明+模型.zip
11-03
基于PyQt5+yolov5+RestNet18分类实现的骨龄识别检测源码+项目使用说明+模型.zip 数据集来自百度飞浆， 数据集下载后，先对数据进行预处理 先运行yolov5_master/images_tag.py 并先修改里面的图片路径 ![image](66...
基于PyQt5+YOLOv5算法实现的人体识别检测追踪系统python源码+模型+项目说明.zip
最新发布
06-19
基于PyQt5+YOLOv5算法实现的人体识别检测追踪系统python源码+模型+项目说明.zip基于PyQt5+YOLOv5算法实现的人体识别检测追踪系统python源码+模型+项目说明.zip基于PyQt5+YOLOv5算法实现的人体识别检测追踪系统python...
YOLOv5行人重识别+PYQT5操作画面
04-07
将YOLOv5行人重识别+PYQT5操作画面相结合，可以实现以下功能： 1. 基于已训练的行人重识别模型，实时进行行人框标注，以进行后续的行人识别。 2. 进行行人框标注后，自动保存到query文件中，得到行人特征，并在输入视频帧中进行行人重识别 3. 操作画面使用PYQT进行简单设计，可以选择输入视频、照片等，并实时显示画面 4. 引用大佬博客https://blog.csdn.net/z240626191s/article/details/129221510?spm=1001.2014.3001.5501
linux桌面小程序开发日记4(pyqt5+yolov5)
Liuchengzhizhi的博客
02-13 1664
linux桌面小程序开发日记4(pyqt5+yolov5)
使用YOLOv5模型进行目标检测！AI大佬手撕源码带你学
m0_74693860的博客
11-15 3587
YOLO主要做了什么? 通俗的讲目标检测是计算机视觉领域的一大任务，大致分为一阶段目标检测与两阶段目标检测。其中一阶段目标检测模型以YOLO系列为代表。
当YOLOv5碰上PyQt5 ...
热门推荐
06-30 1万+
环境windows 10 64位gtx 1050Tiyolov5 5.0 + pytorch1.7 + cu110pyqt5前言有小伙伴在问，能不能给yolov5做个GUI界面，比如使用p...
qt界面上显示yolov5检测结果
weixin_35756637的博客
01-10 971
在 Qt 界面上显示 YOLOv5 检测结果可以使用以下步骤进行： 使用 YOLOv5 进行检测，将检测到的结果存储为一个数组或列表。 在 Qt 界面上创建一个 QLabel 控件，用来显示图像。 将检测到的结果在图像上绘制出来，使用 QPainter 类。 将处理后的图像设置为 QLabel 控件的图像。 更新 QLabel 控件，使得修改后的图像能够显示在界面上。 你可以使用 O...
从零开始使用YOLOv5+PyQt5+OpenCV+爬虫实现是否佩戴口罩检测
qq_37553692的博客
02-20 4678
从零开始使用YOLOv5+PyQt5+OpenCV实现是否佩戴口罩检测。全流程教程，从数据采集YOLO框架（You Only Look Once）与RCNN系列算法不一样，是以不同的方式处理对象检测。它将整个图像放在一个实例中，并预测这些框的边界框坐标和及所属类别概率。使用YOLO算法最大优的点是速度极快，每秒可处理45帧，也能够理解一般的对象表示。首先，输入图像：然后，YOLO将输入图像划分为网格形式（例如3 X 3）：最后，对每个网格应用图像分类和定位处理，获得预测对象的边界框及其对应的类概率。
基于Qt的yolov5工程
qq_44798484的博客
06-26 1953
基于Qt的工程
yolov5调用海康相机
10-23
Yolo是一种目标检测算法，而Yolov5是Yolo系列中最新的版本。海康相机是海康威视公司生产的一种高质量监控相机。要将Yolov5与海康相机进行结合使用，可以按照以下步骤进行: 1. 确保你有一台配备海康相机的计算机，并已连接到网络。对于Yolov5的安装和配置，你可以参考官方文档进行操作。 2. 下载Yolov5的代码库，并通过命令行进入该代码库所在的文件夹。 3. 在命令行中运行`pip install -r requirements.txt`安装所需的依赖库。 4. 创建一个Python脚本，导入所需的模块。例如，`import cv2`用于处理图像，`import torch`用于模型的加载和预测。 5. 在脚本中，通过使用`torch.hub.load('ultralytics/yolov5', 'yolov5s')`加载Yolov5模型。这会从Torch Hub中下载预训练模型。 6. 通过`cap = cv2.VideoCapture('<海康相机的URL>')`打开海康相机的视频流。 7. 在一个循环中，不断读取相机的视频帧，并将其传递给Yolov5模型进行目标检测。 8. 可以通过对检测到的目标进行后续处理，例如绘制边界框和显示检测结果。 9. 最后，记得释放相机资源和关闭窗口。通过`cap.release()`释放相机资源，通过`cv2.destroyAllWindows()`关闭窗口。 综上所述，这是将Yolov5与海康相机进行结合使用的一般步骤。具体的实现可能会因实际环境和需求而有所不同，可以根据具体情况进行调整。

“相关推荐”对你有帮助么？

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论 5
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付：10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明：

1.余额是钱包充值的虚拟货币，按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载，可以购买VIP、付费专栏及课程。

余额充值