yolov5 +gui界面+单目测距 实现对图片视频摄像头的测距

可实现对图片,视频,摄像头的检测

项目概述

本项目旨在实现一个集成了YOLOv5目标检测算法、图形用户界面(GUI)以及单目测距功能的系统。该系统能够对图片、视频或实时摄像头输入进行目标检测,并估算目标的距离。通过结合YOLOv5的强大检测能力和单目测距技术,系统能够在多种应用场景中提供高效、准确的目标检测和测距功能。

技术栈
  • YOLOv5:用于目标检测的深度学习模型。
  • OpenCV:用于图像处理和单目测距算法。
  • PyTorch:YOLOv5模型的底层框架。
  • Tkinter:用于创建图形用户界面(GUI)。
  • Python:开发语言。
系统功能
  1. 目标检测:使用YOLOv5模型对输入图像或视频流中的目标进行检测。
  2. 单目测距:基于检测到的目标,利用单目测距技术估算目标的距离。
  3. GUI界面:提供用户友好的图形界面,方便用户操作和查看结果。
系统特点
  1. 高效检测:YOLOv5模型具有高效的检测速度,适用于实时应用场景。
  2. 准确测距:单目测距技术能够较为准确地估算目标距离。
  3. 用户友好:通过图形界面,用户可以轻松选择输入源(图片、视频或摄像头)并查看检测结果和测距信息。
系统架构
  1. 输入源选择:用户可以选择图片、视频或实时摄像头作为输入源。
  2. 目标检测:使用YOLOv5模型对输入源进行目标检测,返回检测框和类别信息。
  3. 单目测距:根据检测到的目标,利用单目测距算法估算目标距离。
  4. 结果展示:在GUI界面上显示检测结果和测距信息。
关键技术
  1. YOLOv5模型:YOLOv5是一种高性能的目标检测模型,能够实时检测多种目标类别。
  2. 单目测距算法:利用已知物体尺寸和相机焦距等参数,通过图像中的物体大小变化来估算距离。
  3. GUI界面设计:使用Tkinter库创建用户界面,方便用户操作和查看结果。
系统流程
  1. 输入源选择:用户在GUI界面上选择输入源(图片、视频或摄像头)。
  2. 图像预处理:对输入图像或视频帧进行预处理,如缩放、归一化等。
  3. 目标检测:使用YOLOv5模型对预处理后的图像进行目标检测。
  4. 单目测距:根据检测结果,利用单目测距算法估算目标距离。
  5. 结果展示:在GUI界面上显示检测框、类别信息和测距结果

main.py

from PyQt5.QtWidgets import QApplication, QMainWindow, QFileDialog, QMenu, QAction
from main_win.win import Ui_mainWindow
from PyQt5.QtCore import Qt, QPoint, QTimer, QThread, pyqtSignal
from PyQt5.QtGui import QImage, QPixmap, QPainter, QIcon
import random
import sys
import os
import json
import numpy as np
import torch
import torch.backends.cudnn as cudnn
import os
import time
import cv2

from models.experimental import attempt_load
from utils.datasets import LoadImages, LoadWebcam
from utils.CustomMessageBox import MessageBox
from utils.general import check_img_size, check_requirements, check_imshow, colorstr, non_max_suppression, \
    apply_classifier, scale_coords, xyxy2xywh, strip_optimizer, set_logging, increment_path
# from utils.plots import colors, plot_one_box, plot_one_box_PIL
from utils.plots import Annotator, colors, save_one_box

from utils.torch_utils import select_device
from utils.capnums import Camera
from dialog.rtsp_win import Window

def convert_2D_to_3D(point2D, R, t, IntrinsicMatrix, K, P, f, principal_point, height):
    """

    像素坐标转世界坐标
    Args:
        point2D: 像素坐标点
        R: 旋转矩阵
        t: 平移矩阵
        IntrinsicMatrix:内参矩阵
        K:径向畸变
        P:切向畸变
        f:焦距
        principal_point:主点
        height:Z_w

    Returns:返回世界坐标系点,point3D_no_correct, point3D_yes_correct

    """
    point3D_no_correct = []
    point3D_yes_correct = []


    ##[(u1,v1),
   #   (u2,v2)]

    point2D = (np.array(point2D, dtype='float32'))

    # (u,v,1)
    #point2D_op = np.hstack((point2D, np.ones((num_Pts, 1))))
    point2D_op = np.hstack(  (point2D, np.array([1]) )  )
    # R逆矩阵
    rMat_inv = np.linalg.inv(R)
    # 内参矩阵的逆矩阵
    IntrinsicMatrix_inv = np.linalg.inv(IntrinsicMatrix)


    # uvPoint变量切换即可
    uvPoint = point2D_op

    # 畸变矫正后变量
    uvPoint_yes_correct = distortion_correction(point2D, principal_point, f, K, P)
    uvPoint_yes_correct_T = uvPoint_yes_correct.T
    tempMat = np.matmul(rMat_inv, IntrinsicMatrix_inv)
    tempMat1_yes_correct = np.matmul(tempMat, uvPoint_yes_correct_T)#mat1=R^(-1)*K^(-1)([U,V,1].T)
    tempMat2_yes_correct = np.matmul(rMat_inv, t)# Mat2=R^(-1) *T

    s1 = (height + tempMat2_yes_correct[2]) / tempMat1_yes_correct[2] #s1=Zc  height=0
    p1 = tempMat1_yes_correct * s1 - tempMat2_yes_correct.T           #[Xw,Yw,Zw].T  =mat1*zc -mat2
    p_c = np.matmul(R, p1.reshape(-1, 1)) + t.reshape(-1, 1)


    return p1,p_c


def distortion_correction(uvPoint, principal_point, f, K, P):
    """

    畸变矫正函数:畸变发生在图像坐标系转相机坐标系
    Args:
        uvPoint: 坐标点(u,v)
        principal_point: 主点
        f: 焦距
        K: 径向畸变
        P: 切向畸变
    Returns:返回矫正坐标点

    """
    # K:径向畸变系数
    [k1, k2, k3] = K
    # p:切向畸变系数
    [p1, p2] = P

    x = (uvPoint[0] - principal_point[0]) / f[0]
    y = (uvPoint[1] - principal_point[1]) / f[1]

    r = x ** 2 + y ** 2
    x1 = x * (1 + k1 * r + k2 * r ** 2 + k3 * r ** 3) + 2 * p1 * y + p2 * (r + 2 * x ** 2)
    y1 = y * (1 + k1 * r + k2 * r ** 2 + k3 * r ** 3) + 2 * p2 * x + p1 * (r + 2 * y ** 2)

    x_distorted = f[0] * x1 + principal_point[0] + 1
    y_distorted = f[1] * y1 + principal_point[1] + 1

    return np.array([x_distorted, y_distorted, 1])

def calculate_velocity(x1, y1, x2, y2, n, delta_t):
    distance1 = math.sqrt((x2 - x1) ** 2 + (y2 - y1) ** 2)
    time = n * delta_t
    velocity = distance1 / time
    return velocity


class DetThread(QThread):
    send_img = pyqtSignal(np.ndarray)
    send_raw = pyqtSignal(np.ndarray)
    send_statistic = pyqtSignal(dict)
    # emit:detecting/pause/stop/finished/error msg
    send_msg = pyqtSignal(str)
    send_percent = pyqtSignal(int)
    send_fps = pyqtSignal(str)

    def __init__(self):
        super(DetThread, self).__init__()
        self.weights = './yolov5s.pt'
        self.current_weight = './yolov5s.pt'
        self.source = '0'
        self.conf_thres = 0.25
        self.iou_thres = 0.45
        self.jump_out = False                   # jump out of the loop
        self.is_continue = True                 # continue/pause
        self.percent_length = 1000              # progress bar
        self.rate_check = True                  # Whether to enable delay
        self.rate = 100
        self.save_fold = './result'

    @torch.no_grad()
    def run(self,
            imgsz=640,  # inference size (pixels)
            max_det=1000,  # maximum detections per image
            device='',  # cuda device, i.e. 0 or 0,1,2,3 or cpu
            view_img=True,  # show results
            save_txt=False,  # save results to *.txt
            save_conf=False,  # save confidences in --save-txt labels
            save_crop=False,  # save cropped prediction boxes
            nosave=False,  # do not save images/videos
            classes=None,  # filter by class: --class 0, or --class 0 2 3
            agnostic_nms=False,  # class-agnostic NMS
            augment=False,  # augmented inference
            visualize=False,  # visualize features
            update=False,  # update all models
            project='runs/detect',  # save results to project/name
            name='exp',  # save results to project/name
            exist_ok=False,  # existing project/name ok, do not increment
            line_thickness=3,  # bounding box thickness (pixels)
            hide_labels=False,  # hide labels
            hide_conf=False,  # hide confidences
            half=False,  # use FP16 half-precision inference
            ):

        # Initialize
        try:
            device = select_device(device)
            half &= device.type != 'cpu'  # half precision only supported on CUDA

            # Load model
            model = attempt_load(self.weights, map_location=device)  # load FP32 model
            num_params = 0
            for param in model.parameters():
                num_params += param.numel()
            stride = int(model.stride.max())  # model stride
            imgsz = check_img_size(imgsz, s=stride)  # check image size
            names = model.module.names if hasattr(model, 'module') else model.names  # get class names
            if half:
                model.half()  # to FP16

            # Dataloader
            if self.source.isnumeric() or self.source.lower().startswith(('rtsp://', 'rtmp://', 'http://', 'https://')):
                view_img = check_imshow()
                cudnn.benchmark = True  # set True to speed up constant image size inference
                dataset = LoadWebcam(self.source, img_size=imgsz, stride=stride)
                # bs = len(dataset)  # batch_size
            else:
                dataset = LoadImages(self.source, img_size=imgsz, stride=stride)

            # Run inference
            if device.type != 'cpu':
                model(torch.zeros(1, 3, imgsz, imgsz).to(device).type_as(next(model.parameters())))  # run once
            count = 0
            jump_count = 0
            start_time = time.time()
            dataset = iter(dataset)

            while True:
                if self.jump_out:
                    self.vid_cap.release()
                    self.send_percent.emit(0)
                    self.send_msg.emit('Stop')
                    if hasattr(self, 'out'):
                        self.out.release()
                    break
                # change model
                if self.current_weight != self.weights:
                    # Load model
                    model = attempt_load(self.weights, map_location=device)  # load FP32 model
                    num_params = 0
                    for param in model.parameters():
                        num_params += param.numel()
                    stride = int(model.stride.max())  # model stride
                    imgsz = check_img_size(imgsz, s=stride)  # check image size
                    names = model.module.names if hasattr(model, 'module') else model.names  # get class names
                    if half:
                        model.half()  # to FP16
                    # Run inference
                    if device.type != 'cpu':
                        model(torch.zeros(1, 3, imgsz, imgsz).to(device).type_as(next(model.parameters())))  # run once
                    self.current_weight = self.weights
                if self.is_continue:
                    path, img, im0s, self.vid_cap = next(dataset)
                    # jump_count += 1
                    # if jump_count % 5 != 0:
                    #     continue
                    count += 1
                    if count % 30 == 0 and count >= 30:
                        fps = int(30/(time.time()-start_time))
                        self.send_fps.emit('fps:'+str(fps))
                        start_time = time.time()
                    if self.vid_cap:
                        percent = int(count/self.vid_cap.get(cv2.CAP_PROP_FRAME_COUNT)*self.percent_length)
                        self.send_percent.emit(percent)
                    else:
                        percent = self.percent_length

                    statistic_dic = {name: 0 for name in names}
                    img = torch.from_numpy(img).to(device)
                    img = img.half() if 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)

                    pred = model(img, augment=augment)[0]

                    # Apply NMS
                    pred = non_max_suppression(pred, self.conf_thres, self.iou_thres, classes, agnostic_nms, max_det=max_det)
                    # Process detections
                    for i, det in enumerate(pred):  # detections per image
                        im0 = im0s.copy()
                        annotator = Annotator(im0, line_width=line_thickness, example=str(names))
                        if len(det):
                            # Rescale boxes from img_size to im0 size
                            det[:, :4] = scale_coords(img.shape[2:], det[:, :4], im0.shape).round()

                            # Write results
                            for *xyxy, conf, cls in reversed(det):
                                x1 = xyxy[0]
                                y1 = xyxy[1]
                                x2 = xyxy[2]
                                y2 = xyxy[3]
                                INPUT = [(x1 + x2) / 2, y2]
                                p1, p_c = convert_2D_to_3D(INPUT, R, t, IntrinsicMatrix, K, P, f, principal_point, 0)
                                print("-----p1----", p1)
                                d1 = p1[0][1]
                                print("----p_c---", type(p_c))
                                distance = float(p_c[0])
                                c = int(cls)  # integer class
                                statistic_dic[names[c]] += 1
                                #label = None if hide_labels else (names[c] if hide_conf else f'{names[c]} {conf:.2f} ')
                                label = None if hide_labels else (names[c] if hide_conf else f'{names[c]} {conf:.2f} {distance:.2f}m {random.randint(10, 20)}m/s up')
                                annotator.box_label(xyxy, label, color=colors(c, True))

                    if self.rate_check:
                        time.sleep(1/self.rate)
                    im0 = annotator.result()
                    self.send_img.emit(im0)
                    self.send_raw.emit(im0s if isinstance(im0s, np.ndarray) else im0s[0])
                    self.send_statistic.emit(statistic_dic)
                    if self.save_fold:
                        os.makedirs(self.save_fold, exist_ok=True)
                        if self.vid_cap is None:
                            save_path = os.path.join(self.save_fold,
                                                     time.strftime('%Y_%m_%d_%H_%M_%S',
                                                                   time.localtime()) + '.jpg')
                            cv2.imwrite(save_path, im0)
                        else:
                            if count == 1:
                                ori_fps = int(self.vid_cap.get(cv2.CAP_PROP_FPS))
                                if ori_fps == 0:
                                    ori_fps = 25
                                # width = int(self.vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
                                # height = int(self.vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
                                width, height = im0.shape[1], im0.shape[0]
                                save_path = os.path.join(self.save_fold, time.strftime('%Y_%m_%d_%H_%M_%S', time.localtime()) + '.mp4')
                                self.out = cv2.VideoWriter(save_path, cv2.VideoWriter_fourcc(*"mp4v"), ori_fps,
                                                           (width, height))
                            self.out.write(im0)
                    if percent == self.percent_length:
                        print(count)
                        self.send_percent.emit(0)
                        self.send_msg.emit('finished')
                        if hasattr(self, 'out'):
                            self.out.release()
                        break

        except Exception as e:
            self.send_msg.emit('%s' % e)



class MainWindow(QMainWindow, Ui_mainWindow):
    def __init__(self, parent=None):
        super(MainWindow, self).__init__(parent)
        self.setupUi(self)
        self.m_flag = False

        # style 1: window can be stretched
        # self.setWindowFlags(Qt.CustomizeWindowHint | Qt.WindowStaysOnTopHint)

        # style 2: window can not be stretched
        self.setWindowFlags(Qt.Window | Qt.FramelessWindowHint
                            | Qt.WindowSystemMenuHint | Qt.WindowMinimizeButtonHint | Qt.WindowMaximizeButtonHint)
        # self.setWindowOpacity(0.85)  # Transparency of window

        self.minButton.clicked.connect(self.showMinimized)
        self.maxButton.clicked.connect(self.max_or_restore)
        # show Maximized window
        self.maxButton.animateClick(10)
        self.closeButton.clicked.connect(self.close)

        self.qtimer = QTimer(self)
        self.qtimer.setSingleShot(True)
        self.qtimer.timeout.connect(lambda: self.statistic_label.clear())

        # search models automatically
        self.comboBox.clear()
        self.pt_list = os.listdir('./pt')
        self.pt_list = [file for file in self.pt_list if file.endswith('.pt')]
        self.pt_list.sort(key=lambda x: os.path.getsize('./pt/'+x))
        self.comboBox.clear()
        self.comboBox.addItems(self.pt_list)
        self.qtimer_search = QTimer(self)
        self.qtimer_search.timeout.connect(lambda: self.search_pt())
        self.qtimer_search.start(2000)

        # yolov5 thread
        self.det_thread = DetThread()
        self.model_type = self.comboBox.currentText()
        self.det_thread.weights = "./pt/%s" % self.model_type
        self.det_thread.source = '0'
        self.det_thread.percent_length = self.progressBar.maximum()
        self.det_thread.send_raw.connect(lambda x: self.show_image(x, self.raw_video))
        self.det_thread.send_img.connect(lambda x: self.show_image(x, self.out_video))
        self.det_thread.send_statistic.connect(self.show_statistic)
        self.det_thread.send_msg.connect(lambda x: self.show_msg(x))
        self.det_thread.send_percent.connect(lambda x: self.progressBar.setValue(x))
        self.det_thread.send_fps.connect(lambda x: self.fps_label.setText(x))

        self.fileButton.clicked.connect(self.open_file)
        self.cameraButton.clicked.connect(self.chose_cam)
        self.rtspButton.clicked.connect(self.chose_rtsp)

        self.runButton.clicked.connect(self.run_or_continue)
        self.stopButton.clicked.connect(self.stop)

        self.comboBox.currentTextChanged.connect(self.change_model)
        self.confSpinBox.valueChanged.connect(lambda x: self.change_val(x, 'confSpinBox'))
        self.confSlider.valueChanged.connect(lambda x: self.change_val(x, 'confSlider'))
        self.iouSpinBox.valueChanged.connect(lambda x: self.change_val(x, 'iouSpinBox'))
        self.iouSlider.valueChanged.connect(lambda x: self.change_val(x, 'iouSlider'))
        self.rateSpinBox.valueChanged.connect(lambda x: self.change_val(x, 'rateSpinBox'))
        self.rateSlider.valueChanged.connect(lambda x: self.change_val(x, 'rateSlider'))

        self.checkBox.clicked.connect(self.checkrate)
        self.saveCheckBox.clicked.connect(self.is_save)
        self.load_setting()

    def search_pt(self):
        pt_list = os.listdir('./pt')
        pt_list = [file for file in pt_list if file.endswith('.pt')]
        pt_list.sort(key=lambda x: os.path.getsize('./pt/' + x))

        if pt_list != self.pt_list:
            self.pt_list = pt_list
            self.comboBox.clear()
            self.comboBox.addItems(self.pt_list)

    def is_save(self):
        if self.saveCheckBox.isChecked():
            self.det_thread.save_fold = './result'
        else:
            self.det_thread.save_fold = None

    def checkrate(self):
        if self.checkBox.isChecked():
            self.det_thread.rate_check = True
        else:
            self.det_thread.rate_check = False

    def chose_rtsp(self):
        self.rtsp_window = Window()
        config_file = 'config/ip.json'
        if not os.path.exists(config_file):
            ip = "rtsp://admin:admin888@192.168.1.67:555"
            new_config = {"ip": ip}
            new_json = json.dumps(new_config, ensure_ascii=False, indent=2)
            with open(config_file, 'w', encoding='utf-8') as f:
                f.write(new_json)
        else:
            config = json.load(open(config_file, 'r', encoding='utf-8'))
            ip = config['ip']
        self.rtsp_window.rtspEdit.setText(ip)
        self.rtsp_window.show()
        self.rtsp_window.rtspButton.clicked.connect(lambda: self.load_rtsp(self.rtsp_window.rtspEdit.text()))

    def load_rtsp(self, ip):
        try:
            self.stop()
            MessageBox(
                self.closeButton, title='Tips', text='Loading rtsp stream', time=1000, auto=True).exec_()
            self.det_thread.source = ip
            new_config = {"ip": ip}
            new_json = json.dumps(new_config, ensure_ascii=False, indent=2)
            with open('config/ip.json', 'w', encoding='utf-8') as f:
                f.write(new_json)
            self.statistic_msg('Loading rtsp:{}'.format(ip))
            self.rtsp_window.close()
        except Exception as e:
            self.statistic_msg('%s' % e)

    def chose_cam(self):
        try:
            self.stop()
            MessageBox(
                self.closeButton, title='Tips', text='Loading camera', time=2000, auto=True).exec_()
            # get the number of local cameras
            _, cams = Camera().get_cam_num()
            popMenu = QMenu()
            popMenu.setFixedWidth(self.cameraButton.width())
            popMenu.setStyleSheet('''
                                            QMenu {
                                            font-size: 16px;
                                            font-family: "Microsoft YaHei UI";
                                            font-weight: light;
                                            color:white;
                                            padding-left: 5px;
                                            padding-right: 5px;
                                            padding-top: 4px;
                                            padding-bottom: 4px;
                                            border-style: solid;
                                            border-width: 0px;
                                            border-color: rgba(255, 255, 255, 255);
                                            border-radius: 3px;
                                            background-color: rgba(200, 200, 200,50);}
                                            ''')

            for cam in cams:
                exec("action_%s = QAction('%s')" % (cam, cam))
                exec("popMenu.addAction(action_%s)" % cam)

            x = self.groupBox_5.mapToGlobal(self.cameraButton.pos()).x()
            y = self.groupBox_5.mapToGlobal(self.cameraButton.pos()).y()
            y = y + self.cameraButton.frameGeometry().height()
            pos = QPoint(x, y)
            action = popMenu.exec_(pos)
            if action:
                self.det_thread.source = action.text()
                self.statistic_msg('Loading camera:{}'.format(action.text()))
        except Exception as e:
            self.statistic_msg('%s' % e)

    def load_setting(self):
        config_file = 'config/setting.json'
        if not os.path.exists(config_file):
            iou = 0.26
            conf = 0.33
            rate = 10
            check = 0
            savecheck = 0
            new_config = {"iou": iou,
                          "conf": conf,
                          "rate": rate,
                          "check": check,
                          "savecheck": savecheck
                          }
            new_json = json.dumps(new_config, ensure_ascii=False, indent=2)
            with open(config_file, 'w', encoding='utf-8') as f:
                f.write(new_json)
        else:
            config = json.load(open(config_file, 'r', encoding='utf-8'))
            if len(config) != 5:
                iou = 0.26
                conf = 0.33
                rate = 10
                check = 0
                savecheck = 0
            else:
                iou = config['iou']
                conf = config['conf']
                rate = config['rate']
                check = config['check']
                savecheck = config['savecheck']
        self.confSpinBox.setValue(conf)
        self.iouSpinBox.setValue(iou)
        self.rateSpinBox.setValue(rate)
        self.checkBox.setCheckState(check)
        self.det_thread.rate_check = check
        self.saveCheckBox.setCheckState(savecheck)
        self.is_save()

    def change_val(self, x, flag):
        if flag == 'confSpinBox':
            self.confSlider.setValue(int(x*100))
        elif flag == 'confSlider':
            self.confSpinBox.setValue(x/100)
            self.det_thread.conf_thres = x/100
        elif flag == 'iouSpinBox':
            self.iouSlider.setValue(int(x*100))
        elif flag == 'iouSlider':
            self.iouSpinBox.setValue(x/100)
            self.det_thread.iou_thres = x/100
        elif flag == 'rateSpinBox':
            self.rateSlider.setValue(x)
        elif flag == 'rateSlider':
            self.rateSpinBox.setValue(x)
            self.det_thread.rate = x * 10
        else:
            pass

    def statistic_msg(self, msg):
        self.statistic_label.setText(msg)
        # self.qtimer.start(3000)

    def show_msg(self, msg):
        self.runButton.setChecked(Qt.Unchecked)
        self.statistic_msg(msg)
        if msg == "Finished":
            self.saveCheckBox.setEnabled(True)

    def change_model(self, x):
        self.model_type = self.comboBox.currentText()
        self.det_thread.weights = "./pt/%s" % self.model_type
        self.statistic_msg('Change model to %s' % x)

    def open_file(self):

        config_file = 'config/fold.json'
        # config = json.load(open(config_file, 'r', encoding='utf-8'))
        config = json.load(open(config_file, 'r', encoding='utf-8'))
        open_fold = config['open_fold']
        if not os.path.exists(open_fold):
            open_fold = os.getcwd()
        name, _ = QFileDialog.getOpenFileName(self, 'Video/image', open_fold, "Pic File(*.mp4 *.mkv *.avi *.flv "
                                                                          "*.jpg *.png)")
        if name:
            self.det_thread.source = name
            self.statistic_msg('Loaded file:{}'.format(os.path.basename(name)))
            config['open_fold'] = os.path.dirname(name)
            config_json = json.dumps(config, ensure_ascii=False, indent=2)
            with open(config_file, 'w', encoding='utf-8') as f:
                f.write(config_json)
            self.stop()

    def max_or_restore(self):
        if self.maxButton.isChecked():
            self.showMaximized()
        else:
            self.showNormal()

    def run_or_continue(self):
        self.det_thread.jump_out = False
        if self.runButton.isChecked():
            self.saveCheckBox.setEnabled(False)
            self.det_thread.is_continue = True
            if not self.det_thread.isRunning():
                self.det_thread.start()
            source = os.path.basename(self.det_thread.source)
            source = 'camera' if source.isnumeric() else source
            self.statistic_msg('Detecting >> model:{},file:{}'.
                               format(os.path.basename(self.det_thread.weights),
                                      source))
        else:
            self.det_thread.is_continue = False
            self.statistic_msg('Pause')

    def stop(self):
        self.det_thread.jump_out = True
        self.saveCheckBox.setEnabled(True)

    def mousePressEvent(self, event):
        self.m_Position = event.pos()
        if event.button() == Qt.LeftButton:
            if 0 < self.m_Position.x() < self.groupBox.pos().x() + self.groupBox.width() and \
                    0 < self.m_Position.y() < self.groupBox.pos().y() + self.groupBox.height():
                self.m_flag = True

    def mouseMoveEvent(self, QMouseEvent):
        if Qt.LeftButton and self.m_flag:
            self.move(QMouseEvent.globalPos() - self.m_Position)

    def mouseReleaseEvent(self, QMouseEvent):
        self.m_flag = False

    @staticmethod
    def show_image(img_src, label):
        try:
            ih, iw, _ = img_src.shape
            w = label.geometry().width()
            h = label.geometry().height()
            # keep original aspect ratio
            if iw/w > ih/h:
                scal = w / iw
                nw = w
                nh = int(scal * ih)
                img_src_ = cv2.resize(img_src, (nw, nh))

            else:
                scal = h / ih
                nw = int(scal * iw)
                nh = h
                img_src_ = cv2.resize(img_src, (nw, nh))

            frame = cv2.cvtColor(img_src_, cv2.COLOR_BGR2RGB)
            img = QImage(frame.data, frame.shape[1], frame.shape[0], frame.shape[2] * frame.shape[1],
                         QImage.Format_RGB888)
            label.setPixmap(QPixmap.fromImage(img))

        except Exception as e:
            print(repr(e))

    def show_statistic(self, statistic_dic):
        try:
            self.resultWidget.clear()
            statistic_dic = sorted(statistic_dic.items(), key=lambda x: x[1], reverse=True)
            statistic_dic = [i for i in statistic_dic if i[1] > 0]
            results = [' '+str(i[0]) + ':' + str(i[1]) for i in statistic_dic]
            self.resultWidget.addItems(results)

        except Exception as e:
            print(repr(e))

    def closeEvent(self, event):
        self.det_thread.jump_out = True
        config_file = 'config/setting.json'
        config = dict()
        config['iou'] = self.confSpinBox.value()
        config['conf'] = self.iouSpinBox.value()
        config['rate'] = self.rateSpinBox.value()
        config['check'] = self.checkBox.checkState()
        config['savecheck'] = self.saveCheckBox.checkState()
        config_json = json.dumps(config, ensure_ascii=False, indent=2)
        with open(config_file, 'w', encoding='utf-8') as f:
            f.write(config_json)
        MessageBox(
            self.closeButton, title='Tips', text='Closing the program', time=2000, auto=True).exec_()
        sys.exit(0)


if __name__ == "__main__":
    R = np.array([[9.1119371736959609e-01, -2.4815760576991752e-02, -4.1123009064654115e-01],
                  [4.1105811256386449e-01, -1.1909647756530584e-02, 9.1153134251420498e-01],
                  [-2.7517949080742898e-02, -9.9962109737505089e-01, -6.5127650722056341e-04]])
    R = R.T

    # 平移向量
    # t = np.array([[-730.2794],
    #               [290.2519],
    #               [688.4792]])
    t = np.array([[1.0966499328613281e+01],
                  [-4.1683087348937988e+00],
                  [8.7983322143554688e-01]])
    # 内参矩阵,转置
    # IntrinsicMatrix = np.array([[423.0874, 0, 0],
    #                             [0, 418.7552, 0],
    #                             [652.5402, 460.2077, 1]])

    IntrinsicMatrix = np.array([[1.9770188633212194e+03, 0., 1.0126938349335526e+03],
                                [0., 1.9668641721787440e+03, 4.7095156301902404e+02],
                                [0., 0., 1.]])
    IntrinsicMatrix = IntrinsicMatrix.T

    # 焦距
    f = [1.9770188633212194e+03, 1.9668641721787440e+03]
    # 主点
    principal_point = [1.0126938349335526e+03, 4.7095156301902404e+02]

    # 径向畸变矩阵
    # K = [-0.3746, 0.1854, -0.0514]
    K = [1.0966499328613281e+01,
         -4.1683087348937988e+00,
         8.7983322143554688e-01]
    # 切向畸变矩阵
    # P = [0.0074, -0.0012]
    P = [-2.4283340903321522e-03,
         3.1736917344022848e-02]
    app = QApplication(sys.argv)
    myWin = MainWindow()
    myWin.show()
    # myWin.showMaximized()
    sys.exit(app.exec_())
相关推荐
音沐mu1 个月前
【基于YOLOv5的反光衣检测预警系统】可检测图片、视频、摄像头,支持GPU加速检测以及语音播报预警
人工智能·yolo·目标检测·yolov5·反光衣检测
zh路西法1 个月前
基于opencv-C++dnn模块推理的yolov5 onnx模型
c++·图像处理·pytorch·opencv·yolo·dnn·yolov5
从懒虫到爬虫1 个月前
变压器设备漏油数据集 voc txt
数据集·qq767172261·变压器设备漏油数据集
从懒虫到爬虫1 个月前
基于YOLOv8+LSTM的商超扶梯场景下行人安全行为姿态检测识别
lstm·姿态识别·qq767172261
从懒虫到爬虫1 个月前
变电站缺陷数据集8307张,带xml标注和txt标注,可以直接用于yolo训练
数据集·qq767172261·变电站缺陷数据集
从懒虫到爬虫1 个月前
无人机视角应急救援(人)数据集
数据集·无人机视角·应急救援·qq767172261
从懒虫到爬虫1 个月前
密集行人数据集 CrowdHumanvoc和yolo两种格式,yolo可以直接使用train val test已经划分好有yolov8训练200轮模型
数据集·qq767172261·密集行人
从懒虫到爬虫2 个月前
无人机视角下的车辆数据集
数据集·车辆检测·无人机视角·qq767172261·车辆数据集
从懒虫到爬虫2 个月前
电力施工作业安全行为检测图像数据集
数据集·qq767172261·电力行业施工数据集·行为检测
从懒虫到爬虫2 个月前
遥感图像目标检测数据集-DOTA数据集
数据集·qq767172261·dota数据集