yolov3目标检测可视化框架——火焰识别为例

yolov3目标检测可视化框架（轻量级）——以火焰识别为例

---

a.其中主要包含有图片检测、权重加载、可信度调整、检测结果保存，日志输出功能（可扩充）
b.文件中的权重和数据集可以更换，训练介绍后不影响GUI的使用

一、功能介绍

1.主界面布局

2.权重加载

3.读取文件 -> 目标检测

4.更改可信度 -> 目标检测

5.保存图片

---

二、更换检测数据

可以参考：https://blog.csdn.net/qq_21578849/article/details/84980298

三、源码

def open_weight(self):
    self.weightName, weightType = QFileDialog.getOpenFileName(self, "打开权重文件", "", "*.pth;;All Files(*)")
    self.lineEdit.setText(str(self.weightName))
    self.txt += f'[+]权重文件加载成功!地址:{self.weightName}\n'
    self.textEdit.setPlainText(self.txt)
    self.pushButton.setEnabled(True)
def status_msg(self, msg, time):
    self.statusBar.showMessage(str(msg), int(time))
    QApplication.processEvents()
def open_image(self):
    self.imgName, imgType = QFileDialog.getOpenFileName(self, "打开图片", "", "*.jpg;;*.png;;All Files(*)")
    imgName_cv2 = cv2.imread(self.imgName)
    imgName_cv2 = cv2.resize(imgName_cv2, (450, 360))
    try:
        if imgName_cv2.all != None:
            im0 = cv2.cvtColor(imgName_cv2, cv2.COLOR_RGB2BGR)
            showImage = QtGui.QImage(im0, im0.shape[1], im0.shape[0], 3 * im0.shape[1], QtGui.QImage.Format_RGB888)
            self.label.setPixmap(QtGui.QPixmap.fromImage(showImage))
            self.pushButton_3.setEnabled(True)
            self.txt+=f'[+]图片加载成功!地址:{self.imgName}\n'
            self.textEdit.setPlainText(self.txt)
    except Exception as e:
        print(str(e))
        self.status_msg('[-]不能含有中午路径'+str(e),5000)
        QApplication.processEvents()
def cvImgtoQtImg(self,cvImg):
    QtImgBuf = cv2.cvtColor(cvImg, cv2.COLOR_BGR2BGRA)
    QtImg = QtGui.QImage(QtImgBuf.data, QtImgBuf.shape[1], QtImgBuf.shape[0], QtGui.QImage.Format_RGB32)
    return QtImg

---

源文件下载：传送门