agv/vision
8
0
Форк 0
Код Задачі Запити на злиття Actions Packages Проєкти Релізи Вікі Активність

origin

Переглянути джерело
This commit is contained in:
ty
2025-04-06 17:27:22 +08:00
зафіксовано wuxianfu
коміт 7a4cb1fef1
8 змінених файлів з 742 додано та 0 видалено
Завантажити патч Завантажити файл різниці Expand all files Collapse all files

302
YOLOPv2.cpp Normal file
Переглянути файл

@@ -0,0 +1,302 @@
#include "YOLOPv2.h"
#include <QLoggingCategory>
YOLOPv2::YOLOPv2(Net_config config)
{
this->confThreshold = config.confThreshold;
this->nmsThreshold = config.nmsThreshold;
//string model_path = config.modelpath;
//std::wstring widestr = std::wstring(model_path.begin(), model_path.end());
//CUDA option set
OrtCUDAProviderOptions cuda_option;
cuda_option.device_id = 0;
cuda_option.arena_extend_strategy = 0;
cuda_option.cudnn_conv_algo_search = OrtCudnnConvAlgoSearchExhaustive;
cuda_option.gpu_mem_limit = SIZE_MAX;
cuda_option.do_copy_in_default_stream = 1;
//CUDA 加速
sessionOptions.SetIntraOpNumThreads(1);//设置线程数
sessionOptions.SetGraphOptimizationLevel(GraphOptimizationLevel::ORT_ENABLE_ALL); //函数用于设置在使用 ORT 库执行模型时应用的图优化级别。ORT_ENABLE_ALL 选项启用所有可用的优化，这可以导致更快速和更高效的执行。
sessionOptions.AppendExecutionProvider_CUDA(cuda_option);
const char *modelpath = "/home/wuxianfu/Projects/Fast-YolopV2/build/onnx/yolopv2_192x320.onnx" ;
ort_session = new Session(env, modelpath, sessionOptions);
size_t numInputNodes = ort_session->GetInputCount();
size_t numOutputNodes = ort_session->GetOutputCount();
AllocatorWithDefaultOptions allocator;
for (int i = 0; i < numInputNodes; i++)
{
//input_names.push_back(ort_session->GetInputName(i, allocator));
AllocatedStringPtr input_name_Ptr = ort_session->GetInputNameAllocated(i, allocator);
input_names.push_back(input_name_Ptr.get());
qDebug() << "Input Name: " << input_name_Ptr.get();
Ort::TypeInfo input_type_info = ort_session->GetInputTypeInfo(i);
auto input_tensor_info = input_type_info.GetTensorTypeAndShapeInfo();
auto input_dims = input_tensor_info.GetShape();
input_node_dims.push_back(input_dims);
}
for (int i = 0; i < numOutputNodes; i++)
{
//output_names.push_back(ort_session->GetOutputName(i, allocator));
AllocatedStringPtr output_name_Ptr= ort_session->GetOutputNameAllocated(i, allocator);
output_names.push_back(output_name_Ptr.get());
qDebug() << "Output Name: " << output_name_Ptr.get();
Ort::TypeInfo output_type_info = ort_session->GetOutputTypeInfo(i);
auto output_tensor_info = output_type_info.GetTensorTypeAndShapeInfo();
auto output_dims = output_tensor_info.GetShape();
output_node_dims.push_back(output_dims);
}
this->inpHeight = input_node_dims[0][2];
this->inpWidth = input_node_dims[0][3];
string classesFile = "/home/wuxianfu/Projects/Fast-YolopV2/build/coco.names";
ifstream ifs(classesFile.c_str());
string line;
while (getline(ifs, line)) this->class_names.push_back(line);
this->num_class = class_names.size();
}
void YOLOPv2::normalize_(Mat img)
{
// img.convertTo(img, CV_32F);
int row = img.rows;
int col = img.cols;
this->input_image_.resize(row * col * img.channels());
for (int c = 0; c < 3; c++)
{
for (int i = 0; i < row; i++)
{
for (int j = 0; j < col; j++)
{
float pix = img.ptr<uchar>(i)[j * 3 + 2 - c];
this->input_image_[c * row * col + i * col + j] = pix / 255.0;
}
}
}
}
void YOLOPv2::nms(vector<BoxInfo>& input_boxes)
{
sort(input_boxes.begin(), input_boxes.end(), [](BoxInfo a, BoxInfo b) { return a.score > b.score; });
vector<float> vArea(input_boxes.size());
for (int i = 0; i < int(input_boxes.size()); ++i)
{
vArea[i] = (input_boxes.at(i).x2 - input_boxes.at(i).x1 + 1)
* (input_boxes.at(i).y2 - input_boxes.at(i).y1 + 1);
}
vector<bool> isSuppressed(input_boxes.size(), false);
for (int i = 0; i < int(input_boxes.size()); ++i)
{
if (isSuppressed[i]) { continue; }
for (int j = i + 1; j < int(input_boxes.size()); ++j)
{
if (isSuppressed[j]) { continue; }
float xx1 = (max)(input_boxes[i].x1, input_boxes[j].x1);
float yy1 = (max)(input_boxes[i].y1, input_boxes[j].y1);
float xx2 = (min)(input_boxes[i].x2, input_boxes[j].x2);
float yy2 = (min)(input_boxes[i].y2, input_boxes[j].y2);
float w = (max)(float(0), xx2 - xx1 + 1);
float h = (max)(float(0), yy2 - yy1 + 1);
float inter = w * h;
float ovr = inter / (vArea[i] + vArea[j] - inter);
if (ovr >= this->nmsThreshold)
{
isSuppressed[j] = true;
}
}
}
// return post_nms;
int idx_t = 0;
input_boxes.erase(remove_if(input_boxes.begin(), input_boxes.end(), [&idx_t, &isSuppressed](const BoxInfo& f) { return isSuppressed[idx_t++]; }), input_boxes.end());
}
inline float sigmoid(float x)
{
return 1.0 / (1 + exp(-x));
}
Mat YOLOPv2::detect(Mat& frame)
{
Mat dstimg;
resize(frame, dstimg, Size(this->inpWidth, this->inpHeight));
this->normalize_(dstimg);
array<int64_t, 4> input_shape_{ 1, 3, this->inpHeight, this->inpWidth };
auto allocator_info = MemoryInfo::CreateCpu(OrtDeviceAllocator, OrtMemTypeDefault);
Value input_tensor_ = Value::CreateTensor<float>(allocator_info, input_image_.data(), input_image_.size(), input_shape_.data(), input_shape_.size());
// 开始推理
/*qDebug() << " output_names size: " << output_names.size()<< " sec \n";
qDebug() << " input_names: " << input_names[0]<< " sec \n";
qDebug() << " output_names: " << output_names[1]<< " sec \n";
vector<Value> ort_outputs = ort_session->Run(RunOptions{nullptr}, input_names.data(), &input_tensor_, 1, output_names.data(), output_names.size());*/
const char* inputNames[] = { "input" };//这两个值是根据netron查看onnx格式得到的输入输出名称
const char* outputNames[] = { "seg" , "ll" , "pred0" , "pred1" , "pred2" , };
vector<Value> ort_outputs = ort_session->Run(RunOptions{nullptr}, inputNames, &input_tensor_, 1, outputNames, 5);
/////generate proposals
vector<BoxInfo> generate_boxes;
float ratioh = (float)frame.rows / this->inpHeight, ratiow = (float)frame.cols / this->inpWidth;
int n = 0, q = 0, i = 0, j = 0, nout = this->class_names.size() + 5, c = 0, area = 0;
for (n = 0; n < 3; n++) ///尺度
{
int num_grid_x = (int)(this->inpWidth / this->stride[n]);
int num_grid_y = (int)(this->inpHeight / this->stride[n]);
area = num_grid_x * num_grid_y;
const float* pdata = ort_outputs[n + 2].GetTensorMutableData<float>();
for (q = 0; q < 3; q++) ///anchor数
{
const float anchor_w = this->anchors[n][q * 2];
const float anchor_h = this->anchors[n][q * 2 + 1];
for (i = 0; i < num_grid_y; i++)
{
for (j = 0; j < num_grid_x; j++)
{
float box_score = sigmoid(pdata[4 * area + i * num_grid_x + j]);
if (box_score > this->confThreshold)
{
float max_class_socre = -100000, class_socre = 0;
int max_class_id = 0;
for (c = 0; c < this->class_names.size(); c++) //// get max socre
{
class_socre = pdata[(c + 5) * area + i * num_grid_x + j];
if (class_socre > max_class_socre)
{
max_class_socre = class_socre;
max_class_id = c;
}
}
max_class_socre = sigmoid(max_class_socre) * box_score;
if (max_class_socre > this->confThreshold)
{
float cx = (sigmoid(pdata[i * num_grid_x + j]) * 2.f - 0.5f + j) * this->stride[n]; ///cx
float cy = (sigmoid(pdata[area + i * num_grid_x + j]) * 2.f - 0.5f + i) * this->stride[n]; ///cy
float w = powf(sigmoid(pdata[2 * area + i * num_grid_x + j]) * 2.f, 2.f) * anchor_w; ///w
float h = powf(sigmoid(pdata[3 * area + i * num_grid_x + j]) * 2.f, 2.f) * anchor_h; ///h
float xmin = (cx - 0.5*w)*ratiow;
float ymin = (cy - 0.5*h)*ratioh;
float xmax = (cx + 0.5*w)*ratiow;
float ymax = (cy + 0.5*h)*ratioh;
generate_boxes.push_back(BoxInfo{ xmin, ymin, xmax, ymax, max_class_socre, max_class_id });
}
}
}
}
pdata += area * nout;
}
}
nms(generate_boxes);
Mat outimg = frame.clone();
for (size_t i = 0; i < generate_boxes.size(); ++i)
{
int xmin = int(generate_boxes[i].x1);
int ymin = int(generate_boxes[i].y1);
rectangle(outimg, Point(xmin, ymin), Point(int(generate_boxes[i].x2), int(generate_boxes[i].y2)), Scalar(0, 0, 255), 2);
string label = format("%.2f", generate_boxes[i].score);
label = this->class_names[generate_boxes[i].label-1] + ":" + label;
putText(outimg, label, Point(xmin, ymin - 5), FONT_HERSHEY_SIMPLEX, 0.75, Scalar(0, 255, 0), 1);
}
const float* pdrive_area = ort_outputs[0].GetTensorMutableData<float>();
const float* plane_line = ort_outputs[1].GetTensorMutableData<float>();
area = this->inpHeight*this->inpWidth;
int min_y = -1;
vector<Point2f> points_L, points_R;
for (i = 0; i < frame.rows; i++)
{
bool flg = false;
int left = -1, right = -1;
for (j = 0; j < frame.cols; j++)
{
const int x = int(j / ratiow);
const int y = int(i / ratioh);
if (pdrive_area[y * this->inpWidth + x] < pdrive_area[area + y * this->inpWidth + x])
{
outimg.at<Vec3b>(i, j)[0] = 0;
outimg.at<Vec3b>(i, j)[1] = 255;
outimg.at<Vec3b>(i, j)[2] = 0;
}
if (plane_line[y * this->inpWidth + x] > 0.5)
{
outimg.at<Vec3b>(i, j)[0] = 255;
outimg.at<Vec3b>(i, j)[1] = 0;
outimg.at<Vec3b>(i, j)[2] = 0;
if (!flg && j >= frame.cols / 2 && right == -1) { // 记录图像右半部分最靠左的车道线的左边缘坐标
right = j;
}
flg = true;
} else {
if (flg && j - 1 < frame.cols / 2) { //记录图像左半部分最靠右的车道线的右边缘坐标
left = j - 1;
}
flg = false;
}
}
if (min_y == -1 && (left != -1 || right != -1)) {
min_y = i;
}
if (left != -1){
points_L.push_back(Point2f(left, i));
}
if (right != -1){
points_R.push_back(Point2f(right, i));
}
//若左右参考车道线均存在，计算并标记中心点
if (left > -1 && right > -1) {
int mid = (left + right) / 2;
for (int k = -5; k <= 5; k++) {
outimg.at<Vec3b>(i, mid+k)[0] = 255;
outimg.at<Vec3b>(i, mid+k)[1] = 255;
outimg.at<Vec3b>(i, mid+k)[2] = 0;
}
}
//(需要考虑的问题 1.双车道3条线 2.拐角处曲线 3.近处显示不全 4.两条线粘连)
}
//备选方案，对左右车道线分别拟合直线并计算中心线解析式 泛化 鲁棒 （目前有bug
if (points_L.size() && points_R.size()) {
Vec4f line_L, line_R;
float kL, bL, kR, bR, kM, bM; // x=ky+b
fitLine(points_L, line_L, DIST_WELSCH, 0, 0.01, 0.01);
fitLine(points_R, line_R, DIST_WELSCH, 0, 0.01, 0.01);
kL = line_L[0] / line_L[1];
bL = line_L[2] - kL * line_L[3];
kR = line_R[0] / line_R[1];
bR = line_R[2] - kR * line_R[3];
kM = (kL + kR) / 2;
bM = (bL + bR) / 2;
for (int i = min_y; i < frame.rows; i++) {
int mid = round(kM * i + bM);
for (int k = -5; k <= 5; k++) {
outimg.at<Vec3b>(i, mid+k)[0] = 255;
outimg.at<Vec3b>(i, mid+k)[1] = 0;
outimg.at<Vec3b>(i, mid+k)[2] = 255;
}
}
}
return outimg;
}

62
YOLOPv2.h Normal file
Переглянути файл

@@ -0,0 +1,62 @@
#ifndef YOLOPV2_H
#define YOLOPV2_H
#include <fstream>
#include <sstream>
#include <iostream>
#include <opencv2/imgproc.hpp>
#include <opencv2/highgui.hpp>
#include <onnxruntime_cxx_api.h>
using namespace cv;
using namespace Ort;
using namespace std;
struct Net_config
{
float confThreshold; // Confidence threshold
float nmsThreshold; // Non-maximum suppression threshold
string modelpath;
};
typedef struct BoxInfo
{
float x1;
float y1;
float x2;
float y2;
float score;
int label;
} BoxInfo;
class YOLOPv2
{
public:
YOLOPv2(Net_config config);
Mat detect(Mat& frame);
private:
int inpWidth;
int inpHeight;
int nout;
int num_proposal;
vector<string> class_names;
int num_class;
float confThreshold;
float nmsThreshold;
vector<float> input_image_;
void normalize_(Mat img);
void nms(vector<BoxInfo>& input_boxes);
const float anchors[3][6] = { {12, 16, 19, 36, 40, 28}, {36, 75, 76, 55, 72, 146},{142, 110, 192, 243, 459, 401} };
const float stride[3] = { 8.0, 16.0, 32.0 };
Env env = Env(ORT_LOGGING_LEVEL_ERROR, "YOLOPv2");
Ort::Session *ort_session = nullptr;
SessionOptions sessionOptions = SessionOptions();
vector<char*> input_names;
vector<char*> output_names;
vector<vector<int64_t>> input_node_dims; // >=1 outputs
vector<vector<int64_t>> output_node_dims; // >=1 outputs
};
#endif // YOLOPV2_H

61
build/ui_mainwindow.h Normal file
Переглянути файл

@@ -0,0 +1,61 @@
/********************************************************************************
** Form generated from reading UI file 'mainwindow.ui'
**
** Created by: Qt User Interface Compiler version 5.15.3
**
** WARNING! All changes made in this file will be lost when recompiling UI file!
********************************************************************************/
#ifndef UI_MAINWINDOW_H
#define UI_MAINWINDOW_H
#include <QtCore/QVariant>
#include <QtWidgets/QApplication>
#include <QtWidgets/QMainWindow>
#include <QtWidgets/QMenuBar>
#include <QtWidgets/QStatusBar>
#include <QtWidgets/QWidget>
QT_BEGIN_NAMESPACE
class Ui_MainWindow
{
public:
QWidget *centralwidget;
QMenuBar *menubar;
QStatusBar *statusbar;
void setupUi(QMainWindow *MainWindow)
{
if (MainWindow->objectName().isEmpty())
MainWindow->setObjectName(QString::fromUtf8("MainWindow"));
MainWindow->resize(800, 600);
centralwidget = new QWidget(MainWindow);
centralwidget->setObjectName(QString::fromUtf8("centralwidget"));
MainWindow->setCentralWidget(centralwidget);
menubar = new QMenuBar(MainWindow);
menubar->setObjectName(QString::fromUtf8("menubar"));
MainWindow->setMenuBar(menubar);
statusbar = new QStatusBar(MainWindow);
statusbar->setObjectName(QString::fromUtf8("statusbar"));
MainWindow->setStatusBar(statusbar);
retranslateUi(MainWindow);
QMetaObject::connectSlotsByName(MainWindow);
} // setupUi
void retranslateUi(QMainWindow *MainWindow)
{
MainWindow->setWindowTitle(QCoreApplication::translate("MainWindow", "MainWindow", nullptr));
} // retranslateUi
};
namespace Ui {
class MainWindow: public Ui_MainWindow {};
} // namespace Ui
QT_END_NAMESPACE
#endif // UI_MAINWINDOW_H

51
fast-yolopv2.pro Normal file
Переглянути файл

@@ -0,0 +1,51 @@
QT += core gui widgets
greaterThan(QT_MAJOR_VERSION, 4): QT += widgets
CONFIG += c++11
DEFINES += OPENCV
DEFINES += GPU
DEFINES += CUDNN
# You can make your code fail to compile if it uses deprecated APIs.
# In order to do so, uncomment the following line.
#DEFINES += QT_DISABLE_DEPRECATED_BEFORE=0x060000 # disables all the APIs deprecated before Qt 6.0.0
# 包含 OpenCV 的头文件路径
INCLUDEPATH += /usr/local/Opencv-4.10.0/include/opencv4
# 包含 Cuda 的头文件路径
INCLUDEPATH += /usr/local/cuda-12.6/include
# 包含 onnx 的头文件路径
INCLUDEPATH += /usr/local/include/onnxruntime
# 链接到 OpenCV 的库文件
LIBS += -L/usr/local/Opencv-4.10.0/lib \
-lopencv_core \
-lopencv_imgproc \
-lopencv_highgui \
-lopencv_imgcodecs \
-lopencv_videoio
# 链接到 onnx 的库文件
LIBS += -L/usr/local/lib -lonnxruntime
SOURCES += \
YOLOPv2.cpp \
main.cpp \
mainwindow.cpp
HEADERS += \
YOLOPv2.h \
mainwindow.h
FORMS += \
mainwindow.ui
# Default rules for deployment.
qnx: target.path = /tmp/$${TARGET}/bin
else: unix:!android: target.path = /opt/$${TARGET}/bin
!isEmpty(target.path): INSTALLS += target

11
main.cpp Normal file
Переглянути файл

@@ -0,0 +1,11 @@
#include "mainwindow.h"
#include <QApplication>
int main(int argc, char *argv[])
{
QApplication a(argc, argv);
MainWindow w;
w.show();
return a.exec();
}

201
mainwindow.cpp Normal file
Переглянути файл

@@ -0,0 +1,201 @@
#include "mainwindow.h"
#include "ui_mainwindow.h"
#include "YOLOPv2.h"
#include <string>
#include <QFile>
#include <QLoggingCategory>
string gstreamer_pipeline (int capture_width, int capture_height, int display_width, int display_height, int framerate, int flip_method)
{
return "nvarguscamerasrc ! video/x-raw(memory:NVMM), width=(int)" + to_string(capture_width) + ", height=(int)" +
to_string(capture_height) + ", format=(string)NV12, framerate=(fraction)" + to_string(framerate) +
"/1 ! nvvidconv flip-method=" + to_string(flip_method) + " ! video/x-raw, width=(int)" + to_string(display_width) + ", height=(int)" +
to_string(display_height) + ", format=(string)BGRx ! videoconvert ! video/x-raw, format=(string)BGR ! appsink";
}
MainWindow::MainWindow(QWidget *parent)
: QMainWindow(parent)
, ui(new Ui::MainWindow)
{
ui->setupUi(this);
//ShowImage();
ShowVideo();
//OpenCSICamera();
}
MainWindow::~MainWindow()
{
delete ui;
}
void MainWindow::ShowImage()
{
Net_config YOLOPv2_nets = { 0.5, 0.5, "/home/wuxianfu/Projects/Fast-YolopV2/build/onnx/yolopv2_192x320.onnx" }; ////choices = onnx文件夹里的文件
YOLOPv2 net(YOLOPv2_nets);
string imgpath = "/home/wuxianfu/Projects/Fast-YolopV2/build/images/3c0e7240-96e390d2.jpg";
static const string kWinName = "Deep learning object detection in ONNXRuntime";
namedWindow(kWinName, WINDOW_NORMAL);
Mat srcimg = imread(imgpath);
imshow(kWinName, srcimg);
Mat outimg = net.detect(srcimg);
imshow(kWinName, outimg);
waitKey(0);
destroyAllWindows();
}
void MainWindow::ShowVideo()
{
Net_config YOLOPv2_nets = { 0.5, 0.5, "/home/wuxianfu/Projects/Fast-YolopV2/build/onnx/yolopv2_736x1280.onnx" }; ////choices = onnx文件夹里的文件
YOLOPv2 net(YOLOPv2_nets);
static const string kWinName = "Deep learning object detection in ONNXRuntime";
namedWindow(kWinName, WINDOW_NORMAL);
cv::VideoCapture cap("/home/wuxianfu/Projects/Fast-YolopV2/build/videos/566a351c29c00924a337e91e85fa7dec.mp4");
if (!cap.isOpened()) {
std::cerr << "Error opening video stream" << std::endl;
return;
}
cv::Mat frame;
while (true) {
cap >> frame; // 读取一帧
if (frame.empty()) {
std::cerr << "Error reading frame" << std::endl;
break;
}
auto start = std::chrono::steady_clock::now();
Mat outimg = net.detect(frame);
auto end = std::chrono::steady_clock::now();
std::chrono::duration<double> spent = end - start;
qDebug()<< " Time: " << spent.count() << " sec \n";
imshow(kWinName, outimg);
if (cv::waitKey(5) >= 0) break; // 按任意键退出循环
}
cap.release(); // 释放资源
cv::destroyAllWindows(); // 关闭所有OpenCV窗口
}
void MainWindow::OpenUSBCamera() {
Net_config YOLOPv2_nets = { 0.5, 0.5, "/home/wuxianfu/Projects/Fast-YolopV2/build/onnx/yolopv2_192x320.onnx" }; ////choices = onnx文件夹里的文件
YOLOPv2 net(YOLOPv2_nets);
cv::VideoCapture cap(1); // 使用默认的摄像头索引（通常是0）
if (!cap.isOpened()) {
std::cerr << "Error opening video stream" << std::endl;
return;
}
cv::Mat frame;
while (true) {
cap >> frame; // 读取一帧
if (frame.empty()) {
std::cerr << "Error reading frame" << std::endl;
break;
}
Mat outimg = net.detect(frame);
cv::imshow("USB Camera", outimg); // 显示帧
if (cv::waitKey(10) >= 0) break; // 按任意键退出循环
}
cap.release(); // 释放资源
cv::destroyAllWindows(); // 关闭所有OpenCV窗口
}
void MainWindow::OpenCSICamera() {
Net_config YOLOPv2_nets = { 0.5, 0.5, "/home/wuxianfu/Projects/Fast-YolopV2/build/onnx/yolopv2_736x1280.onnx" }; ////choices = onnx文件夹里的文件
YOLOPv2 net(YOLOPv2_nets);
string imgpath = "/home/wuxianfu/Projects/Fast-YolopV2/build/images/0ace96c3-48481887.jpg";
Mat srcimg = imread(imgpath);
int capture_width = 3280 ;
int capture_height = 1848 ;
int display_width = 3280 ;
int display_height = 1848 ;
//3280*2464最大支持21帧
//3280*1848最大支持28帧
//1920*1080最大支持29帧
//1640*1232最大支持29帧
//1280*720 最大支持59帧
int framerate = 10;
int flip_method = 0;
//创建管道
string pipeline = gstreamer_pipeline(capture_width,
capture_height,
display_width,
display_height,
framerate,
flip_method);
std::cout << "使用gstreamer管道: \n\t" << pipeline << "\n";
//管道与视频流绑定
VideoCapture cap(pipeline, CAP_GSTREAMER);
if(!cap.isOpened())
{
std::cout<<"打开摄像头失败."<<std::endl;
return ;
}
qDebug()<<"打开摄像头成功.";
//创建显示窗口
namedWindow("CSI Camera", WINDOW_AUTOSIZE);
Mat img;
//逐帧显示
while(true)
{
qDebug()<<"开始捕获摄像头.";
auto start = std::chrono::steady_clock::now();
if (!cap.read(img))
{
std::cout<<"捕获失败"<<std::endl;
break;
}
int new_width,new_height,width,height;
width=img.cols;
height=img.rows;
//调整图像大小
new_width=1000;
if(width>800)
{
new_height=int(new_width*1.0/width*height);
}
cv::resize(img, img, cv::Size(new_width, new_height));
Mat outimg = net.detect(img);
auto end = std::chrono::steady_clock::now();
std::chrono::duration<double> spent = end - start;
qDebug()<< " Time: " << spent.count() << " sec \n";
imshow("CSI Camera",outimg);
int keycode = cv::waitKey(2) & 0xff ; //ESC键退出
if (keycode == 27) break ;
}
cap.release();
destroyAllWindows() ;
}

32
mainwindow.h Normal file
Переглянути файл

@@ -0,0 +1,32 @@
#ifndef MAINWINDOW_H
#define MAINWINDOW_H
#include <QMainWindow>
QT_BEGIN_NAMESPACE
namespace Ui { class MainWindow; }
QT_END_NAMESPACE
class Detector;
class MainWindow : public QMainWindow
{
Q_OBJECT
public:
MainWindow(QWidget *parent = nullptr);
~MainWindow();
public:
void OpenUSBCamera();
void OpenCSICamera();
void ShowImage();
void ShowVideo();
std::vector<std::string> m_vObjects_Names; //检测目标名称
Detector *m_pDetector;
private:
Ui::MainWindow *ui;
};
#endif // MAINWINDOW_H

22
mainwindow.ui Normal file
Переглянути файл

@@ -0,0 +1,22 @@
<?xml version="1.0" encoding="UTF-8"?>
<ui version="4.0">
<class>MainWindow</class>
<widget class="QMainWindow" name="MainWindow">
<property name="geometry">
<rect>
<x>0</x>
<y>0</y>
<width>800</width>
<height>600</height>
</rect>
</property>
<property name="windowTitle">
<string>MainWindow</string>
</property>
<widget class="QWidget" name="centralwidget"/>
<widget class="QMenuBar" name="menubar"/>
<widget class="QStatusBar" name="statusbar"/>
</widget>
<resources/>
<connections/>
</ui>