yolov5动态链接库DLL导出(TensorRT)

    延续上一篇tTensorRT部署yolov5，大家可以使用生成的yolov5.exe进行终端命令或者VS里面使用命令代码进行检测，但是这样看起来很繁琐很臃肿，有些同学想调用他做一个QT界面啥的，直接调用这个dll就可以进行推理又方便还很快，大家也可以去原博主下面查看，

首选i保证你看了我的第一篇tensort推理yolov5，我们打开cmake编译程序的工程目录：

一.文件创建:

导出库必要文件：dllmain.cpp,framework.h,就在当前工程下面建立这两个文件

// dllmain.cpp : 定义 DLL 应用程序的入口点。
#pragma once
#include "pch.h"

BOOL APIENTRY DllMain(HMODULE hModule,
	DWORD  ul_reason_for_call,
	LPVOID lpReserved
)
{
	switch (ul_reason_for_call)
	{
	case DLL_PROCESS_ATTACH:
	case DLL_THREAD_ATTACH:
	case DLL_THREAD_DETACH:
	case DLL_PROCESS_DETACH:
		break;
	}
	return TRUE;
}

// framework.h
#pragma once

#define WIN32_LEAN_AND_MEAN             // 从 Windows 头文件中排除极少使用的内容
// Windows 头文件
#include <windows.h>

新建导出类文件：pch.h,Detection.h,Detection.cpp

文件：pch.h 声明虚基类，定义了模型参数宏：

//hcp.h
#pragma once
#ifndef PCH_H
#define PCH_H

// 添加要在此处预编译的标头
#include "framework.h"
#include <opencv2/opencv.hpp>
#include <opencv2/dnn.hpp>
#include <string.h>
#include <vector>
#include <fstream>
#include <sstream>
#include <iostream>

#define USE_FP16  // set USE_INT8 or USE_FP16 or USE_FP32
#define DEVICE 0  // GPU id
#define NMS_THRESH 0.4
#define CONF_THRESH 0.5
#define BATCH_SIZE 1

#define CLASS_DECLSPEC __declspec(dllexport)//表示这里要把类导出//

struct Net_config
{
	float gd; // engine threshold
	float gw;  // engine threshold
	const char* netname;
};


class CLASS_DECLSPEC YOLOV5
{
public:
	YOLOV5() {};
	virtual ~YOLOV5() {};
public:
	virtual void Initialize(const char* model_path, int num) = 0;
	virtual int Detecting(cv::Mat& frame, std::vector<cv::Rect>& Boxes, std::vector<const char*>& ClassLables) = 0;
};

#endif //PCH_H

导出类头文件：Detection.h，声明导出类，声明关联类导出类与虚基类，

大家注意下这里的类别和类别数量根据自己训练的情况来定，如果是官网80类，则把类别复制过来，

//Detection.h
#pragma once
#include "pch.h"
#include "yololayer.h"
#include <chrono>
#include "cuda_utils.h"
#include "logging.h"
#include "common.hpp"
#include "utils.h"
#include "calibrator.h"



class CLASS_DECLSPEC Connect
{
public:
	Connect();
	~Connect();

public:
	YOLOV5* Create_YOLOV5_Object();
	void Delete_YOLOV5_Object(YOLOV5* _bp);
};



class Detection :public YOLOV5
{
public:
	Detection();
	~Detection();

	void Initialize(const char* model_path, int num);
	void setClassNum(int num);
	int Detecting(cv::Mat& frame, std::vector<cv::Rect>& Boxes, std::vector<const char*>& ClassLables);

private:
	char netname[20] = { 0 };
	float gd = 0.0f, gw = 0.0f;
	const char* classes[2] = { "J_Deformation", "J_Splitting" };
	Net_config yolo_nets[4] = {
		{0.33, 0.50, "yolov5s"},
		{0.67, 0.75, "yolov5m"},
		{1.00, 1.00, "yolov5l"},
		{1.33, 1.25, "yolov5x"}
	};

	int CLASS_NUM = 2;
	float data[1 * 3 * 640 * 640];
	float prob[1 * 6001];
	size_t size = 0;

	int inputIndex = 0;
	int outputIndex = 0;

	char* trtModelStream = nullptr;
	void* buffers[2] = { 0 };

	nvinfer1::IExecutionContext* context;
	cudaStream_t stream;
	nvinfer1::IRuntime* runtime;
	nvinfer1::ICudaEngine* engine;
};

Detection.cpp 实现导出类，实现关联类导出类与虚基类

//Detection.cpp
#pragma once
#include "pch.h"
#include "Detection.h"


using namespace std;
static const int INPUT_H = Yolo::INPUT_H;
static const int INPUT_W = Yolo::INPUT_W;
static const int OUTPUT_SIZE = Yolo::MAX_OUTPUT_BBOX_COUNT * sizeof(Yolo::Detection) / sizeof(float) + 1;  // we assume the yololayer outputs no more than MAX_OUTPUT_BBOX_COUNT boxes that conf >= 0.1
const char* INPUT_BLOB_NAME = "data";
const char* OUTPUT_BLOB_NAME = "prob";
static Logger gLogger;



static int get_width(int x, float gw, int divisor = 8) {
	//return math.ceil(x / divisor) * divisor
	if (int(x * gw) % divisor == 0) {
		return int(x * gw);
	}
	return (int(x * gw / divisor) + 1) * divisor;
}

static int get_depth(int x, float gd) {
	if (x == 1) {
		return 1;
	}
	else {
		return round(x * gd) > 1 ? round(x * gd) : 1;
	}
}

ICudaEngine* build_engine(unsigned int maxBatchSize, IBuilder* builder, IBuilderConfig* config, DataType dt, float& gd, float& gw, std::string& wts_name) {
	INetworkDefinition* network = builder->createNetworkV2(0U);

	// Create input tensor of shape {3, INPUT_H, INPUT_W} with name INPUT_BLOB_NAME
	ITensor* data = network->addInput(INPUT_BLOB_NAME, dt, Dims3{ 3, INPUT_H, INPUT_W });
	assert(data);

	std::map<std::string, Weights> weightMap = loadWeights(wts_name);

	/* ------ yolov5 backbone------ */
	auto focus0 = focus(network, weightMap, *data, 3, get_width(64, gw), 3, "model.0");
	auto conv1 = convBlock(network, weightMap, *focus0->getOutput(0), get_width(128, gw), 3, 2, 1, "model.1");
	auto bottleneck_CSP2 = C3(network, weightMap, *conv1->getOutput(0), get_width(128, gw), get_width(128, gw), get_depth(3, gd), true, 1, 0.5, "model.2");
	auto conv3 = convBlock(network, weightMap, *bottleneck_CSP2->getOutput(0), get_width(256, gw), 3, 2, 1, "model.3");
	auto bottleneck_csp4 = C3(network, weightMap, *conv3->getOutput(0), get_width(256, gw), get_width(256, gw), get_depth(9, gd), true, 1, 0.5, "model.4");
	auto conv5 = convBlock(network, weightMap, *bottleneck_csp4->getOutput(0), get_width(512, gw), 3, 2, 1, "model.5");
	auto bottleneck_csp6 = C3(network, weightMap, *conv5->getOutput(0), get_width(512, gw), get_width(512, gw), get_depth(9, gd), true, 1, 0.5, "model.6");
	auto conv7 = convBlock(network, weightMap, *bottleneck_csp6->getOutput(0), get_width(1024, gw), 3, 2, 1, "model.7");
	auto spp8 = SPP(network, weightMap, *conv7->getOutput(0), get_width(1024, gw), get_width(1024, gw), 5, 9, 13, "model.8");

	/* ------ yolov5 head ------ */
	auto bottleneck_csp9 = C3(network, weightMap, *spp8->getOutput(0), get_width(1024, gw), get_width(1024, gw), get_depth(3, gd), false, 1, 0.5, "model.9");
	auto conv10 = convBlock(network, weightMap, *bottleneck_csp9->getOutput(0), get_width(512, gw), 1, 1, 1, "model.10");

	auto upsample11 = network->addResize(*conv10->getOutput(0));
	assert(upsample11);
	upsample11->setResizeMode(ResizeMode::kNEAREST);
	upsample11->setOutputDimensions(bottleneck_csp6->getOutput(0)->getDimensions());

	ITensor* inputTensors12[] = { upsample11->getOutput(0), bottleneck_csp6->getOutput(0) };
	auto cat12 = network->addConcatenation(inputTensors12, 2);
	auto bottleneck_csp13 = C3(network, weightMap, *cat12->getOutput(0), get_width(1024, gw), get_width(512, gw), get_depth(3, gd), false, 1, 0.5, "model.13");
	auto conv14 = convBlock(network, weightMap, *bottleneck_csp13->getOutput(0), get_width(256, gw), 1, 1, 1, "model.14");

	auto upsample15 = network->addResize(*conv14->getOutput(0));
	assert(upsample15);
	upsample15->setResizeMode(ResizeMode::kNEAREST);
	upsample15->setOutputDimensions(bottleneck_csp4->getOutput(0)->getDimensions());

	ITensor* inputTensors16[] = { upsample15->getOutput(0), bottleneck_csp4->getOutput(0) };
	auto cat16 = network->addConcatenation(inputTensors16, 2);

	auto bottleneck_csp17 = C3(network, weightMap, *cat16->getOutput(0), get_width(512, gw), get_width(256, gw), get_depth(3, gd), false, 1, 0.5, "model.17");

	// yolo layer 0
	IConvolutionLayer* det0 = network->addConvolutionNd(*bottleneck_csp17->getOutput(0), 3 * (Yolo::CLASS_NUM + 5), DimsHW{ 1, 1 }, weightMap["model.24.m.0.weight"], weightMap["model.24.m.0.bias"]);
	auto conv18 = convBlock(network, weightMap, *bottleneck_csp17->getOutput(0), get_width(256, gw), 3, 2, 1, "model.18");
	ITensor* inputTensors19[] = { conv18->getOutput(0), conv14->getOutput(0) };
	auto cat19 = network->addConcatenation(inputTensors19, 2);
	auto bottleneck_csp20 = C3(network, weightMap, *cat19->getOutput(0), get_width(512, gw), get_width(512, gw), get_depth(3, gd), false, 1, 0.5, "model.20");
	//yolo layer 1
	IConvolutionLayer* det1 = network->addConvolutionNd(*bottleneck_csp20->getOutput(0), 3 * (Yolo::CLASS_NUM + 5), DimsHW{ 1, 1 }, weightMap["model.24.m.1.weight"], weightMap["model.24.m.1.bias"]);
	auto conv21 = convBlock(network, weightMap, *bottleneck_csp20->getOutput(0), get_width(512, gw), 3, 2, 1, "model.21");
	ITensor* inputTensors22[] = { conv21->getOutput(0), conv10->getOutput(0) };
	auto cat22 = network->addConcatenation(inputTensors22, 2);
	auto bottleneck_csp23 = C3(network, weightMap, *cat22->getOutput(0), get_width(1024, gw), get_width(1024, gw), get_depth(3, gd), false, 1, 0.5, "model.23");
	IConvolutionLayer* det2 = network->addConvolutionNd(*bottleneck_csp23->getOutput(0), 3 * (Yolo::CLASS_NUM + 5), DimsHW{ 1, 1 }, weightMap["model.24.m.2.weight"], weightMap["model.24.m.2.bias"]);

	auto yolo = addYoLoLayer(network, weightMap, det0, det1, det2);
	yolo->getOutput(0)->setName(OUTPUT_BLOB_NAME);
	network->markOutput(*yolo->getOutput(0));

	// Build engine
	builder->setMaxBatchSize(maxBatchSize);
	config->setMaxWorkspaceSize(16 * (1 << 20));  // 16MB
#if defined(USE_FP16)
	config->setFlag(BuilderFlag::kFP16);
#elif defined(USE_INT8)
	std::cout << "Your platform support int8: " << (builder->platformHasFastInt8() ? "true" : "false") << std::endl;
	assert(builder->platformHasFastInt8());
	config->setFlag(BuilderFlag::kINT8);
	Int8EntropyCalibrator2* calibrator = new Int8EntropyCalibrator2(1, INPUT_W, INPUT_H, "./coco_calib/", "int8calib.table", INPUT_BLOB_NAME);
	config->setInt8Calibrator(calibrator);
#endif

	std::cout << "Building engine, please wait for a while..." << std::endl;
	ICudaEngine* engine = builder->buildEngineWithConfig(*network, *config);
	std::cout << "Build engine successfully!" << std::endl;

	// Don't need the network any more
	network->destroy();

	// Release host memory
	for (auto& mem : weightMap)
	{
		free((void*)(mem.second.values));
	}

	return engine;
}

void APIToModel(unsigned int maxBatchSize, IHostMemory** modelStream, float& gd, float& gw, std::string& wts_name) {
	// Create builder
	IBuilder* builder = createInferBuilder(gLogger);
	IBuilderConfig* config = builder->createBuilderConfig();

	// Create model to populate the network, then set the outputs and create an engine
	ICudaEngine* engine = build_engine(maxBatchSize, builder, config, DataType::kFLOAT, gd, gw, wts_name);
	assert(engine != nullptr);

	// Serialize the engine
	(*modelStream) = engine->serialize();

	// Close everything down
	engine->destroy();
	builder->destroy();
	config->destroy();
}

inline void doInference(IExecutionContext& context, cudaStream_t& stream, void** buffers, float* input, float* output, int batchSize) {
	// DMA input batch data to device, infer on the batch asynchronously, and DMA output back to host
	CUDA_CHECK(cudaMemcpyAsync(buffers[0], input, batchSize * 3 * INPUT_H * INPUT_W * sizeof(float), cudaMemcpyHostToDevice, stream));
	context.enqueue(batchSize, buffers, stream, nullptr);
	CUDA_CHECK(cudaMemcpyAsync(output, buffers[1], batchSize * OUTPUT_SIZE * sizeof(float), cudaMemcpyDeviceToHost, stream));
	cudaStreamSynchronize(stream);
}


void Detection::Initialize(const char* model_path, int num)
{
	if (num < 0 || num>3) {
		cout << "=================="
			"0, yolov5s"
			"1, yolov5m"
			"2, yolov5l"
			"3, yolov5x" << endl;
		return;
	}
	cout << "Net use :" << yolo_nets[num].netname << endl;
	this->gd = yolo_nets[num].gd;
	this->gw = yolo_nets[num].gw;

	//初始化GPU引擎
	cudaSetDevice(DEVICE);
	std::ifstream file(model_path, std::ios::binary);
	if (!file.good()) {
		std::cerr << "read " << model_path << " error!" << std::endl;
		return;
	}

	file.seekg(0, file.end);
	size = file.tellg();                //统计模型字节流大小
	file.seekg(0, file.beg);
	trtModelStream = new char[size];    // 申请模型字节流大小的空间
	assert(trtModelStream);
	file.read(trtModelStream, size);    // 读取字节流到trtModelStream
	file.close();


	// prepare input data ------NCHW---------------------
	runtime = createInferRuntime(gLogger);
	assert(runtime != nullptr);
	engine = runtime->deserializeCudaEngine(trtModelStream, size);
	assert(engine != nullptr);
	context = engine->createExecutionContext();
	assert(context != nullptr);
	delete[] trtModelStream;
	assert(engine->getNbBindings() == 2);
	inputIndex = engine->getBindingIndex(INPUT_BLOB_NAME);
	outputIndex = engine->getBindingIndex(OUTPUT_BLOB_NAME);
	assert(inputIndex == 0);
	assert(outputIndex == 1);
	// Create GPU buffers on device
	CUDA_CHECK(cudaMalloc(&buffers[inputIndex], BATCH_SIZE * 3 * INPUT_H * INPUT_W * sizeof(float)));
	CUDA_CHECK(cudaMalloc(&buffers[outputIndex], BATCH_SIZE * OUTPUT_SIZE * sizeof(float)));
	CUDA_CHECK(cudaStreamCreate(&stream));
	std::cout << "Engine Initialize successfully!" << endl;
}


void Detection::setClassNum(int num) 
{
	CLASS_NUM = num;
}


int Detection::Detecting(cv::Mat& img, std::vector<cv::Rect>& Boxes, std::vector<const char*>& ClassLables)
{
	if (img.empty()) {
		std::cout << "read image failed!" << std::endl;
		return -1;
	}
	if (img.rows < 640 || img.cols < 640) {
		std::cout << "img.rows: "<< img.rows <<"\timg.cols: "<< img.cols << std::endl;
		std::cout << "image height<640||width<640!" << std::endl;
		return -1;
	}
	cv::Mat pr_img = preprocess_img(img, INPUT_W, INPUT_H); // letterbox BGR to RGB
	int i = 0;
	for (int row = 0; row < INPUT_H; ++row) {
		uchar* uc_pixel = pr_img.data + row * pr_img.step;
		for (int col = 0; col < INPUT_W; ++col) {
			data[i] = (float)uc_pixel[2] / 255.0;
			data[i + INPUT_H * INPUT_W] = (float)uc_pixel[1] / 255.0;
			data[i + 2 * INPUT_H * INPUT_W] = (float)uc_pixel[0] / 255.0;
			uc_pixel += 3;
			++i;
		}
	}

	// Run inference
	auto start = std::chrono::system_clock::now();
	doInference(*context, stream, buffers, data, prob, BATCH_SIZE);
	auto end = std::chrono::system_clock::now();
	std::cout << std::chrono::duration_cast<std::chrono::milliseconds>(end - start).count() << "ms" << std::endl;
	std::vector<Yolo::Detection> batch_res;
	nms(batch_res, &prob[0], CONF_THRESH, NMS_THRESH);

	for (size_t j = 0; j < batch_res.size(); j++) {
		cv::Rect r = get_rect(img, batch_res[j].bbox);
		Boxes.push_back(r);
		ClassLables.push_back(classes[(int)batch_res[j].class_id]);
		cv::rectangle(img, r, cv::Scalar(0x27, 0xC1, 0x36), 2);
		cv::putText(
			img, 
			classes[(int)batch_res[j].class_id], 
			cv::Point(r.x, r.y - 2), 
			cv::FONT_HERSHEY_COMPLEX, 
			1.8, 
			cv::Scalar(0xFF, 0xFF, 0xFF), 
			2
		);
	}
	return 0;
}



Detection::Detection() {}
Detection::~Detection() 
{
	// Release stream and buffers
	cudaStreamDestroy(stream);
	CUDA_CHECK(cudaFree(buffers[inputIndex]));
	CUDA_CHECK(cudaFree(buffers[outputIndex]));
	// Destroy the engine
	context->destroy();
	engine->destroy();
	runtime->destroy();
}



Connect::Connect()
{}
Connect::~Connect()
{}


YOLOV5* Connect::Create_YOLOV5_Object()
{
	return new Detection;		//注意此处
}


void Connect::Delete_YOLOV5_Object(YOLOV5* _bp)
{
	if (_bp)
		delete _bp;
}

二.编译

1、修改CMakeLists.txt 修改生成目标为动态链接库。（这里去掉了yolov5.cpp，并新增了新建的文件）

#修改前
add_executable(yolov5 ${PROJECT_SOURCE_DIR}/yolov5.cpp ${PROJECT_SOURCE_DIR}/common.hpp ${PROJECT_SOURCE_DIR}/yololayer.cu ${PROJECT_SOURCE_DIR}/yololayer.h)

#修改后
add_library(yolov5 SHARED ${PROJECT_SOURCE_DIR}/common.hpp ${PROJECT_SOURCE_DIR}/yololayer.cu ${PROJECT_SOURCE_DIR}/yololayer.h "Detection.h" "Detection.cpp" "framework.h" "dllmain.cpp"  )

大家可以新建一个文件夹build_dll：打开cmake

 编译完configure---Generate----open project进行realses和debug编译：

 编译完成后出现无法启动程序大家可以不用管，打开自己的build_dll：

 release:会出现我们生成的dll:

三.测试

 然后大家就可以VS新建自己的项目main.cpp,

将上面三个文件放入此工程中：以及yolov5s.engine权重和图片，

 大家将下面代码复制到main.cpp中：这里的Detection.h和dirent.h根据自己的路径添加：

#pragma once
#include <iostream>
#include <opencv2/opencv.hpp>
#include <opencv2/dnn.hpp>
#include "D:/yolov5_tensort/tensorrtx-yolov5-v4.0/yolov5/Detection.h"
#include "D:/yolov5_tensort/tensorrtx-yolov5-v4.0/yolov5/dirent.h"
#include "yololayer.h"
int main()
{

	Connect connect;
	YOLOV5* yolo_dll = connect.Create_YOLOV5_Object();

	cv::VideoCapture capture(0);
	if (!capture.isOpened()) {
		std::cout << "Error opening video stream or file" << std::endl;
		return -1;
	}
	yolo_dll->Initialize("./yolov5s.engine", 0);
	while (1)
	{
		cv::Mat frame;
		capture >> frame;
		vector<cv::Rect> Boxes;
		vector<const char*> ClassLables;

		yolo_dll->Detecting(frame, Boxes, ClassLables);
		cv::imshow("output", frame);
		cv::waitKey(1);
	}
	connect.Delete_YOLOV5_Object(yolo_dll);
	return 0;
}

现在我们要配置包含目录、库目录、附加依赖项。

#将此路径加入项目属性包含目录中
D:\yolov5_tensort\tensorrtx-yolov5-v4.0\yolov5

#将此路径加入项目属性的库目录中，也就是我们刚刚生成dll的文件目录
D:\yolov5_tensort\tensorrtx-yolov5-v4.0\yolov5\build_dll\Release

#在输入链接器添加依赖库
yolov5.dll

四.问题解决

1.

 网上很多方法都试过感觉没有什么用：第一个错误我分析为没有相关的依赖库导入，所以我就把tensort的库全部导入：

opencv_world341.lib
opencv_world341d.lib
cudart.lib
cudart_static.lib
yolov5.lib
myelin64_1.lib
nvinfer.lib
nvinfer_plugin.lib
nvonnxparser.lib
nvparsers.lib

再次运行之后发现第一个确实不报错了：

大胆的猜想误打误撞将yololayer.cu导入到文件中，并配置为cuda/c++

 此时代码可以运行了但是还是出错：

 说明模型是初始化了，但是在检测的时候出错了，我们进入源码看一看：detection.cpp

 原来是这里大哥把图片大小设置了我们把这里改成320：

再次运行发现还是检测部分出现错误：

 传入的classlables错误，OK检测视频，调用摄像头没问题，把这两个注释掉：

 差不多问题解决了可以运行了；

测试视频、照片、摄像头大家直接用opencv就可以实现。

1

 大家可以在QT上面使用，设计一个界面很NICE。感谢大家有什么建议进群交流：471550878