目标跟踪算法（bytetrack)-tensorrt部署教程

2024-06-22 03:36
文章标签 算法 部署 教程 目标 bytetrack 跟踪 tensorrt

本文主要是介绍目标跟踪算法（bytetrack)-tensorrt部署教程，希望对大家解决编程问题提供一定的参考价值，需要的开发者们随着小编来一起学习吧！

一、本机安装python环境

conda create -n bytetrace_env python=3.8
activate bytetrace_env
conda install pytorch torchvision cudatoolkit=10.1 -c

检测GPU是否可用，不可用不行

import torch
print(torch.cuda.is_available())

安装bytetrack

git clone https://github.com/ifzhang/ByteTrack.git
cd ByteTrack
pip install -r requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple
python setup.py develop

在这里插入图片描述
上述即安装成功。
安装pycocotools

pip install cython
pip install 'git+https://github.com/cocodataset/cocoapi.git#subdirectory=PythonAPI'

或者（Linux）

pip install git+https://gitee.com/pursuit_zhangyu/cocoapi.git#subdirectory=PythonAPI

windows下

pip install pycocotools-windows

二、安装tensorrt环境

下载tensorrt包
TensorRT-8.4.3.1.Windows10.x86_64.cuda-10.2.cudnn8.4.zip
在这里插入图片描述
将所有的dll复制到
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\bin
并设置环境变量。
在这里插入图片描述
虚拟环境中python版本为python3.8

pip install tensorrt-8.4.3.1-cp38-none-win_amd64.whl

三、转换模型

https://pan.baidu.com/s/1PiP1kQfgxAIrnGUbFP6Wfg
qflm
获取bytetrack_s_mot17.pth.tar并创建pretrained进行存放

python tools/trt.py -f exps/example/mot/yolox_s_mix_det.py -c pretrained/bytetrack_s_mot17.pth.tar

最后在D:\git_clone\ByteTrack-main\YOLOX_outputs\yolox_s_mix_det目录下生成tensorrt模型与pth模型：
在这里插入图片描述

四、cmake生成eigen库并使用VS2015编译

https://pan.baidu.com/s/15kEfCxpy-T7tz60msxxExg
ueq4
在这里插入图片描述
在这里插入图片描述
在这里插入图片描述
在这里插入图片描述
在这里插入图片描述

五、下载opencv450

https://nchc.dl.sourceforge.net/project/opencvlibrary/4.5.0/opencv-4.5.0-vc14_vc15.exe?viasf=1
安装D:\opencv450

六、cmake生成bytetrack并使用VS2015编译

修改CMakeLists.txt

cmake_minimum_required(VERSION 2.6)project(bytetrack)add_definitions(-std=c++11)option(CUDA_USE_STATIC_CUDA_RUNTIME OFF)
set(CMAKE_CXX_STANDARD 11)
set(CMAKE_BUILD_TYPE Debug)find_package(CUDA REQUIRED)include_directories(${PROJECT_SOURCE_DIR}/include)
include_directories(D:\VS2015_CUDA\TensorRT\eigen-3.3.9\build\install\include\eigen3)
link_directories(${PROJECT_SOURCE_DIR}/include)
# include and link dirs of cuda and tensorrt, you need adapt them if yours are different
# cuda
include_directories(C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\include)
link_directories(C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\lib\x64)
# cudnn
include_directories(C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\include)
link_directories(C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\lib\x64)
# tensorrt
include_directories(D:\VS2015_CUDA\TensorRT\TensorRT-8.4.3.1\include)
link_directories(D:\VS2015_CUDA\TensorRT\TensorRT-8.4.3.1\lib)set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11 -Wall -D_MWAITXINTRIN_H_INCLUDED")set(OpenCV_INCLUDE_DIRS_DIRS D:\opencv450\build\include)
set(OpenCV_LIBS D:\opencv450\build\x64\vc14\lib)
include_directories(${OpenCV_INCLUDE_DIRS})file(GLOB My_Source_Files ${PROJECT_SOURCE_DIR}/src/*.cpp)
add_executable(bytetrack ${My_Source_Files})
target_link_libraries(bytetrack nvinfer)
target_link_libraries(bytetrack cudart)
target_link_libraries(bytetrack ${OpenCV_LIBS})add_definitions(-O2 -pthread)

使用CMake进行配置生成。
打开VS2015工程进行配置
在这里插入图片描述
在这里插入图片描述
在这里插入图片描述
opencv可以自己加进去编译。
在这里插入图片描述
修改bytetrack.cpp

int img_w = cap.get(cv::CAP_PROP_FRAME_WIDTH);
int img_h = cap.get(cv::CAP_PROP_FRAME_HEIGHT);

这里有个bug，高的参数不对，会影响后面的demo.mp4。
编译生成
在这里插入图片描述

七、使用Dependency Walker查看exe依赖

在这里插入图片描述
在这里插入图片描述

D:\VS2015_CUDA\TensorRT\cpp\model_trt.engine -i D:\VS2015_CUDA\TensorRT\cpp\palace.mp4

八、windows源码

bytetrack.cpp

#include <fstream>
#include <iostream>
#include <sstream>
#include <numeric>
#include <chrono>
#include <vector>
#include <opencv2/opencv.hpp>
#include "NvInfer.h"
#include "cuda_runtime_api.h"
#include "logging.h"
#include "BYTETracker.h"#define CHECK(status) \do\{\auto ret = (status);\if (ret != 0)\{\cerr << "Cuda failure: " << ret << endl;\abort();\}\} while (0)#define DEVICE 0  // GPU id
#define NMS_THRESH 0.7
#define BBOX_CONF_THRESH 0.1using namespace nvinfer1;// stuff we know about the network and the input/output blobs
static const int INPUT_W = 1088;
static const int INPUT_H = 608;
const char* INPUT_BLOB_NAME = "input_0";
const char* OUTPUT_BLOB_NAME = "output_0";
static Logger gLogger;Mat static_resize(Mat& img) {float r = min(INPUT_W / (img.cols*1.0), INPUT_H / (img.rows*1.0));// r = std::min(r, 1.0f);int unpad_w = r * img.cols;int unpad_h = r * img.rows;Mat re(unpad_h, unpad_w, CV_8UC3);resize(img, re, re.size());Mat out(INPUT_H, INPUT_W, CV_8UC3, Scalar(114, 114, 114));re.copyTo(out(Rect(0, 0, re.cols, re.rows)));return out;
}struct GridAndStride
{int grid0;int grid1;int stride;
};static void generate_grids_and_stride(const int target_w, const int target_h, vector<int>& strides, vector<GridAndStride>& grid_strides)
{for (auto stride : strides){GridAndStride GS;int num_grid_w = target_w / stride;int num_grid_h = target_h / stride;for (int g1 = 0; g1 < num_grid_h; g1++){for (int g0 = 0; g0 < num_grid_w; g0++){GS.grid0 = g0;GS.grid1 = g1;GS.stride = stride;grid_strides.push_back(GS);}}}
}static inline float intersection_area(const Object& a, const Object& b)
{Rect_<float> inter = a.rect & b.rect;return inter.area();
}static void qsort_descent_inplace(vector<Object>& faceobjects, int left, int right)
{int i = left;int j = right;float p = faceobjects[(left + right) / 2].prob;while (i <= j){while (faceobjects[i].prob > p)i++;while (faceobjects[j].prob < p)j--;if (i <= j){// swapswap(faceobjects[i], faceobjects[j]);i++;j--;}}#pragma omp parallel sections{#pragma omp section{if (left < j) qsort_descent_inplace(faceobjects, left, j);}#pragma omp section{if (i < right) qsort_descent_inplace(faceobjects, i, right);}}
}static void qsort_descent_inplace(vector<Object>& objects)
{if (objects.empty())return;qsort_descent_inplace(objects, 0, objects.size() - 1);
}static void nms_sorted_bboxes(const vector<Object>& faceobjects, vector<int>& picked, float nms_threshold)
{picked.clear();const int n = faceobjects.size();vector<float> areas(n);for (int i = 0; i < n; i++){areas[i] = faceobjects[i].rect.area();}for (int i = 0; i < n; i++){const Object& a = faceobjects[i];int keep = 1;for (int j = 0; j < (int)picked.size(); j++){const Object& b = faceobjects[picked[j]];// intersection over unionfloat inter_area = intersection_area(a, b);float union_area = areas[i] + areas[picked[j]] - inter_area;// float IoU = inter_area / union_areaif (inter_area / union_area > nms_threshold)keep = 0;}if (keep)picked.push_back(i);}
}static void generate_yolox_proposals(vector<GridAndStride> grid_strides, float* feat_blob, float prob_threshold, vector<Object>& objects)
{const int num_class = 1;const int num_anchors = grid_strides.size();for (int anchor_idx = 0; anchor_idx < num_anchors; anchor_idx++){const int grid0 = grid_strides[anchor_idx].grid0;const int grid1 = grid_strides[anchor_idx].grid1;const int stride = grid_strides[anchor_idx].stride;const int basic_pos = anchor_idx * (num_class + 5);// yolox/models/yolo_head.py decode logicfloat x_center = (feat_blob[basic_pos+0] + grid0) * stride;float y_center = (feat_blob[basic_pos+1] + grid1) * stride;float w = exp(feat_blob[basic_pos+2]) * stride;float h = exp(feat_blob[basic_pos+3]) * stride;float x0 = x_center - w * 0.5f;float y0 = y_center - h * 0.5f;float box_objectness = feat_blob[basic_pos+4];for (int class_idx = 0; class_idx < num_class; class_idx++){float box_cls_score = feat_blob[basic_pos + 5 + class_idx];float box_prob = box_objectness * box_cls_score;if (box_prob > prob_threshold){Object obj;obj.rect.x = x0;obj.rect.y = y0;obj.rect.width = w;obj.rect.height = h;obj.label = class_idx;obj.prob = box_prob;objects.push_back(obj);}} // class loop} // point anchor loop
}float* blobFromImage(Mat& img){cvtColor(img, img, COLOR_BGR2RGB);float* blob = new float[img.total()*3];int channels = 3;int img_h = img.rows;int img_w = img.cols;vector<float> mean = {0.485f, 0.456f, 0.406f};vector<float> std = {0.229f, 0.224f, 0.225f};for (size_t c = 0; c < channels; c++) {for (size_t  h = 0; h < img_h; h++) {for (size_t w = 0; w < img_w; w++) {blob[c * img_w * img_h + h * img_w + w] =(((float)img.at<Vec3b>(h, w)[c]) / 255.0f - mean[c]) / std[c];}}}return blob;
}static void decode_outputs(float* prob, vector<Object>& objects, float scale, const int img_w, const int img_h) {vector<Object> proposals;vector<int> strides = {8, 16, 32};vector<GridAndStride> grid_strides;generate_grids_and_stride(INPUT_W, INPUT_H, strides, grid_strides);generate_yolox_proposals(grid_strides, prob,  BBOX_CONF_THRESH, proposals);//std::cout << "num of boxes before nms: " << proposals.size() << std::endl;qsort_descent_inplace(proposals);vector<int> picked;nms_sorted_bboxes(proposals, picked, NMS_THRESH);int count = picked.size();//std::cout << "num of boxes: " << count << std::endl;objects.resize(count);for (int i = 0; i < count; i++){objects[i] = proposals[picked[i]];// adjust offset to original unpaddedfloat x0 = (objects[i].rect.x) / scale;float y0 = (objects[i].rect.y) / scale;float x1 = (objects[i].rect.x + objects[i].rect.width) / scale;float y1 = (objects[i].rect.y + objects[i].rect.height) / scale;// clip// x0 = std::max(std::min(x0, (float)(img_w - 1)), 0.f);// y0 = std::max(std::min(y0, (float)(img_h - 1)), 0.f);// x1 = std::max(std::min(x1, (float)(img_w - 1)), 0.f);// y1 = std::max(std::min(y1, (float)(img_h - 1)), 0.f);objects[i].rect.x = x0;objects[i].rect.y = y0;objects[i].rect.width = x1 - x0;objects[i].rect.height = y1 - y0;}
}const float color_list[80][3] =
{{0.000, 0.447, 0.741},{0.850, 0.325, 0.098},{0.929, 0.694, 0.125},{0.494, 0.184, 0.556},{0.466, 0.674, 0.188},{0.301, 0.745, 0.933},{0.635, 0.078, 0.184},{0.300, 0.300, 0.300},{0.600, 0.600, 0.600},{1.000, 0.000, 0.000},{1.000, 0.500, 0.000},{0.749, 0.749, 0.000},{0.000, 1.000, 0.000},{0.000, 0.000, 1.000},{0.667, 0.000, 1.000},{0.333, 0.333, 0.000},{0.333, 0.667, 0.000},{0.333, 1.000, 0.000},{0.667, 0.333, 0.000},{0.667, 0.667, 0.000},{0.667, 1.000, 0.000},{1.000, 0.333, 0.000},{1.000, 0.667, 0.000},{1.000, 1.000, 0.000},{0.000, 0.333, 0.500},{0.000, 0.667, 0.500},{0.000, 1.000, 0.500},{0.333, 0.000, 0.500},{0.333, 0.333, 0.500},{0.333, 0.667, 0.500},{0.333, 1.000, 0.500},{0.667, 0.000, 0.500},{0.667, 0.333, 0.500},{0.667, 0.667, 0.500},{0.667, 1.000, 0.500},{1.000, 0.000, 0.500},{1.000, 0.333, 0.500},{1.000, 0.667, 0.500},{1.000, 1.000, 0.500},{0.000, 0.333, 1.000},{0.000, 0.667, 1.000},{0.000, 1.000, 1.000},{0.333, 0.000, 1.000},{0.333, 0.333, 1.000},{0.333, 0.667, 1.000},{0.333, 1.000, 1.000},{0.667, 0.000, 1.000},{0.667, 0.333, 1.000},{0.667, 0.667, 1.000},{0.667, 1.000, 1.000},{1.000, 0.000, 1.000},{1.000, 0.333, 1.000},{1.000, 0.667, 1.000},{0.333, 0.000, 0.000},{0.500, 0.000, 0.000},{0.667, 0.000, 0.000},{0.833, 0.000, 0.000},{1.000, 0.000, 0.000},{0.000, 0.167, 0.000},{0.000, 0.333, 0.000},{0.000, 0.500, 0.000},{0.000, 0.667, 0.000},{0.000, 0.833, 0.000},{0.000, 1.000, 0.000},{0.000, 0.000, 0.167},{0.000, 0.000, 0.333},{0.000, 0.000, 0.500},{0.000, 0.000, 0.667},{0.000, 0.000, 0.833},{0.000, 0.000, 1.000},{0.000, 0.000, 0.000},{0.143, 0.143, 0.143},{0.286, 0.286, 0.286},{0.429, 0.429, 0.429},{0.571, 0.571, 0.571},{0.714, 0.714, 0.714},{0.857, 0.857, 0.857},{0.000, 0.447, 0.741},{0.314, 0.717, 0.741},{0.50, 0.5, 0}
};void doInference(IExecutionContext& context, float* input, float* output, const int output_size, Size input_shape) {const ICudaEngine& engine = context.getEngine();// Pointers to input and output device buffers to pass to engine.// Engine requires exactly IEngine::getNbBindings() number of buffers.assert(engine.getNbBindings() == 2);void* buffers[2];// In order to bind the buffers, we need to know the names of the input and output tensors.// Note that indices are guaranteed to be less than IEngine::getNbBindings()const int inputIndex = engine.getBindingIndex(INPUT_BLOB_NAME);assert(engine.getBindingDataType(inputIndex) == nvinfer1::DataType::kFLOAT);const int outputIndex = engine.getBindingIndex(OUTPUT_BLOB_NAME);assert(engine.getBindingDataType(outputIndex) == nvinfer1::DataType::kFLOAT);int mBatchSize = engine.getMaxBatchSize();// Create GPU buffers on deviceCHECK(cudaMalloc(&buffers[inputIndex], 3 * input_shape.height * input_shape.width * sizeof(float)));CHECK(cudaMalloc(&buffers[outputIndex], output_size*sizeof(float)));// Create streamcudaStream_t stream;CHECK(cudaStreamCreate(&stream));// DMA input batch data to device, infer on the batch asynchronously, and DMA output back to hostCHECK(cudaMemcpyAsync(buffers[inputIndex], input, 3 * input_shape.height * input_shape.width * sizeof(float), cudaMemcpyHostToDevice, stream));context.enqueue(1, buffers, stream, nullptr);CHECK(cudaMemcpyAsync(output, buffers[outputIndex], output_size * sizeof(float), cudaMemcpyDeviceToHost, stream));cudaStreamSynchronize(stream);// Release stream and bufferscudaStreamDestroy(stream);CHECK(cudaFree(buffers[inputIndex]));CHECK(cudaFree(buffers[outputIndex]));
}int main(int argc, char** argv) {cudaSetDevice(DEVICE);// create a model using the API directly and serialize it to a streamchar *trtModelStream{nullptr};size_t size{0};if (argc == 4 && string(argv[2]) == "-i") {const string engine_file_path {argv[1]};ifstream file(engine_file_path, ios::binary);if (file.good()) {file.seekg(0, file.end);size = file.tellg();file.seekg(0, file.beg);trtModelStream = new char[size];assert(trtModelStream);file.read(trtModelStream, size);file.close();}} else {cerr << "arguments not right!" &

这篇关于目标跟踪算法（bytetrack)-tensorrt部署教程的文章就介绍到这儿，希望我们推荐的文章对编程师们有所帮助！

原文地址:
本文来自互联网用户投稿，该文观点仅代表作者本人，不代表本站立场。本站仅提供信息存储空间服务，不拥有所有权，不承担相关法律责任。如若转载，请注明出处：http://www.chinasem.cn/article/1083181。

相关文章

Elasticsearch 在 Java 中的使用教程

Elasticsearch 在 Java 中的使用教程

《Elasticsearch在Java中的使用教程》Elasticsearch是一个分布式搜索和分析引擎,基于ApacheLucene构建,能够实现实时数据的存储、搜索、和分析,它广泛应用于全文... 目录1. Elasticsearch 简介2. 环境准备2.1 安装 Elasticsearch2.2 J
阅读更多...
Linux系统中卸载与安装JDK的详细教程

Linux系统中卸载与安装JDK的详细教程

《Linux系统中卸载与安装JDK的详细教程》本文详细介绍了如何在Linux系统中通过Xshell和Xftp工具连接与传输文件,然后进行JDK的安装与卸载,安装步骤包括连接Linux、传输JDK安装包... 目录1、卸载1.1 linux删除自带的JDK1.2 Linux上卸载自己安装的JDK2、安装2.1
阅读更多...
Linux卸载自带jdk并安装新jdk版本的图文教程

Linux卸载自带jdk并安装新jdk版本的图文教程

《Linux卸载自带jdk并安装新jdk版本的图文教程》在Linux系统中,有时需要卸载预装的OpenJDK并安装特定版本的JDK,例如JDK1.8,所以本文给大家详细介绍了Linux卸载自带jdk并... 目录Ⅰ、卸载自带jdkⅡ、安装新版jdkⅠ、卸载自带jdk1、输入命令查看旧jdkrpm -qa
阅读更多...
Java使用Curator进行ZooKeeper操作的详细教程

Java使用Curator进行ZooKeeper操作的详细教程

《Java使用Curator进行ZooKeeper操作的详细教程》ApacheCurator是一个基于ZooKeeper的Java客户端库,它极大地简化了使用ZooKeeper的开发工作,在分布式系统... 目录1、简述2、核心功能2.1 CuratorFramework2.2 Recipes3、示例实践3
阅读更多...
springboot简单集成Security配置的教程

springboot简单集成Security配置的教程

《springboot简单集成Security配置的教程》：本文主要介绍springboot简单集成Security配置的教程,具有很好的参考价值,希望对大家有所帮助,如有错误或未考虑完全的地方,... 目录集成Security安全框架引入依赖编写配置类WebSecurityConfig（自定义资源权限规则
阅读更多...
SpringBoot实现MD5加盐算法的示例代码

SpringBoot实现MD5加盐算法的示例代码

《SpringBoot实现MD5加盐算法的示例代码》加盐算法是一种用于增强密码安全性的技术,本文主要介绍了SpringBoot实现MD5加盐算法的示例代码,文中通过示例代码介绍的非常详细,对大家的学习... 目录一、什么是加盐算法二、如何实现加盐算法2.1 加盐算法代码实现2.2 注册页面中进行密码加盐2.
阅读更多...
tomcat多实例部署的项目实践

tomcat多实例部署的项目实践

《tomcat多实例部署的项目实践》Tomcat多实例是指在一台设备上运行多个Tomcat服务,这些Tomcat相互独立,本文主要介绍了tomcat多实例部署的项目实践,具有一定的参考价值,感兴趣的可... 目录1.创建项目目录，测试文China编程件2js.创建实例的安装目录3.准备实例的配置文件4.编辑实例的
阅读更多...
MySQL Workbench 安装教程(保姆级)

MySQL Workbench 安装教程(保姆级)

《MySQLWorkbench安装教程(保姆级)》MySQLWorkbench是一款强大的数据库设计和管理工具,本文主要介绍了MySQLWorkbench安装教程,文中通过图文介绍的非常详细,对大... 目录前言：详细步骤：一、检查安装的数据库版本二、在官网下载对应的mysql Workbench版本，要是
阅读更多...
SpringBoot配置Ollama实现本地部署DeepSeek

SpringBoot配置Ollama实现本地部署DeepSeek

《SpringBoot配置Ollama实现本地部署DeepSeek》本文主要介绍了在本地环境中使用Ollama配置DeepSeek模型,并在IntelliJIDEA中创建一个Sprin... 目录前言详细步骤一、本地配置DeepSeek二、SpringBoot项目调用本地DeepSeek前言随着人工智能技
阅读更多...
python+opencv处理颜色之将目标颜色转换实例代码

python+opencv处理颜色之将目标颜色转换实例代码

《python+opencv处理颜色之将目标颜色转换实例代码》OpenCV是一个的跨平台计算机视觉库,可以运行在Linux、Windows和MacOS操作系统上,：本文主要介绍python+ope... 目录下面是代码+ 效果 + 解释转HSV： 关于颜色总是要转HSV的掩膜再标注总结 目标：将红色的部分滤
阅读更多...

Copyright China编程 23002807@qq.com

沪ICP备2023033072号-2