[Model] add pptracking model

.gitignore

@@ -13,12 +13,14 @@ fastdeploy/version.py
 fastdeploy/core/config.h
 python/fastdeploy/c_lib_wrap.py
 python/fastdeploy/LICENSE*
+python/build_cpu.sh
 python/fastdeploy/ThirdPartyNotices*
 *.so*
 fpython/astdeploy/libs/third_libs
 fastdeploy/core/config.h
 fastdeploy/pybind/main.cc
 python/fastdeploy/libs/lib*
+python/fastdeploy/libs/third_libs
 __pycache__
 build_fd_android.sh
 python/scripts/process_libraries.py

cmake/opencv.cmake

@@ -186,6 +186,6 @@ else()
     endif()
     find_package(OpenCV REQUIRED PATHS ${OpenCV_DIR})
     include_directories(${OpenCV_INCLUDE_DIRS})
-    list(APPEND DEPEND_LIBS opencv_core opencv_highgui opencv_imgproc opencv_imgcodecs)
+    list(APPEND DEPEND_LIBS opencv_core opencv_video opencv_highgui opencv_imgproc opencv_imgcodecs)
   endif()
 endif()

docs/api/vision_results/README.md

@@ -2,13 +2,14 @@
 
 FastDeploy根据视觉模型的任务类型，定义了不同的结构体(`fastdeploy/vision/common/result.h`)来表达模型预测结果，具体如下表所示
 
-| 结构体 | 文档 | 说明 | 相应模型 |
-| :----- | :--- | :---- | :------- |
-| ClassifyResult | [C++/Python文档](./classification_result.md) | 图像分类返回结果 | ResNet50、MobileNetV3等 |
-| SegmentationResult | [C++/Python文档](./segmentation_result.md) | 图像分割返回结果 | PP-HumanSeg、PP-LiteSeg等 |
-| DetectionResult | [C++/Python文档](./detection_result.md) | 目标检测返回结果 | PP-YOLOE、YOLOv7系列模型等 |
-| FaceDetectionResult | [C++/Python文档](./face_detection_result.md) | 目标检测返回结果 | SCRFD、RetinaFace系列模型等 |
-| KeyPointDetectionResult | [C++/Python文档](./keypointdetection_result.md) | 关键点检测返回结果 | PP-Tinypose系列模型等 |
-| FaceRecognitionResult | [C++/Python文档](./face_recognition_result.md) | 目标检测返回结果 | ArcFace、CosFace系列模型等 |
-| MattingResult | [C++/Python文档](./matting_result.md) | 目标检测返回结果 | MODNet系列模型等 |
-| OCRResult | [C++/Python文档](./ocr_result.md) | 文本框检测，分类和文本识别返回结果 | OCR系列模型等 |
+| 结构体                     | 文档                                            | 说明                | 相应模型                    |
+|:------------------------|:----------------------------------------------|:------------------|:------------------------|
+| ClassifyResult          | [C++/Python文档](./classification_result.md)    | 图像分类返回结果          | ResNet50、MobileNetV3等   |
+| SegmentationResult      | [C++/Python文档](./segmentation_result.md)      | 图像分割返回结果          | PP-HumanSeg、PP-LiteSeg等 |
+| DetectionResult         | [C++/Python文档](./detection_result.md)         | 目标检测返回结果          | PP-YOLOE、YOLOv7系列模型等    |
+| FaceDetectionResult     | [C++/Python文档](./face_detection_result.md)    | 目标检测返回结果          | SCRFD、RetinaFace系列模型等   |
+| KeyPointDetectionResult | [C++/Python文档](./keypointdetection_result.md) | 关键点检测返回结果         | PP-Tinypose系列模型等        |
+| FaceRecognitionResult   | [C++/Python文档](./face_recognition_result.md)  | 目标检测返回结果          | ArcFace、CosFace系列模型等    |
+| MattingResult           | [C++/Python文档](./matting_result.md)           | 目标检测返回结果          | MODNet系列模型等             |
+| OCRResult               | [C++/Python文档](./ocr_result.md)               | 文本框检测，分类和文本识别返回结果 | OCR系列模型等                |
+| MOTResult               | [C++/Python文档](./mot_result.md)               | 多目标跟踪返回结果         | pptracking系列模型等         |

docs/api/vision_results/mot_result.md

@@ -0,0 +1,40 @@
+# MOTResult 多目标跟踪结果
+
+MOTResult代码定义在`fastdeploy/vision/common/result.h`中，用于表明多目标跟踪中的检测出来的目标框、目标跟踪id、目标类别和目标置信度。
+
+## C++ 定义
+
+```c++
+fastdeploy::vision::MOTResult
+```  
+
+```c++
+struct MOTResult{
+  // left top right bottom
+  std::vector<std::array<int, 4>> boxes;
+  std::vector<int> ids;
+  std::vector<float> scores;
+  std::vector<int> class_ids;
+  void Clear();
+  std::string Str();
+};
+```
+
+- **boxes**: 成员变量，表示单帧画面中检测出来的所有目标框坐标，`boxes.size()`表示框的个数，每个框以4个float数值依次表示xmin, ymin, xmax, ymax， 即左上角和右下角坐标
+- **ids**: 成员变量，表示单帧画面中所有目标的id，其元素个数与`boxes.size()`一致
+- **scores**: 成员变量，表示单帧画面检测出来的所有目标置信度，其元素个数与`boxes.size()`一致
+- **class_ids**: 成员变量，表示单帧画面出来的所有目标类别，其元素个数与`boxes.size()`一致
+- **Clear()**: 成员函数，用于清除结构体中存储的结果
+- **Str()**: 成员函数，将结构体中的信息以字符串形式输出（用于Debug）
+
+## Python 定义
+
+```python
+fastdeploy.vision.MOTResult
+```
+
+- **boxes**(list of list(float)): 成员变量，表示单帧画面中检测出来的所有目标框坐标。boxes是一个list，其每个元素为一个长度为4的list， 表示为一个框，每个框以4个float数值依次表示xmin, ymin, xmax, ymax， 即左上角和右下角坐标
+- **ids**(list of list(float)):成员变量，表示单帧画面中所有目标的id，其元素个数与`boxes`一致
+- **scores**(list of float): 成员变量，表示单帧画面检测出来的所有目标置信度
+- **class_ids**(list of int): 成员变量，表示单帧画面出来的所有目标类别
+

examples/vision/tracking/pptracking/cpp/CMakeLists.txt

@@ -0,0 +1,14 @@
+PROJECT(infer_demo C CXX)
+CMAKE_MINIMUM_REQUIRED (VERSION 3.10)
+
+# 指定下载解压后的fastdeploy库路径
+option(FASTDEPLOY_INSTALL_DIR "Path of downloaded fastdeploy sdk.")
+
+include(${FASTDEPLOY_INSTALL_DIR}/FastDeploy.cmake)
+
+# 添加FastDeploy依赖头文件
+include_directories(${FASTDEPLOY_INCS})
+
+add_executable(infer_demo ${PROJECT_SOURCE_DIR}/infer.cc)
+# 添加FastDeploy库依赖
+target_link_libraries(infer_demo ${FASTDEPLOY_LIBS})

examples/vision/tracking/pptracking/cpp/README.md

@@ -0,0 +1,79 @@
+# PP-Tracking C++部署示例
+
+本目录下提供`infer.cc`快速完成PP-Tracking在CPU/GPU，以及GPU上通过TensorRT加速部署的示例。
+
+在部署前，需确认以下两个步骤
+
+- 1. 软硬件环境满足要求，参考[FastDeploy环境要求](../../../../../docs/cn/build_and_install/download_prebuilt_libraries.md)  
+- 2. 根据开发环境，下载预编译部署库和samples代码，参考[FastDeploy预编译库](../../../../../docs/cn/build_and_install/download_prebuilt_libraries.md)
+
+以Linux上 PP-Tracking 推理为例，在本目录执行如下命令即可完成编译测试（如若只需在CPU上部署，可在[Fastdeploy C++预编译库](../../../../../docs/cn/build_and_install/download_prebuilt_libraries.md/CPP_prebuilt_libraries.md)下载CPU推理库）
+
+```bash
+#下载SDK，编译模型examples代码（SDK中包含了examples代码）
+wget https://bj.bcebos.com/fastdeploy/release/cpp/fastdeploy-linux-x64-gpu-0.3.0.tgz
+tar xvf fastdeploy-linux-x64-gpu-0.3.0.tgz
+cd fastdeploy-linux-x64-gpu-0.3.0/examples/vision/tracking/pptracking/cpp/
+mkdir build && cd build
+cmake .. -DFASTDEPLOY_INSTALL_DIR=${PWD}/../../../../../../../fastdeploy-linux-x64-gpu-0.3.0
+make -j
+
+# 下载PP-Tracking模型文件和测试视频
+wget https://bj.bcebos.com/paddlehub/fastdeploy/fairmot_hrnetv2_w18_dlafpn_30e_576x320.tgz
+tar -xvf fairmot_hrnetv2_w18_dlafpn_30e_576x320.tgz
+wget https://bj.bcebos.com/paddlehub/fastdeploy/person.mp4
+wget https://bj.bcebos.com/paddlehub/fastdeploy/person.mp4
+
+
+# CPU推理
+./infer_demo fairmot_hrnetv2_w18_dlafpn_30e_576x320 person.mp4 0
+# GPU推理
+./infer_demo fairmot_hrnetv2_w18_dlafpn_30e_576x320 person.mp4 1
+# GPU上TensorRT推理
+./infer_demo fairmot_hrnetv2_w18_dlafpn_30e_576x320 person.mp4 2
+```
+
+以上命令只适用于Linux或MacOS, Windows下SDK的使用方式请参考:  
+- [如何在Windows中使用FastDeploy C++ SDK](../../../../../docs/cn/faq/use_sdk_on_windows.md)
+
+## PP-Tracking C++接口
+
+### PPTracking类
+
+```c++
+fastdeploy::vision::tracking::PPTracking(
+        const string& model_file,
+        const string& params_file = "",
+        const string& config_file,
+        const RuntimeOption& runtime_option = RuntimeOption(),
+        const ModelFormat& model_format = ModelFormat::PADDLE)
+```
+
+PP-Tracking模型加载和初始化，其中model_file为导出的Paddle模型格式。
+
+**参数**
+
+> * **model_file**(str): 模型文件路径
+> * **params_file**(str): 参数文件路径
+> * **config_file**(str): 推理部署配置文件
+> * **runtime_option**(RuntimeOption): 后端推理配置，默认为None，即采用默认配置
+> * **model_format**(ModelFormat): 模型格式，默认为Paddle格式
+
+#### Predict函数
+
+> ```c++
+> PPTracking::Predict(cv::Mat* im, MOTResult* result)
+> ```
+>
+> 模型预测接口，输入图像直接输出检测结果。
+>
+> **参数**
+>
+> > * **im**: 输入图像，注意需为HWC，BGR格式
+> > * **result**: 检测结果，包括检测框，跟踪id，各个框的置信度，对象类别id，MOTResult说明参考[视觉模型预测结果](../../../../../docs/api/vision_results/)
+
+
+- [模型介绍](../../)
+- [Python部署](../python)
+- [视觉模型预测结果](../../../../../docs/api/vision_results/)
+- [如何切换模型推理后端引擎](../../../../../docs/cn/faq/how_to_change_backend.md)

examples/vision/tracking/pptracking/cpp/infer.cc

@@ -0,0 +1,158 @@
+// Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#include "fastdeploy/vision.h"
+
+#ifdef WIN32
+const char sep = '\\';
+#else
+const char sep = '/';
+#endif
+
+void CpuInfer(const std::string& model_dir, const std::string& video_file) {
+  auto model_file = model_dir + sep + "model.pdmodel";
+  auto params_file = model_dir + sep + "model.pdiparams";
+  auto config_file = model_dir + sep + "infer_cfg.yml";
+  auto model = fastdeploy::vision::tracking::PPTracking(
+      model_file, params_file, config_file);
+
+  if (!model.Initialized()) {
+    std::cerr << "Failed to initialize." << std::endl;
+    return;
+  }
+
+  fastdeploy::vision::MOTResult result;
+  cv::Mat frame;
+  int frame_id=0;
+  cv::VideoCapture capture(video_file);
+  // according to the time of prediction to calculate fps
+  float fps= 0.0f;
+  while (capture.read(frame)) {
+    if (frame.empty()) {
+        break;
+    }
+    if (!model.Predict(&frame, &result)) {
+        std::cerr << "Failed to predict." << std::endl;
+        return;
+    }
+    // std::cout << result.Str() << std::endl;
+    cv::Mat out_img = fastdeploy::vision::VisMOT(frame, result, fps , frame_id);
+    cv::imshow("mot",out_img);
+    cv::waitKey(30);
+    frame_id++;
+  }
+  capture.release();
+  cv::destroyAllWindows();
+}
+
+void GpuInfer(const std::string& model_dir, const std::string& video_file) {
+  auto model_file = model_dir + sep + "model.pdmodel";
+  auto params_file = model_dir + sep + "model.pdiparams";
+  auto config_file = model_dir + sep + "infer_cfg.yml";
+
+  auto option = fastdeploy::RuntimeOption();
+  option.UseGpu();
+  auto model = fastdeploy::vision::tracking::PPTracking(
+      model_file, params_file, config_file, option);
+
+  if (!model.Initialized()) {
+    std::cerr << "Failed to initialize." << std::endl;
+    return;
+  }
+
+  fastdeploy::vision::MOTResult result;
+  cv::Mat frame;
+  int frame_id=0;
+  cv::VideoCapture capture(video_file);
+  // according to the time of prediction to calculate fps
+  float fps= 0.0f;
+  while (capture.read(frame)) {
+    if (frame.empty()) {
+        break;
+    }
+    if (!model.Predict(&frame, &result)) {
+        std::cerr << "Failed to predict." << std::endl;
+        return;
+    }
+    // std::cout << result.Str() << std::endl;
+    cv::Mat out_img = fastdeploy::vision::VisMOT(frame, result, fps , frame_id);
+    cv::imshow("mot",out_img);
+    cv::waitKey(30);
+    frame_id++;
+  }
+  capture.release();
+  cv::destroyAllWindows();
+}
+
+void TrtInfer(const std::string& model_dir, const std::string& video_file) {
+  auto model_file = model_dir + sep + "model.pdmodel";
+  auto params_file = model_dir + sep + "model.pdiparams";
+  auto config_file = model_dir + sep + "infer_cfg.yml";
+
+  auto option = fastdeploy::RuntimeOption();
+  option.UseGpu();
+  option.UseTrtBackend();
+  auto model = fastdeploy::vision::tracking::PPTracking(
+      model_file, params_file, config_file, option);
+
+  if (!model.Initialized()) {
+    std::cerr << "Failed to initialize." << std::endl;
+    return;
+  }
+
+  fastdeploy::vision::MOTResult result;
+  cv::Mat frame;
+  int frame_id=0;
+  cv::VideoCapture capture(video_file);
+  // according to the time of prediction to calculate fps
+  float fps= 0.0f;
+  while (capture.read(frame)) {
+    if (frame.empty()) {
+        break;
+    }
+    if (!model.Predict(&frame, &result)) {
+        std::cerr << "Failed to predict." << std::endl;
+        return;
+    }
+    // std::cout << result.Str() << std::endl;
+    cv::Mat out_img = fastdeploy::vision::VisMOT(frame, result, fps , frame_id);
+    cv::imshow("mot",out_img);
+    cv::waitKey(30);
+    frame_id++;
+  }
+  capture.release();
+  cv::destroyAllWindows();
+}
+
+int main(int argc, char* argv[]) {
+  if (argc < 4) {
+    std::cout
+        << "Usage: infer_demo path/to/model_dir path/to/video run_option, "
+           "e.g ./infer_model ./pptracking_model_dir ./person.mp4 0"
+        << std::endl;
+    std::cout << "The data type of run_option is int, 0: run with cpu; 1: run "
+                 "with gpu; 2: run with gpu and use tensorrt backend."
+              << std::endl;
+    return -1;
+  }
+
+  if (std::atoi(argv[3]) == 0) {
+    CpuInfer(argv[1], argv[2]);
+  } else if (std::atoi(argv[3]) == 1) {
+    GpuInfer(argv[1], argv[2]);
+  } else if (std::atoi(argv[3]) == 2) {
+    TrtInfer(argv[1], argv[2]);
+  }
+  return 0;
+}