devbox
我家自动化
这个屌丝很懒，什么也没留下！
关注作者
热门标签
热门文章
当前位置:   article > 正文

快速识别你家的猫猫狗狗，教你用ModelBox开发AI萌宠应用

作者：我家自动化 | 2024-06-17 03:40:39

赞

踩

快速识别你家的猫猫狗狗，教你用ModelBox开发AI萌宠应用

本文分享自华为云社区《ModelBox-AI应用开发：动物目标检测【玩转华为云】》，作者：阳光大猫。

一、准备环境

ModelBox端云协同AI开发套件（Windows）环境准备【视频教程】

二、应用开发

1. 创建工程

在ModelBox sdk目录下使用create.bat创建yolov7_pet工程

  1. (tensorflow) PS D:\modelbox-win10-x64-1.5.3> .\create.bat -t server -n yolov7_pet 
  2.  
  3. (tensorflow) D:\modelbox-win10-x64-1.5.3>set BASE_PATH=D:\modelbox-win10-x64-1.5.3\ 
  4.  
  5. (tensorflow) D:\modelbox-win10-x64-1.5.3>set PATH=D:\modelbox-win10-x64-1.5.3\\python-embed;C:\Users\yanso\miniconda3\envs\tensorflow\lib\site-packages\pywin32_system32;C:\Users\yanso\miniconda3\envs\tensorflow;C:\Users\yanso\miniconda3\envs\tensorflow\Library\mingw-w64\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Library\usr\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Library\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Scripts;C:\Users\yanso\miniconda3\envs\tensorflow\bin;C:\Users\yanso\miniconda3\condabin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin;C:\Users\yanso\miniconda3\envs\tensorflow\lib\site-packages\pywin32_system32;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Library\mingw-w64\bin;C:\Users\yanso\miniconda3\Library\usr\bin;C:\Users\yanso\miniconda3\Library\bin;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\bin;C:\Users\yanso\miniconda3\condabin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin
  6.  
  7. (tensorflow) D:\modelbox-win10-x64-1.5.3>set PYTHONPATH=
  8.  
  9. (tensorflow) D:\modelbox-win10-x64-1.5.3>set PYTHONHOME=
  10.  
  11. (tensorflow) D:\modelbox-win10-x64-1.5.3>python.exe -u D:\modelbox-win10-x64-1.5.3\\create.py -t server -n yolov7_pet
  12. sdk version is modelbox-win10-x64-1.5.3
  13. dos2unix: converting file D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet/graph\modelbox.conf to Unix format...
  14. dos2unix: converting file D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet/graph\yolov7_pet.toml to Unix format...
  15. dos2unix: converting file D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet/bin\mock_task.toml to Unix format...
  16. success: create yolov7_pet in D:\modelbox-win10-x64-1.5.3\workspace

create.bat工具的参数中，-t表示所创建实例的类型，包括server（ModelBox工程）、python（Python功能单元）、c++（C++功能单元）、infer（推理功能单元）等；-n表示所创建实例的名称，开发者自行命名。

2. 创建推理功能单元

在ModelBox sdk目录下使用create.bat创建yolov7_infer推理功能单元

  1. (tensorflow) PS D:\modelbox-win10-x64-1.5.3> .\create.bat -t infer -n yolov7_infer -p yolov7_pet  
  2.  
  3. (tensorflow) D:\modelbox-win10-x64-1.5.3>set BASE_PATH=D:\modelbox-win10-x64-1.5.3\ 
  4.  
  5. (tensorflow) D:\modelbox-win10-x64-1.5.3>set PATH=D:\modelbox-win10-x64-1.5.3\\python-embed;C:\Users\yanso\miniconda3\envs\tensorflow\lib\site-packages\pywin32_system32;C:\Users\yanso\miniconda3\envs\tensorflow;C:\Users\yanso\miniconda3\envs\tensorflow\Library\mingw-w64\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Library\usr\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Library\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Scripts;C:\Users\yanso\miniconda3\envs\tensorflow\bin;C:\Users\yanso\miniconda3\condabin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin;C:\Users\yanso\miniconda3\envs\tensorflow\lib\site-packages\pywin32_system32;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Library\mingw-w64\bin;C:\Users\yanso\miniconda3\Library\usr\bin;C:\Users\yanso\miniconda3\Library\bin;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\bin;C:\Users\yanso\miniconda3\condabin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin 
  6.  
  7. (tensorflow) D:\modelbox-win10-x64-1.5.3>set PYTHONPATH= 
  8.  
  9. (tensorflow) D:\modelbox-win10-x64-1.5.3>set PYTHONHOME= 
  10.  
  11. (tensorflow) D:\modelbox-win10-x64-1.5.3>python.exe -u D:\modelbox-win10-x64-1.5.3\\create.py -t infer -n yolov7_infer -p yolov7_pet    
  12. sdk version is modelbox-win10-x64-1.5.3
  13. success: create infer yolov7_infer in D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet/model/yolov7_infer

create.bat工具使用时，-t infer 即表示创建的是推理功能单元；-n xxx_infer 表示创建的功能单元名称为xxx_infer；-p yolov7_infer 表示所创建的功能单元属于yolov7_infer应用。

a. 下载转换好的模型

运行此Notebook下载转换好的ONNX格式模型

b. 修改模型配置文件

模型和配置文件保持在同级目录下

  1. # Copyright (C) 2020 Huawei Technologies Co., Ltd. All rights reserved.
  2.  
  3. [base]
  4. name = "yolov7_infer"
  5. device = "cpu"
  6. version = "1.0.0"
  7. description = "your description"
  8. entry = "./best.onnx"  # model file path, use relative path
  9. type = "inference" 
  10. virtual_type = "onnx" # inference engine type: win10 now only support onnx
  11. group_type = "Inference"  # flowunit group attribution, do not change
  12.  
  13. # Input ports description
  14. [input]
  15. [input.input1]  # input port number, Format is input.input[N]
  16. name = "Input"  # input port name
  17. type = "float"  # input port data type ,e.g. float or uint8
  18. device = "cpu"  # input buffer type: cpu, win10 now copy input from cpu
  19.  
  20. # Output ports description
  21. [output]
  22. [output.output1] # output port number, Format is output.output[N]
  23. name = "Output"  # output port name
  24. type = "float"   # output port data type ,e.g. float or uint8

3. 创建后处理功能单元

在ModelBox sdk目录下使用create.bat创建yolov7_post后处理功能单元

  1. (tensorflow) PS D:\modelbox-win10-x64-1.5.3> .\create.bat -t python -n yolov7_post -p yolov7_pet  
  2.  
  3. (tensorflow) D:\modelbox-win10-x64-1.5.3>set BASE_PATH=D:\modelbox-win10-x64-1.5.3\ 
  4.  
  5. (tensorflow) D:\modelbox-win10-x64-1.5.3>set PATH=D:\modelbox-win10-x64-1.5.3\\python-embed;C:\Users\yanso\miniconda3\envs\tensorflow\lib\site-packages\pywin32_system32;C:\Users\yanso\miniconda3\envs\tensorflow;C:\Users\yanso\miniconda3\envs\tensorflow\Library\mingw-w64\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Library\usr\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Library\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Scripts;C:\Users\yanso\miniconda3\envs\tensorflow\bin;C:\Users\yanso\miniconda3\condabin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin;C:\Users\yanso\miniconda3\envs\tensorflow\lib\site-packages\pywin32_system32;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Library\mingw-w64\bin;C:\Users\yanso\miniconda3\Library\usr\bin;C:\Users\yanso\miniconda3\Library\bin;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\bin;C:\Users\yanso\miniconda3\condabin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin
  6.  
  7. (tensorflow) D:\modelbox-win10-x64-1.5.3>set PYTHONPATH=
  8.  
  9. (tensorflow) D:\modelbox-win10-x64-1.5.3>set PYTHONHOME=
  10.  
  11. (tensorflow) D:\modelbox-win10-x64-1.5.3>python.exe -u D:\modelbox-win10-x64-1.5.3\\create.py -t python -n yolov7_post -p yolov7_pet
  12. sdk version is modelbox-win10-x64-1.5.3
  13. success: create python yolov7_post in D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet/etc/flowunit/yolov7_post

a. 修改配置文件

  1. # Copyright (c) Huawei Technologies Co., Ltd. 2022. All rights reserved.
  2.  
  3. # Basic config
  4. [base]
  5. name = "yolov7_post" # The FlowUnit name
  6. device = "cpu" # The flowunit runs on cpu
  7. version = "1.0.0" # The version of the flowunit
  8. type = "python" # Fixed value, do not change
  9. description = "description" # The description of the flowunit
  10. entry = "yolov7_post@yolov7_postFlowUnit" # Python flowunit entry function
  11. group_type = "Generic"  # flowunit group attribution, change as Input/Output/Image/Generic ...
  12.  
  13. # Flowunit Type
  14. stream = false # Whether the flowunit is a stream flowunit
  15. condition = false # Whether the flowunit is a condition flowunit
  16. collapse = false # Whether the flowunit is a collapse flowunit
  17. collapse_all = false # Whether the flowunit will collapse all the data
  18. expand = false #  Whether the flowunit is a expand flowunit
  19.  
  20. # The default Flowunit config
  21. [config]
  22. net_h = 640
  23. net_w = 640
  24. num_classes = 2
  25. conf_threshold = 0.5
  26. iou_threshold = 0.45
  27.  
  28. # Input ports description
  29. [input]
  30. [input.input1] # Input port number, the format is input.input[N]
  31. name = "in_feat" # Input port name
  32. type = "float" # Input port type
  33.  
  34. # Output ports description
  35. [output]
  36. [output.output1] # Output port number, the format is output.output[N]
  37. name = "out_data" # Output port name
  38. type = "string" # Output port type

b. 修改逻辑代码

  1. # Copyright (c) Huawei Technologies Co., Ltd. 2022. All rights reserved.
  2.  
  3. #!/usr/bin/env python
  4. # -*- coding: utf-8 -*-
  5. import _flowunit as modelbox
  6. import numpy as np
  7. import json
  8. import cv2
  9.  
  10. class yolov7_postFlowUnit(modelbox.FlowUnit):
  11.     # Derived from modelbox.FlowUnit
  12.     def __init__(self):
  13.         super().__init__()
  14.  
  15.     # Open the flowunit to obtain configuration information
  16.     def open(self, config):
  17.         # 获取功能单元的配置参数
  18.         self.params = {}
  19.         self.params['net_h'] = config.get_int('net_h')
  20.         self.params['net_w'] = config.get_int('net_w')
  21.         self.params['num_classes'] = config.get_int('num_classes')
  22.         self.params['conf_thre'] = config.get_float('conf_threshold')
  23.         self.params['nms_thre'] = config.get_float('iou_threshold')
  24.         self.num_classes = config.get_int('num_classes')
  25.  
  26.         return modelbox.Status.StatusCode.STATUS_SUCCESS
  27.  
  28.     # Process the data
  29.     def process(self, data_context):
  30.         # 从DataContext中获取输入输出BufferList对象
  31.         in_feat = data_context.input("in_feat")
  32.         out_data = data_context.output("out_data")
  33.  
  34.         # yolov7_post process code.
  35.         # 循环处理每一个输入Buffer数据
  36.         for buffer_feat in in_feat:
  37.             # 将输入Buffer转换为numpy对象
  38.             feat_data = np.array(buffer_feat.as_object(), copy=False)
  39.             feat_data = feat_data.reshape((-1, self.num_classes + 5))
  40.  
  41.             # 业务处理：解码yolov7模型的输出数据，得到检测框，转化为json数据
  42.             bboxes = self.postprocess(feat_data, self.params)
  43.             result = {"det_result": str(bboxes)}
  44.             print(result)
  45.  
  46.             # 将业务处理返回的结果数据转换为Buffer
  47.             result_str = json.dumps(result)
  48.             out_buffer = modelbox.Buffer(self.get_bind_device(), result_str)
  49.  
  50.             # 将输出Buffer放入输出BufferList中
  51.             out_data.push_back(out_buffer)
  52.  
  53.         return modelbox.Status.StatusCode.STATUS_SUCCESS
  54.     
  55.     # model post-processing function
  56.     def postprocess(self, feat_data, params):
  57.         """postprocess for yolo7 model"""
  58.         boxes = []
  59.         class_ids = []
  60.         confidences = []
  61.         for detection in feat_data:
  62.             scores = detection[5:]
  63.             class_id = np.argmax(scores)
  64.             if params['num_classes'] == 1:
  65.                 confidence = detection[4]
  66.             else:
  67.                 confidence = detection[4] * scores[class_id] 
  68.             if confidence > params['conf_thre'] and detection[4] > params['conf_thre']:
  69.                 center_x = detection[0] / params['net_w']
  70.                 center_y = detection[1] / params['net_h']
  71.                 width = detection[2] / params['net_w']
  72.                 height = detection[3] / params['net_h']
  73.  
  74.                 left = center_x - width / 2
  75.                 top = center_y - height / 2
  76.  
  77.                 class_ids.append(class_id)
  78.                 confidences.append(confidence)
  79.                 boxes.append([left, top, width, height])
  80.  
  81.         # use nms algorithm in opencv
  82.         box_idx = cv2.dnn.NMSBoxes(
  83.             boxes, confidences, params['conf_thre'], params['nms_thre'])
  84.  
  85.         detections = []
  86.         for i in box_idx:
  87.             boxes[i][0] = max(0.0, boxes[i][0])  # [0, 1]
  88.             boxes[i][1] = max(0.0, boxes[i][1])  # [0, 1]
  89.             boxes[i][2] = min(1.0, boxes[i][0] + boxes[i][2])  # [0, 1]
  90.             boxes[i][3] = min(1.0, boxes[i][1] + boxes[i][3])  # [0, 1]
  91.             dets = np.concatenate(
  92.                 [boxes[i], np.array([confidences[i]]), np.array([class_ids[i]])], 0).tolist()
  93.             detections.append(dets)
  94.  
  95.         return detections
  96.  
  97.     def close(self):
  98.         # Close the flowunit
  99.         return modelbox.Status()
  100.  
  101.     def data_pre(self, data_context):
  102.         # Before streaming data starts
  103.         return modelbox.Status()
  104.  
  105.     def data_post(self, data_context):
  106.         # After streaming data ends
  107.         return modelbox.Status()
  108.  
  109.     def data_group_pre(self, data_context):
  110.         # Before all streaming data starts
  111.         return modelbox.Status()
  112.  
  113.     def data_group_post(self, data_context):
  114.         # After all streaming data ends
  115.         return modelbox.Status()

4. 修改流程图

yolov7_pet工程graph目录下存放流程图，默认的流程图yolov7_pet.toml与工程同名，其内容为（以Windows版ModelBox为例）：

  1. # Copyright (C) 2020 Huawei Technologies Co., Ltd. All rights reserved.
  2.  
  3. [driver]
  4. dir = ["${HILENS_APP_ROOT}/etc/flowunit",
  5. "${HILENS_APP_ROOT}/etc/flowunit/cpp",
  6. "${HILENS_APP_ROOT}/model",
  7. "${HILENS_MB_SDK_PATH}/flowunit"]
  8. skip-default = true
  9. [profile]
  10. profile=false
  11. trace=false
  12. dir="${HILENS_DATA_DIR}/mb_profile"
  13. [graph]
  14. format = "graphviz"
  15. graphconf = """digraph yolov7_pet {
  16.     node [shape=Mrecord]
  17.     queue_size = 4
  18.     batch_size = 1
  19.     input1[type=input,flowunit=input,device=cpu,deviceid=0]
  21.     httpserver_sync_receive[type=flowunit, flowunit=httpserver_sync_receive_v2, device=cpu, deviceid=0, time_out_ms=5000, endpoint="http://0.0.0.0:8083/v1/yolov7_pet", max_requests=100]
  22.     image_decoder[type=flowunit, flowunit=image_decoder, device=cpu, key="image_base64", queue_size=4]
  23.     image_resize[type=flowunit, flowunit=resize, device=cpu, deviceid=0, image_width=640, image_height=640]
  24.     image_transpose[type=flowunit, flowunit=packed_planar_transpose, device=cpu, deviceid=0]
  25.     normalize[type=flowunit flowunit=normalize device=cpu deviceid=0 standard_deviation_inverse="0.0039215686,0.0039215686,0.0039215686"]
  26.     yolov7_infer[type=flowunit, flowunit=yolov7_infer, device=cpu, deviceid=0, batch_size = 1]
  27.     yolov7_post[type=flowunit, flowunit=yolov7_post, device=cpu, deviceid=0]
  28.     httpserver_sync_reply[type=flowunit, flowunit=httpserver_sync_reply_v2, device=cpu, deviceid=0]
  29.     
  30.     input1:input -> httpserver_sync_receive:in_url
  31.     httpserver_sync_receive:out_request_info -> image_decoder:in_encoded_image
  32.     image_decoder:out_image -> image_resize:in_image
  33.     image_resize:out_image -> image_transpose:in_image
  34.     image_transpose:out_image -> normalize:in_data
  35.     normalize:out_data -> yolov7_infer:Input
  36.     yolov7_infer:Output -> yolov7_post:in_feat
  37.     yolov7_post:out_data -> httpserver_sync_reply:in_reply_info
  38. }"""
  39. [flow]
  40. desc = "yolov7_pet run in modelbox-win10-x64"

5. 准备动物图片和测试脚本

a. 动物图片

yolov7_pet工程data目录下存放动物图片文件夹test_imgs

b. 测试脚本

yolov7_pet工程data目录下存放测试脚本test_http.py

  1. #!/usr/bin/env python
  2. # -*- coding: utf-8 -*-
  3.  
  4. # Copyright (c) Huawei Technologies Co., Ltd. 2022. All rights reserved.
  5.  
  6. import os
  7. import cv2
  8. import json
  9. import base64
  10. import http.client
  11. class HttpConfig:
  12.     '''http调用的参数配置'''
  13.     def __init__(self, host_ip, port, url, img_base64_str):
  14.         self.hostIP = host_ip
  15.         self.Port = port
  16.  
  17.         self.httpMethod = "POST"
  18.         self.requstURL = url
  19.         self.headerdata = {
  20.             "Content-Type": "application/json"
  21.         }
  22.         self.test_data = {
  23.             "image_base64": img_base64_str
  24.         }
  25.         self.body = json.dumps(self.test_data)
  26. def read_image(img_path):
  27.     '''读取图片数据并转为base64编码的字符串'''
  28.     img_data = cv2.imread(img_path)
  29.     img_str = cv2.imencode('.jpg', img_data)[1].tostring()
  30.     img_bin = base64.b64encode(img_str)
  31.     img_base64_str = str(img_bin, encoding='utf8')
  32.     return img_data, img_base64_str
  33. def decode_car_bboxes(bbox_str, input_shape):
  34.     try:
  35.         labels = [0, 1]  # cat, dog
  36.         bboxes = json.loads(json.loads(bbox_str)['det_result'])
  37.         bboxes = list(filter(lambda x: int(x[5]) in labels, bboxes))
  38.     except Exception as ex:
  39.         print(str(ex))
  40.         return []
  41.     else:
  42.         for bbox in bboxes:
  43.             bbox[0] = int(bbox[0] * input_shape[1])
  44.             bbox[1] = int(bbox[1] * input_shape[0])
  45.             bbox[2] = int(bbox[2] * input_shape[1])
  46.             bbox[3] = int(bbox[3] * input_shape[0])
  47.         return bboxes
  48. def draw_bboxes(img_data, bboxes):
  49.     '''画框'''
  50.     for bbox in bboxes:
  51.         x1, y1, x2, y2, score, label = bbox
  52.         color = (0, 0, 255)
  53.         names = ['cat', 'dog']  
  54.         score = '%.2f' % score
  55.         label = '%s:%s' % (names[int(label)], score)
  56.         cv2.rectangle(img_data, (x1, y1), (x2, y2), color, 2)
  57.         cv2.putText(img_data, label, (x1, y1 - 10), cv2.FONT_HERSHEY_TRIPLEX, 0.5, (0, 255, 0), thickness=1)
  58.     return img_data
  59. def test_image(img_path, ip, port, url):
  60.     '''单张图片测试'''
  61.     img_data, img_base64_str = read_image(img_path)
  62.     http_config = HttpConfig(ip, port, url, img_base64_str)
  63.  
  64.     conn = http.client.HTTPConnection(host=http_config.hostIP, port=http_config.Port)
  65.     conn.request(method=http_config.httpMethod, url=http_config.requstURL,
  66.                 body=http_config.body, headers=http_config.headerdata)
  67.  
  68.     response = conn.getresponse().read().decode()
  69.     print('response: ', response)
  70.  
  71.     bboxes = decode_car_bboxes(response, img_data.shape)
  72.     imt_out = draw_bboxes(img_data, bboxes)
  73.     cv2.imwrite('./result-' + os.path.basename(img_path), imt_out)
  74. if __name__ == "__main__":
  75.     port = 8083
  76.     ip = "127.0.0.1"
  77.     url = "/v1/yolov7_pet"
  78.     img_path = "./test.jpg"
  79.     img_folder = './test_imgs'
  80.     file_list = os.listdir(img_folder)
  81.     for img_file in file_list:
  82.         print("\n================ {} ================".format(img_file))
  83.         img_path = os.path.join(img_folder, img_file)
  84.         test_image(img_path, ip, port, url)

三、运行应用

在yolov7_pet工程目录下执行.\bin\main.bat运行应用：

  1. (tensorflow) PS D:\modelbox-win10-x64-1.5.3> cd D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet
  2. (tensorflow) PS D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet> .\bin\main.bat
  3.  
  4. (tensorflow) D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet>set PATH=D:/modelbox-win10-x64-1.5.3/workspace/yolov7_pet/bin/../../../python-embed;D:/modelbox-win10-x64-1.5.3/workspace/yolov7_pet/bin/../../../modelbox-win10-x64/bin;D:/modelbox-win10-x64-1.5.3/workspace/yolov7_pet/bin/../dependence/lib;C:\Users\yanso\miniconda3\envs\tensorflow\lib\site-packages\pywin32_system32;C:\Users\yanso\miniconda3\envs\tensorflow;C:\Users\yanso\miniconda3\envs\tensorflow\Library\mingw-w64\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Library\usr\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Library\bin;C:\Users\yanso\miniconda3\envs\tensorflow\Scripts;C:\Users\yanso\miniconda3\envs\tensorflow\bin;C:\Users\yanso\miniconda3\condabin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin;C:\Users\yanso\miniconda3\envs\tensorflow\lib\site-packages\pywin32_system32;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Library\mingw-w64\bin;C:\Users\yanso\miniconda3\Library\usr\bin;C:\Users\yanso\miniconda3\Library\bin;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\bin;C:\Users\yanso\miniconda3\condabin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin;C:\Windows\System32\HWAudioDriverLibs;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0;C:\Windows\System32\OpenSSH;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0;C:\WINDOWS\System32\OpenSSH;C:\Program Files\Git\cmd;C:\Users\yanso\miniconda3;C:\Users\yanso\miniconda3\Scripts;C:\Users\yanso\miniconda3\Library\bin;.;C:\Program Files\Git LFS;C:\Users\yanso\AppData\Local\Microsoft\WindowsApps;.;C:\Users\yanso\AppData\Local\Programs\Microsoft VS Code\bin 
  5.  
  6. (tensorflow) D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet>modelbox.exe -c D:/modelbox-win10-x64-1.5.3/workspace/yolov7_pet/bin/../graph/modelbox.conf 
  7. [2024-06-10 06:42:50,922][ WARN][    iva_config.cc:143 ] update vas url failed. Fault, no vas projectid or iva endpoint
  8. open log file D:/modelbox-win10-x64-1.5.3/workspace/yolov7_pet/bin/../hilens_data_dir/log/modelbox.log failed, No error
  9. input dims is:1,3,640,640,
  10. output dims is:1,25200,7,

HTTP服务启动后可以在另一个终端进行请求测试，进入yolov7_pet工程目录data文件夹中使用test_http.py脚本发起HTTP请求进行测试：

  1. (tensorflow) PS D:\modelbox-win10-x64-1.5.3> cd D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet\data                                                        
  2. (tensorflow) PS D:\modelbox-win10-x64-1.5.3\workspace\yolov7_pet\data> python .\test_http.py                                                                 
  3.  
  4. ================ Abyssinian_1.jpg ================
  5. .\test_http.py:33: DeprecationWarning: tostring() is deprecated. Use tobytes() instead.
  6.   img_str = cv2.imencode('.jpg', img_data)[1].tostring()
  7. response:  {"det_result": "[[0.554308044910431, 0.1864600658416748, 0.7089953303337098, 0.3776256084442139, 0.82369065284729, 0.0]]"}
  8.  
  9. ================ saint_bernard_143.jpg ================
  10. response:  {"det_result": "[[0.46182055473327643, 0.30239262580871584, 0.8193012714385988, 0.4969032764434815, 0.7603430151939392, 1.0]]"}

四、小结

本章我们介绍了如何使用ModelBox开发一个动物目标检测的AI应用，我们只需要准备模型文件以及简单的配置即可创建一个HTTP服务。同时我们可以了解到图片标注、数据处理和模型训练方法，以及对应的推理应用逻辑。

HDC 2024，6月21日-23日，东莞松山湖，期待与您相见！

更多详情请关注官网：

中文：华为开发者大会2024_HDC 2024

英文：华为开发者大会2024_HDC 2024

点击关注，第一时间了解华为云新鲜技术~

声明：本文内容由网友自发贡献，不代表【wpsshop博客】立场，版权归原作者所有，本站不承担相应法律责任。如您发现有侵权的内容，请联系我们。转载请注明出处：https://www.wpsshop.cn/w/我家自动化/article/detail/729505
推荐阅读
相关标签

Copyright © 2003-2013 www.wpsshop.cn 版权所有，并保留所有权利。

  

闽ICP备14008679号