树莓派3B+(适合树莓派3B)运行Tensorflow Object Detection

一、系统安装

此处省略XX字，在我的博客树莓派3B+(适合树莓派3B) Qt 使用 Cmake C++ OpenCV，此处链接中具有详细的系统安装教程，此处延用当时的环境，源，系统，配置等

二、正式安装

1.Update Raspberry Pi

sudo apt-get update
sudo apt-get upgrade

完成如下，上次升过级，所以这次很快，第一次升级速度会很慢

2.Install TensorFlow

sudo apt-get install libatlas-base-dev
sudo pip3 install tensorflow
sudo pip3 install pillow lxml jupyter matplotlib cython
sudo apt-get install python-tk

TensorFlow 安装特别容易失败，下载内容为：
https://www.piwheels.org/simple/tensorflow/tensorflow-1.14.0-cp35-none-linux_armv7l.whl，注意我的是 Python 3.5，注意查看下载的网址文件


我利用PC通过某种方法下载好了此文件，并利用 FileZilla Pro 上传至树莓派，可以使用命令行安装，需要此文件可以评论留言索取

安装此文件

sudo pip3 install tensorflow-1.14.0-cp35-none-linux_armv7l.whl

ls 为查看当前目录文件

安装完成

导入 tensorflow 包，查看版本，弹出的警告是 tensorflow-1.14.0 在新版本语法更新而已，不用理睬

python
import tensorflow
tensorflow.__version__

安装剩余依赖

sudo pip3 install pillow lxml jupyter matplotlib cython
sudo apt-get install python-tk

安装完成

3.Install OpenCV

此处省略XX字，在我的博客树莓派3B+(适合树莓派3B) Qt 使用 Cmake C++ OpenCV，此处链接中具有详细的 Python OpenCV 安装教程，此处延用当时的环境，源，系统，配置等，由于已经安装，所以我可以直接导入

python3
import cv2
cv2.__version

4.Compile and install Protubuf

安装命令

sudo apt-get install protobuf-compiler

安装完成

5.Set up TensorFlow directory structure

下载太慢，下载不下来的话，可以使用 PC 本地下载并使用 FileZilla Pro 上传至树莓派

执行如下命令，并且我把文件夹重命名为 TensorFlowModels，反正你用啥方法就是要这个 models

unzip models-master.zip

为了方便管理文件夹，我把之前下载的压缩包都移进了这个文件夹下

下面对此 TensorFlowModels 进行环境配置
修改 bashrc 文件

sudo nano ~/.bashrc
export PYTHONPATH=$PYTHONPATH:/home/pi/TensorFlowModels/research:/home/pi/TensorFlowModels/research/slim

然后 Ctrl X → Y 保存退出

输入以下命令使环境变量生效

echo $PYTHONPATH

关闭 Terminal 再打开再次输入以下命令即可看到环境变量

echo $PYTHONPATH

进入 object_detection 文件夹并下载模型，点击此处下载模型，ssd_mobilenet_v2_coco 模型较小，选择并下载，可以 PC 本地下载并通过 FileZilla Pro 上传到树莓派，也可以右键复制链接在 Terminal 下载

cd TensorFlowModels/research/object_detection
wget http://download.tensorflow.org/models/object_detection/ssdlite_mobilenet_v2_coco_2018_05_09.tar.gz

解压

tar -xzvf ssdlite_mobilenet_v2_coco_2018_05_09.tar.gz

解压后添加label_map.pbtxt，只需要将以下文件复制过来并重命名为 label_map.pbtxt 就行

/home/pi/TensorFlowModels/research/object_detection/data/mscoco_label_map.pbtxt

在 research 文件夹下构建 protos

protoc object_detection/protos/*.proto --python_out=.

6.Test out object detector

程序部分
demo.image.py

import numpy as np
import matplotlib
import os
import sys
import tensorflow as tf
from PIL import Image
from matplotlib import pyplot as plt

# This is needed since the python file is stored in the object_detection folder.
sys.path.append("..")

from utils import label_map_util
from utils import visualization_utils as vis_util

matplotlib.use('TkAgg')

# What model to download.
MODEL_NAME = 'ssdlite_mobilenet_v2_coco_2018_05_09'

# Path to frozen detection graph. This is the actual model that is used for the object detection.
PATH_TO_CKPT = MODEL_NAME + '/frozen_inference_graph.pb'

# List of the strings that is used to add correct label for each box.
PATH_TO_LABELS = os.path.join('ssdlite_mobilenet_v2_coco_2018_05_09/label_map.pbtxt')

NUM_CLASSES = 90
    
detection_graph = tf.Graph()
with detection_graph.as_default():
  od_graph_def = tf.GraphDef()
  with tf.gfile.GFile(PATH_TO_CKPT, 'rb') as fid:
    serialized_graph = fid.read()
    od_graph_def.ParseFromString(serialized_graph)
    tf.import_graph_def(od_graph_def, name='')
    
label_map = label_map_util.load_labelmap(PATH_TO_LABELS)
categories = label_map_util.convert_label_map_to_categories(label_map, max_num_classes=NUM_CLASSES, use_display_name=True)
category_index = label_map_util.create_category_index(categories)

def load_image_into_numpy_array(image):
  (im_width, im_height) = image.size
  return np.array(image.getdata()).reshape(
      (im_height, im_width, 3)).astype(np.uint8)

# For the sake of simplicity we will use only 2 images:
# image1.jpg
# image2.jpg
# If you want to test the code with your images, just add path to the images to the TEST_IMAGE_PATHS.
PATH_TO_TEST_IMAGES_DIR = 'images'
TEST_IMAGE_PATHS = [ os.path.join(PATH_TO_TEST_IMAGES_DIR, 'image{}.jpg'.format(i)) for i in range(1, 2) ]

# Size, in inches, of the output images.
IMAGE_SIZE = (12, 8)

with detection_graph.as_default():
  with tf.Session(graph=detection_graph) as sess:
    # Definite input and output Tensors for detection_graph
    image_tensor = detection_graph.get_tensor_by_name('image_tensor:0')
    # Each box represents a part of the image where a particular object was detected.
    detection_boxes = detection_graph.get_tensor_by_name('detection_boxes:0')
    # Each score represent how level of confidence for each of the objects.
    # Score is shown on the result image, together with the class label.
    detection_scores = detection_graph.get_tensor_by_name('detection_scores:0')
    detection_classes = detection_graph.get_tensor_by_name('detection_classes:0')
    num_detections = detection_graph.get_tensor_by_name('num_detections:0')
    for image_path in TEST_IMAGE_PATHS:
      image = Image.open(image_path)
      # the array based representation of the image will be used later in order to prepare the
      # result image with boxes and labels on it.
      
      image_np = load_image_into_numpy_array(image)
      
      # Expand dimensions since the model expects images to have shape: [1, None, None, 3]
      
      #image_np_expanded = np.expand_dims(image_np, axis=0)
      image_np_expanded = np.expand_dims(image_np, axis=0)
      
      # Actual detection.
      (boxes, scores, classes, num) = sess.run(
          [detection_boxes, detection_scores, detection_classes, num_detections],
          feed_dict={image_tensor: image_np_expanded})
      # Visualization of the results of a detection.
      vis_util.visualize_boxes_and_labels_on_image_array(
          image_np,
          np.squeeze(boxes),
          np.squeeze(classes).astype(np.int32),
          np.squeeze(scores),
          category_index,
          use_normalized_coordinates=True,
          line_thickness=16,)
      plt.figure(figsize=IMAGE_SIZE)
      plt.imshow(image_np)
      plt.show()

运行结果

三、TensorFlow 1.14.0源码修正与字体大小修改

留意博主下篇文章

到此为止，在树莓派上运行 TensorFlow 就完成啦