基于ROS应用Tensorflow实现物体识别

本文基于古月居《ROS机器人开发与实践》一书而来，应用robot_vision与tensorflow_object_detection功能包

1）连接摄像头到PC端，可以通过

roslaunch usb_cam usb_cam-test.launch 

测试摄像头

如果有问题可以参考https://blog.csdn.net/felaim/article/details/60146455

解决

2）安装Opencv

sudo apt-get update

sudo apt-get install ros-kinectic-vision-opencv libopencv-dev python-opencv

测试一下

rosrun robot_vision cv_bridge_test.py
rqt_image_view

3)安装Tensorflow

sudo apt-get install python-pip python-dev
sudo -H pip install --upgrade pip

export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-1.4.0-cp27-none-linux_x86_64.whl

pip --default-timeout=1000 install -U pip

sudo pip install --upgrade $TF_BINARY_URL

测试一下

import tensorflow as tf
hello = tf.constant('Hello, TensorFlow!')
sess = tf.Session()
print(sess.run(hello))

4）安装Ｔｅｎｓｏｒｆｌｏｗ　Ｏｂｊｅｃｔ　Ｄｅｔｅｃｔｉｏｎ　ＡＰＩ

sudo apt-get install protobuf-compiler python-pil python-lxml

sudo pip install jupyter

sudo pip install matplotlib

然后下载模型库

git clone https://github.com/tensorflow/models

可以先用迅雷下好再转到home目录下

编译库

（在research下编译）

注：安装jupyter时可能遇到错误，更新一下pip再安装，sudo -H pip install –upgrade pip

如果失败点击打开链接

需要重新更新jupyter

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

export PYTHONPATH=$PYTHONPATH:`pwd`:`pwd`/slim

python object_detection/builders/model_builder_test.py

5)运行识别

roslaunch usb_cam usb_cam-test.launch

roslaunch tensorflow_object_detection ros_tensorflow_object_detection.launch

rqt_image_view

后续学习

https://blog.csdn.net/wizardforcel/article/details/79293944

参考博客

https://blog.csdn.net/mwlwlm/article/details/75126670

https://blog.csdn.net/weixin_35654926/article/details/75452450