【Jetson Tx2 视觉slam开发】入门配置二

https://blog.csdn.net/qq_14925953/article/details/73744719

Jetson Tx2安装ROS
https://github.com/jetsonhacks/installROSTX2
updateRepositories.sh This is an optional step. Adds the repositories universe, multiverse, and restricted and then apt-get update. These repositories are in the standard 27.1 install, so probably not needed.

installROS.sh Adds the ROS sources list, sets the keys and then loads ros-kinetic-ros-base. Edit this file to add other ROS packages for your installation. After loading, rosdep is initialized.

setupCatkinWorkspace.sh Usage:

$ ./setupCatkinWorkspace.sh [optionalWorkspaceName]

具体操作：
安装可以参照 http://wiki.ros.org/kinetic/Installation/Ubuntu，或者根据下面所示进行安装。

#!/bin/bash
# Install Robot Operating System (ROS) on NVIDIA Jetson TX2
# Maintainer of ARM builds for ROS is http://answers.ros.org/users/1034/ahendrix/
# Information from:
# http://wiki.ros.org/kinetic/Installation/UbuntuARM

# Setup sources.lst
sudo sh -c 'echo "deb http://packages.ros.org/ros/ubuntu $(lsb_release -sc) main" > /etc/apt/sources.list.d/ros-latest.list'
# Setup keys
sudo apt-key adv --keyserver hkp://ha.pool.sks-keyservers.net:80 --recv-key 0xB01FA116
# Installation
sudo apt-get update
sudo apt-get install ros-kinetic-ros-base -y or sudo apt-get install ros-kinetic-ros-desktop -y
# Add Individual Packages here
# You can install a specific ROS package (replace underscores with dashes of the package name):
# sudo apt-get install ros-kinetic-PACKAGE
# e.g.
# sudo apt-get install ros-kinetic-navigation
#
# To find available packages:
# apt-cache search ros-kinetic
# 
# Initialize rosdep
sudo apt-get install python-rosdep -y
# Certificates are messed up
sudo c_rehash /etc/ssl/certs
# Initialize rosdep
sudo rosdep init
# To find available packages, use:
rosdep update
# Environment Setup
echo "source /opt/ros/kinetic/setup.bash" >> ~/.bashrc
source ~/.bashrc
# Install rosinstall
sudo apt-get install python-rosinstall -y

这里写图片描述

按照以上教程安装完成即可。

启动自带摄像头
打开Jetson TX2的终端

写入命令：nvgstcapture-1.0 ，摄像头就会起来了
1. –prev_res 预览视屏的分辨率，高度和宽度，用的是CSI摄像头的话范围是 2 to 12 (5632x4224)

nvgstcapture-1.0 --prev-res=3

–cus-prev-res 自定义预览分辨率，宽度和高度，仅支持CSI摄像头

nvgstcapture-1.0 --cus-prev-res=1920x1080

多个命令同时使用的话用！隔开，想关掉摄像头的额话，直接在终端输入q再按回车，想捕获图片的话，在终端输入j再按回车，图片将保存当前目录下。

Tensorflow安装

http://www.ncnynl.com/archives/201706/1754.html

git clone https://github.com/ncnynl/installTensorFlowTX2.git
cd installTensorFlowTX2

安装依赖：

chmod +x installPrerequisites.sh
 ./installPrerequisites.sh

下载tensorflow代码：

 ./cloneTensorFlow.sh

设置环境变量：

 ./setTensorFlowEV.sh

问题：提示junit.4.12.jar不能下载

解决：修改vim ~/tensorflow/tensorflow/workspace.bzl 更换地址，并去掉校验

 native.http_jar(
      name = "junit_jar",
      #url = "https://github.com/junit-team/junit4/releases/download/r4.12/junit-4.12.jar",
      url = "https://github.com/orrsella/bazel-example/blob/master/third_party/junit/junit-4.12.jar",
      #sha256 = "59721f0805e223d84b90677887d9ff567dc534d7c502ca903c0c2b17f05c116a",
      #sha256 = "fe3d4c56388dc3d74049abae83f4520f6703062e174e16bb5551cdf439ca4f81",
  )

编译TensorFlow：

./buildTensorFlow.sh

打包成whl文件，放在$HOME目录下,如：tensorflow-1.0.1-cp27-cp27mu-linux_aarch64.whl

./packageTensorFlow.sh

安装whl文件

pip install $HOME/tensorflow-1.0.1-cp27-cp27mu-linux_aarch64.whl

提示权限错误的时候前边加上sudo即可

Caffee安装（安装成功）
参考：http://blog.csdn.net/jiongnima/article/details/70040262亲测可用

sudo gedit /etc/bash.bashrc

在文件的末尾加入如下

export PATH=/usr/local/cuda-8.0/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-8.0/lib64:$LD_LIBRARY_PATH

然后cuda就配置完毕了，运行

nvcc -V

这里写图片描述
首先安装一般依赖项

sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libhdf5-serial-dev protobuf-compiler
sudo apt-get install --no-install-recommends libboost-all-dev

再安装BLAS依赖项

sudo apt-get install libatlas-base-dev

然后是python和其余依赖项,安装python及其头文件

sudo apt-get install python
sudo apt-get install python-dev

安装python的其他依赖

sudo apt-get install python-numpy
sudo apt-get install ipython
sudo apt-get install ipython-notebook
sudo apt-get install python-sklearn
sudo apt-get install python-skimage
sudo apt-get install python-protobuf

然后是谷歌glog和gflags和lmdb依赖项

sudo apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev

接着安装git，并且下载代码

sudo apt-get install git
git clone https://github.com/BVLC/caffe.git

然后进入到源码目录

cd caffe

编译Caffe

cp Makefile.config.example Makefile.config

需要使用cuDNN的朋友可以将Makefile.config中USE_CUDNN := 1这一行之前的井号注释去掉，
同时，为了匹配cuda8.0的计算能力，请把Makefile.config中CUDA_ARCH中的前两行去掉，如下图所示(保留也行，编译的时候会弹出警告)
这里写图片描述
值得注意的是，笔者在第一次编译的时候遇到了如下问题：
./include/caffe/util/hdf5.hpp:6:18: fatal error: hdf5.h: No such file or directory
解决方法笔者参阅了http://blog.csdn.net/hongye000000/article/details/51043913点击打开链接，在链接博客中，解决方法如下：
在Makefile.config文件中，添加 /usr/include/hdf5/serial/ 到 INCLUDE_DIRS，也就是把下面第一行代码改为第二行代码。

INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include
INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include /usr/include/hdf5/serial/

在Makefile文件中，把 hdf5_hl 和hdf5修改为hdf5_serial_hl 和 hdf5_serial，也就是把下面第一行代码改为第二行代码。

LIBRARIES += glog gflags protobuf boost_system boost_filesystem m hdf5_hl hdf5
LIBRARIES += glog gflags protobuf boost_system boost_filesystem m hdf5_serial_hl hdf5_serial

上述解决方法笔者亲测有效，感谢博主，然后

make clean
make all -j8

即可编译成功。

处江湖之远，再见哦！[email protected]

【Jetson Tx2 视觉slam开发】入门配置二

猜你喜欢