—————————20200509补充————————

1.import gdal报错问题

topsApp.py

报错:找不到libpoppler.so.71
解决:

sudo ln -s libpoppler.so libpoppler.so.71

继续运行:

topsApp.py

报错:

ImportError: /home/lll/anaconda3/lib/python3.6/site-packages/osgeo/../../../libgdal.so.20: undefined symbol: _ZN9OutputDev18beginMarkedContentEPcP4Dict

解决:
后续发现,此错误与ISCE无关,在import gdal会报同样的错误,是gdal包的问题。

  1. 新建虚拟环境(不必须,了解到后,方便不同软件的切换,参考链接)

    #创建虚拟环境
    conda create -n isce_conda_env python=3.6
    #激活虚拟环境
    source activate isce_conda_env
    #退出虚拟环境
    #source deactivate isce_conda_env
    
  2. 重新配置python环境,参考正文4.1中依次install
    2.1. opencv3

    #查看版本
    conda search -c menpo opencv3
    #安装所需版本
    conda install -c menpo opencv3=3.1.0=py36_0
    #查看是否成功(gdal的前车之鉴)
    python3
    >>>import cv2
    

    2.2. basemap
    Anaconda查看、删除、增加channel

    #清华镜像加入channel,速度有提升
    conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/
    #basemap是matplotlib的子包,直接安装一直solving environment没有反应,这样安装就可以了
    conda install matplotlib
    conda install basemap
    #查看是否成功(gdal的前车之鉴)
    python3
    >>>from mpl_toolkits.basemap import Basemap
    

    2.3. gdal

    conda install gdal
    python3
    >>>import gdal
    

    不再报错就是成功了,此处安装的gdal是3.0.4,之前安装的是2.2.2,不确定是否是版本的问题,不确定是否跟opencv,basemap,gdal的安装先后顺序有关系。
    【注】修改cuda配置文件,参考正文4.2。

    2.4. 其它包
    【注】cython的软链接

2. mdx:not found 报错问题

处理数据完成后,可视化

mdx.py filt_topophase.flat.geo

报错: sh:1:mdx: not found
解决:

locate mdx
返回:/home/lll/isce2/install/isce/bin/mdx
#编辑路径
gedit ~/.bashrc
添加: export PATH=$ISCE_HOME/bin:$PATH

—————————以下是原文——————————

1.安装双系统

win10/Ubuntu双系统安装参考链接
ubuntu镜像下载链接

2.安装Anaconda

安装Anaconda教程参考链接
清华大学Anaconda镜像链接
官网历史版本Anaconda镜像链接
python和Anaconda的版本对应关系链接

isce2推荐python3.6,下载对应anaconda版本并安装
在这里插入图片描述
在这里插入图片描述
在这里插入图片描述
在这里插入图片描述
是否添加环境变量,yes👇
在这里插入图片描述
打开一个新终端,检查是否安装成功👇
在这里插入图片描述
若不成功则手动添加环境变量👇

sudo gedit ~/.bashrc
txt末尾添加    export PATH="/home/lll/anaconda3/bin:$PATH"
source ~/.bashrc

3.安装cuda和cudnn(启用gpu所需)

ubuntu18.04安装cuda9.0,cudnn7.6和tensorflow1.9 参考链接

4.ISCE环境配置

ISCE github地址
ISCE InSAR处理软件环境配置 参考链接

4.1. 配置Anaconda中的python环境

查看当前conda的包环境👇,生成的txt中有当前环境中已有的包及其版本

conda list -e > requirements.txt

官网的ISCE python环境要求👇
在这里插入图片描述

【方法1-亲测不能用】按照以下方法一起安装所需要的包时
解决环境冲突会选择安装py27的包,而且conda无法再调用。

将 requirements.txt 修改为上述内容

conda install --yes --file requirements.txt

【方法2】根据requirements.txt 筛选,将缺少的包逐一安装。

conda install gdal
			  git
			  fftw
			  basemap
			  scons
              opencv

Ensure that you create a link in the anaconda bin directory for cython3.
安装完成后在conda的bin目录下对cython3建立软链接.

sudo ln -s cython cython3

4.2. 配置cuda的nvcc.profile文件

NVCC CUDA编译流程 参考链接
找到nvcc.profile并编辑

lll@lll-Lenovo-Legion-Y7000:~$ locate nvcc.profile
/etc/nvcc.profile
/usr/lib/nvidia-cuda-toolkit/bin/nvcc.profile
/usr/local/cuda-9.2/bin/nvcc.profile
lll@lll-Lenovo-Legion-Y7000:~$ sudo gedit /usr/local/cuda-9.2/bin/nvcc.profile

未编辑前👇


TOP              = $(_HERE_)/..

NVVMIR_LIBRARY_DIR = $(TOP)/nvvm/libdevice

LD_LIBRARY_PATH += $(TOP)/lib:

PATH            += $(TOP)/nvvm/bin:$(_HERE_):

INCLUDES        +=  "-I$(TOP)/$(_TARGET_DIR_)/include" $(_SPACE_)

LIBRARIES        =+ $(_SPACE_) "-L$(TOP)/$(_TARGET_DIR_)/lib$(_TARGET_SIZE_)/stubs" "-L$(TOP)/$(_TARGET_DIR_)/lib$(_TARGET_SIZE_)"

CUDAFE_FLAGS    +=
PTXAS_FLAGS     +=

将conda的include和lib目录添加到nvcc.profile中
否则编译中会报错:
在这里插入图片描述编辑后👇


TOP              = $(_HERE_)/..
HOME             = $(_HERE_)/../../../..

NVVMIR_LIBRARY_DIR = $(TOP)/nvvm/libdevice

LD_LIBRARY_PATH += $(HOME)/home/lll/anaconda3/lib:$(TOP)/lib64:

PATH            += $(TOP)/nvvm/bin:$(_HERE_):$(HOME)/home/lll/anaconda3/bin:$(HOME)/usr/local/sbin:$(HOME)/usr/local/bin:$(HOME)/usr/sbin:$(HOME)/usr/bin:/sbin:/bin:$(HOME)/usr/games:$(HOME)/usr/local/games:$(HOME)/snap/bin

INCLUDES        +=  "-I$(HOME)/home/lll/anaconda3/include" "-I$(TOP)/$(_TARGET_DIR_)/include"$(_SPACE_)

LIBRARIES        =+ $(_SPACE_) "-L$(TOP)/$(_TARGET_DIR_)/lib$(_TARGET_SIZE_)/stubs" "-L$(TOP)/$(_TARGET_DIR_)/lib$(_TARGET_SIZE_)"

CUDAFE_FLAGS    +=
PTXAS_FLAGS     +=

4.3.安装编译依赖环境

逐个安装,避免出错(ISCE InSAR处理软件环境配置 参考链接)

apt install -y 
		gfortran 
		libmotif-dev 
		libhdf5-dev 
		libfftw3-dev 
		libgdal-dev 
		scons 
		python3 
		cython3 
		python3-scipy 
		python3-matplotlib 
		python3-h5py 
		python3-gdal 
		python3-pip 
		wget 
		curl 
		gdal-bin 
		libx11-dev
		libxt-dev

设置SCONS_CONFIG_DIR=下载的ISCE安装包路径

sudo gedit ~/.bashrc
在文档末尾添加  export SCONS_CONFIG_DIR=/home/lll/isce2
source gedit ~/.bashrc

官方说明👇
在这里插入图片描述在isce2/docker中找到SConfigISCE,复制到SConstruct同目录下txt,按照👆提示编辑内容
编辑结果👇 注意build目录和install目录都以isce结尾,否则import包和调用时会出现问题

# The directory in which ISCE will be built
PRJ_SCONS_BUILD = /home/lll/isce2/build/isce

# The directory into which ISCE will be installed
PRJ_SCONS_INSTALL = /home/lll/isce2/install/isce

# The location of libraries, such as libstdc++, libfftw3 (for most system
# it's /usr/lib and/or /usr/local/lib/ and/or /opt/local/lib) /usr/lib/x86_64-linux-gnu
LIBPATH = /usr/lib /home/lll/anaconda3/lib 

# The location of Python.h. If you have multiple installations of python
# make sure that it points to the right one /usr/include/hdf5/serial
CPPPATH = /home/lll/anaconda3/include/python3.6m /home/lll/anaconda3/lib/python3.6/site-packages/numpy/core/include /usr/include/gdal /home/lll/anaconda3/include

# The location of the fftw3.h (most likely something like /usr/include or
# /usr/local/include /opt/local/include
FORTRANPATH =  /usr/include 

# The location of your Fortran compiler. If not specified it will use the system one
FORTRAN = /usr/bin/gfortran

# The location of your C compiler. If not specified it will use the system one
CC = /usr/bin/gcc
#在这里找到X11,Xm,omp,fftw3 .h
# The location of your C++ compiler. If not specified it will use the system one
CXX = /usr/bin/g++


#libraries needed for mdx display utility
MOTIFLIBPATH = /usr/lib         # path to libXm.dylib
X11LIBPATH = /usr/lib           # path to libXt.dylib
MOTIFINCPATH = /usr/include     # path to location of the Xm
                                # directory with various include files (.h)
X11INCPATH = /usr/include       # path to location of the X11 directory
                                # with various include files
# list of paths to search for shared libraries when running programs
#RPATH =/usr/lib /home/lll/anaconda3/lib 

#Explicitly enable cuda if needed
ENABLE_CUDA = True
CUDA_TOOLKIT_PATH = /usr/local/cuda  #/usr/local/cuda
cd isce2
scons

如果编译报错,应该是配置文件路径问题,根据报错内容,利用以下命令进行查找包的位置和修改路径

locate 
find
whereis

3.4 添加环境变量
在这里插入图片描述

如果不小心改错了PATH,导致terminal无法使用命令

export PATH=/bin:/usr/bin
sudo gedit ~/.bashrc

就可以打开文件修改回来了

3.5 测试是否成功
在这里插入图片描述

Logo

助力广东及东莞地区开发者,代码托管、在线学习与竞赛、技术交流与分享、资源共享、职业发展,成为松山湖开发者首选的工作与学习平台

更多推荐