ISCE2 and Mintpy Installation in Linux

此博客记录在Linux系统中重新安装ISCE2和Mintpy环境的流程。

1. 系统信息

Opreating System: Ubuntu 22.04
Miniconda: 25.5.1
Python: 3.1335

2. 安装ISCE2和Mintpy

ISCE is open source and on GitHub. There is also extensive documentation.

create a new directory named “miniconda3” in your home directory.
download the Linux Miniconda installation script for your chosen chip architecture and save the script as miniconda.sh in the miniconda3 directory.
run the miniconda.sh installation script in silent mode using bash.
remove the miniconda.sh installation script file after installation is complete.

mkdir -p ~/miniconda3
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O ~/miniconda3/miniconda.sh
bash ~/miniconda3/miniconda.sh -b -u -p ~/miniconda3
rm ~/miniconda3/miniconda.sh

After installing, close and reopen your terminal application or refresh it by running the following command:

1	`source ~/miniconda3/bin/activate`

Then, initialize conda on all available shells by running the following command:

1	`conda init --all`

创建单独的conda environment

1 2	`conda create -n isce2 python=3.13 # or 3.8 and above conda activate isce2`

安装必要的python包

1	`conda install -c conda-forge git scons cython gdal h5py libgdal pytest numpy fftw scipy pybind11 shapely opencv basemap`

to compile/install mdx, you will also need openmotif

1	`conda install -c conda-forge openmotif opencv`

检查是否安装成功：

# to check whether there is one  
ldconfig -p | grep libXm
# if not present, install as follows (need root privilege) 
# Ubuntu/Debian 
sudo apt install libxm4
# Redhat CentOS
dnf install motif, motif-devel

You will need to make a symbolic link for cython3, (this is important, otherwise, some packages will be skipped)

1 2	`cd $CONDA_PREFIX/bin ln -sf cython cython3`

Install conda version isce2 and mintpy

Download source codes

cd ~/tools
mkdir isce2; cd isce2
mkdir build install src; cd src
git clone https://github.com/isce-framework/isce2.git

cd ~/tools
git clone https://github.com/insarlab/MintPy.git
git clone https://github.com/insarlab/PyAPS.git
git clone https://github.com/insarlab/PySolid.git
git clone https://github.com/yunjunz/conda-envs.git

mkdir -p ~/tools/utils; cd ~/tools/utils
git clone https://gitlab.com/yunjunz/SSARA.git
git clone https://github.com/yunjunz/sardem.git

新建一个~/tools/isce_requirments.txt文件

python=3.8.6
cython
gdal
git
h5py
libgdal
pytest
numpy
fftw
scipy
basemap
scons
opencv

Install to insar environment

1
2
3

# create new environment
conda create --name isce2 --file isce_requirements.txt
conda activate insar

使用 `conda` 或 `mamba` 安装 `mintpy`

1	`mamba install conda-forge::mintpy pysolid pyaps3 sardem`

以下安装在国内失败，弃用

# [for developers] overwrite PyAPS and PySolid installation from conda to the editable mode
python -m pip install -e PyAPS
export SETUPTOOLS_ENABLE_FEATURES="legacy-editable"
python -m pip install -e PySolid

# install dependencies not available from conda
ln -s ${CONDA_PREFIX}/bin/cython ${CONDA_PREFIX}/bin/cython3
python -m pip install ipynb     # import functions from *.ipynb files
python -m pip install jupyter_nbextensions_configurator
python -m pip install -e utils/sardem

Setup

Create an alias load_insar in ~/.bash_profile file for easy activation, e.g.:

1	`alias load_insar='conda activate insar; source ~/tools/conda-envs/insar/config.rc'`

在~/tools/conda-envs/insar/config.rc文件中添加如下内容:

##---------------------- Work Dirs ----------------##
export WORK_DIR=~/Documents/00_my_projects/01_isce_test
export ORBIT_PATH=${WORK_DIR}/orbits
export AUX_PATH=${WORK_DIR}/AUXFILE
export DEM_PATH=${WORK_DIR}/DEM
export SLC_PATH=${WORK_DIR}/SLC
export INSAR_WORK_DIR=${WORK_DIR}/stacks

在~/.netrc文件中添加如下内容:

1	`machine urs.earthdata.nasa.gov login mengyuchi password 5212109Mengle`

Test the installation

Run the following to test the installation:

1
2
3

load_insar               # warm up conda environment
topsApp.py -h            # test ISCE-2
smallbaselineApp.py -h   # test MintPy

下载测试数据并测试ISCE与Mingpy
- 下载测试数据：
1
2
3
load_insar cd ~/Documents/00_my_projects/01_isce_project/01_isce_test
剩余测试操作步骤详见

失败Install ISCE2: (此部分在我的linux环境中失败，放弃isce开发者环境的安装)

Download isce2 for GitHub or prepare your own version

# create a directory to save source files
mkdir -p ${HOME}/tools/src 
cd $HOME/tools/src
# glone a copy from github
git clone https://github.com/isce-framework/isce2

此命令将下载ISCE2的github版本到指定目标文件夹/home/tools/src/isce2

在文件夹$HOME/.isce下配置SConfigISCE

PRJ_SCONS_BUILD=$HOME/build/isce_build
PRJ_SCONS_INSTALL=$ISCE_HOME
LIBPATH=$CONDA_PREFIX/lib
CPPPATH=$CONDA_PREFIX/include $CONDA_PREFIX/include/python3.13/ $CONDA_PREFIX/lib/python3.13/site-packages/numpy/_core/include $CONDA_PREFIX/include/opencv4
FORTRAN=gfortran
CC=gcc
CXX=g++
FORTRANPATH=$CONDA_PREFIX/include
# note: if you use system-installed openmotif, change the settings below accordingly
MOTIFLIBPATH=$CONDA_PREFIX/lib
X11LIBPATH=$CONDA_PREFIX/lib
MOTIFINCPATH=$CONDA_PREFIX/include
X11INCPATH=$CONDA_PREFIX/include
RPATH=$CONDA_PREFIX/lib
ENABLE_CUDA = True
CUDA_TOOLKIT_PATH=/usr/local/cuda  # use 'which nvcc' to verify

SConfigISCE文件中参数的解释:

PRJ_SCONS_BUILD is a directory to save temporary compiled files

PRJ_SCONS_INSTALL is where the isce2 will be installed. We use a $ISCE_HOME to be defined later

LIBPATH is where to look for the shared libraries, such as gdal, fftw

CPPPATH is where to look for C/C++ head files (#include) for libraries. Use the following commands to find the proper path(s) if you use a different python version or system

# Python.h location
python3 -c "import sysconfig; print(sysconfig.get_path('include'))"
# numpy include 
python3 -c "import numpy; print(numpy.get_include())"

FORTRAN, CC, CXX are the Fortran/C/C++ compilers to be used

FORTRANPATH, MOTIFLIBPATH … X11INCPATH are to set lib and include paths for motif and x11 libraries.

for Ubuntu 18.04, set MOTIFLIBPATH1 and X11LIBPATH to /usr/lib/x86_64-linux-gnu/ and set MOTIFINCPATH and X11INCPATH to /usr/include.

for conda installed packages, set them as $CONDA_PREFIX/lib and $CONDA_PREFIX/include.

ENABLE_CUDA = True/False whether to include some GPU/CUDA accelerated modules. If enabled, please also specify CUDA_TOOLKIT_PATH to where CUDA SDK is installed. Some manual configuration might be needed.

Some settings for environment variables before compile/install. You need to specify three environment variables:

CONDA_PREFIX where Anaconda3 is installed
ISCE_HOME where isce2 will be installed
SCONS_CONFIG_DIR where the SConfigISCE is located. if they are not set. Check by, e.g., echo $CONDA_PREFIX.

for bash

1 2	`export ISCE_HOME=$HOME/tools/isce export SCONS_CONFIG_DIR=$HOME/.isce`

Compile/install isce2

1 2	`cd $HOME/tools/src/isce2 scons install`

If successful, you should obtain a compiled isce2 at $ISCE_HOME or $HOME/tools/isce.

安装中部分错误信息

sh: 1: gfortran: not found: 原因应该是gfortran包未正确安装

1	`sudo apt-get install gfortran`

error: Xm/Xm.h: No such file or directory

1	`sudo apt-get install libmotif-dev`

Error: Type mismatch between actual argument at (1) and actual argument at (2) (COMPLEX(4)/REAL(4)).

是一个非常经典的 Fortran 编译器严格性问题，尤其在使用新版 gfortran 编译 ISCE2 等较老的 Fortran 代码时。在$HOME/.isce文件中，使用gedit SConfigISCE添加下面一行：

1	`FORTRANFLAGS = '-fallow-argument-mismatch'`

成功安装 ISCE2

![成功安装ISCE](ISCE successful installation.png)

配置ISCE2环境路径

可以在目录/home/yuchi/Documents/00_my_projects/01_isce_project下手动新建isce配置文件isce.rc并在此目录下运行source isce.rc.

# isce2.rc
# 1. 设置 ISCE_HOME 环境变量，指向 ISCE2 的安装根目录（通常是 Conda 环境的根目录）
export ISCE_HOME=$HOME/tools/isce

# 2. 将 ISCE2 的应用程序目录添加到 PATH
# 这样您就可以直接运行 topsApp.py 等命令
export PATH=$ISCE_HOME/bin:$ISCE_HOME/applications:$PATH


export LD_LIBRARY_PATH=$ISCE_HOME/lib:$LD_LIBRARY_PATH

# 3. 将 ISCE2 Python 模块的父目录添加到 PYTHONPATH
# 这样 Python 解释器可以找到并导入 'isce' 模块
export PYTHONPATH=$ISCE_HOME:$ISCE_HOME/applications:$ISCE_HOME/components:$ISCE_HOME/library:$HOME/tools:$PYTHONPATH

也可以直接在系统配置文~/.bashrc中加入以上代码并运行source ~/.bashrc即可。

安装验证：

在terminal中运行：

python3 -c "import isce"
topsApp.py --help --steps
# if you compiled mdx
mdx # or mdx.py

ISCE2 and Mintpy Installation in Linux

1. 系统信息

2. 安装ISCE2和Mintpy

使用 conda 或 mamba 安装 mintpy

使用 `conda` 或 `mamba` 安装 `mintpy`