WaterPollutionContermeasure

/ Overview

/ Aim of Development of This Repository

• , , (); \ Explore quantitative relationship between information derived from optical and infrared RS (e.g. band reflectance) and components related to water quality via combination of RS data and in-situ monitoring data.
  • (); \ Initial target: reproduce inversion model documented in other published academic works.
  • ; \ Further target: Construct novel models conbined with machine learning.

Python / Third-Party Python Libraries Mentioned in This Repository

• , . \ Libraries for mathematical modeling, data mining and general scientific data computation.
  • numpy
  • scipy
• , . \ Libraries for data mining and machine learning.
  • pandas
  • scikit-learn(sklearn)
• IDL / ENVIPython 2.x (Python to IDL Bridge). \ Python 2.x interface of IDL / ENVI (Python to IDL Bridge).
  • idlpy

/ Information of Environment for Execution

/ Environment of System and Packages

|
Item|
Value| |:-:|:-| |
Architecture of processor|Intel x86-64| |
OS kernel version|Windows NT 6.1.7601| |Conda
Conda package manager version|conda 4.7.12 (python 3.7.4, requests 2.22.0)| |Anaconda
Anaconda system release version|anaconda 2019.10| |ArcGIS
ArcGIS release version|ArcGIS Desktop 10.5| |ENVI
ENVI release version|ENVI 5.3|

idlpy, sklearn / Configure Integrated Environment of idlpy and sklearn

anaconda, IDL_Machine_Intelligence_x64, idlpysklearns, : \ This repository creates an environment named IDL_Machine_Intelligence_x64 via anaconda environment manager, which intergrates idlpy and sklearn for data processing and machine learning, criteria are as follow:

• . \ Waters covered by extent of RS images must contain in-situ monitoring data.
  • ; \ Construct programming environment for further creation and solution of inversion model based on machine learning;
  • python, ; \ Programming environment is based on python for full usage of package ecosystem of scientific calculation and machine learning;
  • ENVI/IDL, , . \ Programming environment docks to ENVI/IDL, allowing workflow of effecient input, processing and output of RS data.

[!IMPORTANT]

, , anaconda.

Before execution of following operations, developers are supposed to disable anti-virus softwares and third-party firewalls, which may block anaconda from downloading packages from mirror websites.

IDL / Verify Installation Path of IDL

Anaconda prompt, IDL. \ In Anaconda prompt, set a temporary variable to store the installation path of IDL interpreter.

bash set idl=D:\Exelis\IDL85

[!NOTE] IDL(idl.exeidlrt.exe).

bin, examples, external, help, lib, resource.

Path above is NOT the path where the interpreters of IDL (idl.exe or idlrt.exe) locate.

This path contains subdirectories including bin, examples, external, help, lib, resource.

[!IMPORTANT] ****, .

, , . ****.

Values of temporary variables are valid in CURRENT CLI SESSION ONLY and will be cleared or restored after closing the console.

After excuting command above, do NOT close CLI, so that variables can be reused in following commands. Further operations will be executed in THE SAME CONSOLE.

anaconda / Verify Installation Path of anaconda

Anaconda prompt, anaconda. \ In Anaconda prompt, set a temporary variable to store the installation path of anaconda.

bash set anacon=D:\Anaconda3

/ Create and Initiallize Environment

1. anaconda. \ Initialize parameters of anaconda for creation of new environment.

bash conda clean -i

1. arcpy_idlpy_x32. \ Create and enter new environment arcpy_idlpy_x32.

bash conda create -n IDL_Machine_Intelligence_x64 python==2.7.12 pip==8.1.1 wheel==0.29.0 six==1.10.0 setuptools==27.2.0 cycler==0.10.0 pyparsing==2.1.4 -y conda activate IDL_Machine_Intelligence_x64 3. . \ Configure packages related to data processing and machine learning.

bash conda install numpy==1.11.1 scipy==0.18.1 matplotlib==1.5.3 python-dateutil==2.5.3 pytz==2016.6.1 pandas==0.18.1 -c conda-forge -y conda install scikit-learn==0.17.1 scikit-image==0.12.3 networkx==1.11 pillow==3.3.1 -c conda-forge -y

1. ipython. \ Configure packages related to ipython in new environment.

bash conda install notebook==4.2.3 ipykernel==4.5.0 ipython_genutils==0.1.0 jinja2==2.8 jupyter_client==4.4.0 jupyter_core==4.2.0 nbconvert==4.2.0 nbformat==4.1.0 pyzmq==15.4.0 tornado==4.4.1 -c conda-forge -y

1. idlpy. \ Create path configuration file of idlpy in new environment.

bash chdir /D %anacon%\envs\IDL_Machine_Intelligence_x64\Lib\site-packages echo %idl%\bin\bin.x86-64 >IDL8.5.pth echo %idl%\lib\bridges >>IDL8.5.pth

1. ipython, baseJupyter notebook. \ Register new environment as an ipython kernel for use with Jupyter notebook in base environment.

bash pip install backports.functools_lru_cache ipython kernelspec install-self

!NOTE, pip, , conda.

The faliure of kernel registration is generally caused by faliure of installation of library above via pip, In case of phenomenon above, it is required to install this library via conda again.

bash conda install backports.functools_lru_cache -y

1. . \ Finish configuration of new environment.

bash conda deactivate