This repository has the main propuse to analyse Passive Sonar Signals with Machine Learning Algorithms and other staff
1 - First clone the package lastest version
$ git clone https://github.com/natmourajr/SonarAnalysis.git
2 - Go to package folder
$ cd SonarAnalysis
3 - This package works with a couple of Environment Variables, to change their values, do
$ vim setup.sh
Edit $SONAR_WORKSPACE = analysis path, where the magic happens
4 - After change Environment variables, do
$ source setup.sh
5 - After change Environment variables, do
$ cd Packages
6 - In Signal Processing Lab machines, run matlab with ReadRawData command to generate RawData Files
$ matlab -nodesktop -r ReadRawData
7 - In Signal Processing Lab machines, run matlab with LofarAnalysis command to generate LofarData Files
$ matlab -nodesktop -r LofarAnalysis
8 - In Signal Processing Lab machines, run matlab with LofarAnalysis command to generate LofarData Files
$ python ReadLofarData.py
9 - After run these two Matlab scripts, the two new matlab files should be in $OUTPUTDATAPATH, to check it
$ ls $OUTPUTDATAPATH
10 - Some libraries are necessary, to install them do
$ cd $SONAR_WORKSPACE
11 - In this version, all analysis will be produce in Python. I suggest create a virtualenv and upgrade pip (to update pip)
$ pip install --upgrade pip
12 - In this version, all analysis will be produce in Python. I suggest create a virtualenv and install all libraries listed in requirements.txt
$ pip install -r requirements.txt
Obs: To update all packages
$ pip freeze --local | grep -v '^\-e' | cut -d = -f 1 | xargs pip install -U
13 - Now you can go to Packages folders and access all analysis
$ cd Packages
14 - To access a specific analysis, do
$ cd <Specific Analysis Name>
15 - Run a specific setup.sh
$ source setup.sh
16 - And
$ jupyter notebook