Image Processing Pipeline for Space Robotic Applications
- This repo is intended to be used from the main directory for this repo ie.
cd ~./ippsrathen run all commandspython src/ippsra/rank_images.pyfor example. - Below is the output from the help argument for the ranking_images.py
Help Output for rank_images.py
usage: rank_images.py [-h] [--Directory DIRECTORY]
[--Extension {bmp,dib,jpeg,jpg,png,webp,pbm,pgm,ppm,pxm,pnm,sr,
ras,tiff,tif,exr,hdr,pic}]
[--csvPath CSVPATH] [--csvName CSVNAME] [--fileName FILENAME]
[--save {True,False}] [--imgIdx IMGIDX]
[--showImages {True,False}] [--showPlot {True,False}]
Code for ranking a directory of images on a custom scale that is defined at this
following link https://github.com/collin-love/ippsra/blob/main/README.md
options:
-h, --help show this help message and exit
--Directory DIRECTORY, -D DIRECTORY
The PATH to the directory containing the images that are going
to be processed (default: ./data/test_data)
--Extension {bmp,dib,jpeg,jpg,png,webp,pbm,pgm,ppm,pxm,pnm,sr,ras,tiff,tif,exr,hdr,pic},
-ext {bmp,dib,jpeg,jpg,png,webp,pbm,pgm,ppm,pxm,pnm,sr,ras,tiff,tif,exr,hdr,pic}
A list of the extensions for the image data that is present in
the directory for processing (default: png)
--csvPath CSVPATH The PATH to the directory where the CSV file will be saved
(default: ./data/processed_data)
--csvName CSVNAME Then name of the CSV file that will be created (default:
sorted_test.csv)
--fileName FILENAME The test name for the file. ex: if --fileName test1_ then this
would result in test1_1_scatter.png (default:
./data/processed_data/Test_)
--save {True,False} Boolean for if you want to save the files (default: False)
--imgIdx IMGIDX The index of the image that you want to show at the end of the
script from a sorted list of images from best to worst. The
default is 0 (the best ranked image) (default: 0)
--showImages {True,False}
Boolean for if you want to show images(default: True)
--showPlots {True,False}
Boolean for if you want to plot the data (default: True)
- Clone this repository and cd into it
git clone https://github.com/collin-love/ippsra.gitcd ippsra
- ENV setup for this project below
conda env create -f environment.yml conda activate ippsra - For an update to this already created environment from the environment
file -- run
conda activate ippsra conda env update --file environment.yml --prune
- The images that are used to test are in the data/test directory. This subset of data is a good representation of many cases of normal and outlier images
- Once everything is set up in the repository, the following line can be ran from the home directory for the project.
- The default command
python src/ippsra/rank_images.pyis represented below
python src/ippsra/rank_images.py --Directory ./data/test_data --Extension png --csvPath ./data/processed_data --csvName sorted_test.csv --fileName ./data/processed_data/Test_--save False --imgIdx 0 --showPlots True --showImages True
- The output in the terminal will look like the following and some images will be displayed on the screen
The default Terminal Output
There are 41 images in this dir
0 images have been ranked
20 images have been ranked
40 images have been ranked
There are 26 images with a hazard score of 1
There are 7 images with a hazard score of 2
There are 8 images with a hazard score of 3
- Bellow are the sample images and graphs that can be created out--of--the--box
- Scatter plot for the data
- Violin plot for the data
- Both plots side by side
- This first lunar image is the best image from the ranking
- This second lunar image is the best image from the ranking
- User defined image showing
- A compilation of the top images from each category
- These parameters can be varied
--save {True,False} Boolean for if you want to save the files (default: False)
--imgIdx IMGIDX The index of the image that you want to show at the end
of the script from a sorted list of images from best to
worst. The default is 1 (the best ranked image) (default: 1)
--showImages {True,False}
Boolean for if you want to show images(default: True)
--showPlots {True,False}
Boolean for if you want to plot the data (default: True)
- By specifying True when running
python src/ippsra/rank_images.py --save Truea CSV file with the name --csvName of the rankings will be provided to the --csvPath directory - For the defaults this command will create and populate two CSV files in this "./data/processed_data/" directory called "sorted_test.csv" and "raw_data.csv"
- Also, specifying True or False for saving, plotting, and showing images will
either do the specified thing or not. Ex.
python src/ippsra/rank_images.py --showPlots True --showImages Truewill show all the plots and images that are created - For the default arguments this will save all outputs to the "./data/processed_data/"
directory - To check this CSV the output will look like below
The default Terminal Output
- For sorted_test.csv
Image Name,Number of Hazards,Density of Hazards,Hazard Score
test_01.png,6,0.0,1
test_023.png,11,0.0,1
test_034.png,6,0.0,1
test_039.png,3,0.0,1
test_02.png,7,0.0,1
test_027.png,3,0.0,1
test_032.png,2,0.0,1
test_019.png,14,0.0,1
test_041.png,11,0.0,1
test_03.png,10,0.0,1
test_018.png,25,0.01,1
test_05.png,34,0.01,1
test_033.png,24,0.01,1
test_024.png,24,0.01,1
test_040.png,85,0.02,1
test_031.png,28,0.02,1
test_06.png,65,0.02,1
test_07.png,56,0.02,1
test_025.png,9,0.02,1
test_016.png,70,0.02,1
test_030.png,46,0.03,1
test_022.png,35,0.03,1
test_026.png,44,0.03,1
test_04.png,22,0.03,1
test_011.png,121,0.03,1
test_038.png,29,0.04,1
test_020.png,39,0.05,1
test_036.png,32,0.05,1
test_037.png,91,0.07,2
test_029.png,95,0.07,2
test_08.png,1717,0.07,2
test_042.png,127,0.08,2
test_013.png,311,0.08,2
test_021.png,93,0.09,2
test_010.png,343,0.1,2
test_017.png,207,0.14,3
test_043.png,129,0.14,3
test_035.png,223,0.18,3
test_015.png,793,0.19,3
test_012.png,872,0.23,3
test_014.png,976,0.24,3
test_028.png,333,0.34,3
test_09.png,1312,0.41,3
- For raw_data.csv
Image Name,Number of Hazards,Density of Hazards,Hazard Score
test_01.png,6,0.0,1
test_010.png,343,0.1,2
test_011.png,121,0.03,1
test_012.png,872,0.23,3
test_013.png,311,0.08,2
test_014.png,976,0.24,3
test_015.png,793,0.19,3
test_016.png,70,0.02,1
test_017.png,207,0.14,3
test_018.png,25,0.01,1
test_019.png,14,0.0,1
test_02.png,7,0.0,1
test_020.png,39,0.05,1
test_021.png,93,0.09,2
test_022.png,35,0.03,1
test_023.png,11,0.0,1
test_024.png,24,0.01,1
test_025.png,9,0.02,1
test_026.png,44,0.03,1
test_027.png,3,0.0,1
test_028.png,333,0.34,3
test_029.png,95,0.07,2
test_03.png,10,0.0,1
test_030.png,46,0.03,1
test_031.png,28,0.02,1
test_032.png,2,0.0,1
test_033.png,24,0.01,1
test_034.png,6,0.0,1
test_035.png,223,0.18,3
test_036.png,32,0.05,1
test_037.png,91,0.07,2
test_038.png,29,0.04,1
test_039.png,3,0.0,1
test_04.png,22,0.03,1
test_040.png,85,0.02,1
test_041.png,11,0.0,1
test_042.png,127,0.08,2
test_043.png,129,0.14,3
test_05.png,34,0.01,1
test_06.png,65,0.02,1
test_07.png,56,0.02,1
test_08.png,1717,0.07,2
test_09.png,1312,0.41,3
- v1.0
- This version contains functions and scripts that can be used to determine the rough estimates of the density of obstructions in photos. In our case we are using Lunar surface images and renders to test and evaluate the performance of the project.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent