This repository relates to the publication with title:
"A mobile soil analysis system for sustainable agriculture" Ademir Ferreira da Silva el al.
Available at: TBD
and the corresponding dataset repository archived at:
https://archive.materialscloud.org/record/2022.91
This repository comprises the following Jupyter Notebooks for the analysis and model training with colorimetric data extracted from chemical reactions on paper-based sensing devices.
In sequencial order of application:
This notebook extract the colorimetric information from the images captured of paper devices, and saving the RGB values in a csv file. This notebook uses the images in the repository: https://doi.org/10.24435/materialscloud:vt-4t
This notebook loads the csv file with RGB data per paper device, after adding a column with the 'Class' of that data based on the pH value of the sample applied to the paper device, and trains two openCV Logistic Regression models that are saved into XML files for application with a mobile device.
This notebook uses the csv files with RGB data collected for two pH indicators and available at https://doi.org/10.24435/materialscloud:vt-4t
This notebook shows how each image of a paper device captured outdoors with the mobile device is processed to compensate for illumination differences with the calibrationd dataset collected under laboratory illumination conditions. This notebook also goes through the subsequent steps of importing and then applying the trained logistic regression models to the newly captured and treated color data. Reference logistic regression models can be found under the folder '\ML_models' and illumination references used by the notebook can be found under folder '\Illumination_references'
Library of functions used by the above notebooks.
If you would like to see the detailed LICENSE click here.
#
# Copyright 2020- IBM Inc. All rights reserved
# SPDX-License-Identifier: BSD-3-Clause
#
- Matheus Esteves Ferreira
- Jaione Tirapu Azpiroz