Skip to content

Latest commit

 

History

History
63 lines (46 loc) · 2.74 KB

README.md

File metadata and controls

63 lines (46 loc) · 2.74 KB
Raw

Plant-leaf-infection-detection

This is the source code for our work submitted for review at Inderscience journal (IJSHC). The respository contains the code as well as the data used for the simulations.

The dataset is taken from PlanVillage dataset from Sharada Mohanty https://github.com/spMohanty/PlantVillage-Dataset

The different versions of the dataset are present in the raw directory :
color : Original RGB images
grayscale : grayscaled version of the raw images
segmented : RGB images with just the leaf segmented and color corrected.

Execution steps:

Training Process

1. Place 'data_extraction.py' in Train folder of the selected leaf folder. e.g..
cp data_extraction.py Bell\ Pepper\ Data\ Set/Train_pep_bac/.
2. Create a text file in leaf folder( eg. Bell Pepper Data Set) with naming 'DiseaseType_result.txt'. e.g..
touch Bell\ Pepper\ Data\ Set/bacterial_result.txt
3. Go to Train folder. eg..
cd Bell\ Pepper\ Data\ Set/Train_pep_bac/
4. Update filename on line 162, with file name created in step 2.
5. Execute the data_extraction file
python data_extraction.py
6. Repeat for each Train folder in selected leaf folder

At the end of the text file such as 'bacterial_result.txt', the average infection percentage for that particular disease for the selected leaf type will be mentioned.

After creating training text file for each training folder.

Test Process

1. Place 'data_marking.py' and 'leaf_classification.py' in Test folder of the selected leaf folder. e.g..
cp data_marking.py Bell\ Pepper\ Data\ Set/Test_pep_bac/.
cp leaf_classification.py Bell\ Pepper\ Data\ Set/Test_pep_bac/.

2. Create a text file in leaf folder( eg. Bell Pepper Data Set/Test_pep_healthy/) with naming 'DiseaseType__test_result.txt'. e.g..
touch Bell\ Pepper\ Data\ Set/Test_pep_healthy/healthy_test_result.txt

3. Go to Train folder. eg..
cd Bell\ Pepper\ Data\ Set/Test_pep_healthy/

4. Update filename on line 134 of data_marking, with file name created in step 2.

5. Change leaf marking as 'Healthy' or 'Infected' at line 151 and 164 of data_marking.

6. Execute the data_marking file
python data_marking.py

7. Execute the leaf_classification file
python leaf_classification.py

8. Repeat for each test folder

At the end of the text file such as 'healthy_test_result.txt', the classification accuracy for that particular disease for the selected leaf type will be mentioned.

The link to the original paper : https://dx.doi.org/10.1504/IJSHC.2019.101602

If our work helps you, please cite it as:

N. Paliwal, P. Vanjani, J.-W. Liu, S. Saini, and A. Sharma. ”Image processing-based intelligent robotic system for assistance of agricultural crops.” International Journal of Social and Humanistic Computing,3(2):191–204, 2019.