Skip to main content

A Lightweight Deep Learning Model for Crop Disease Detection on Mobile Devices Cover

.blurhash-client-img { display: none !important; }

A Lightweight Deep Learning Model for Crop Disease Detection on Mobile Devices

International Journal on Smart Sensing and Intelligent Systems

Volume 19 (2026): Issue 1 (January 2026)

By: Qi Jing

Open Access

|Jul 2026

Figures & tables

Figures & Tables

Workflow of the proposed crop disease detection system.

Decision tree.

CNN model accuracy and loss. CNN, Convolutional Neural Network.

Crop Disease Detection Using Simulated Data

Step 1: Generate synthetic values for soil health, weather conditions, and plant health
Step 2: Assign disease label using the rule plant health <50
Step 3: Split the dataset into training set (80%) and testing set (20%)
Step 4: Train Random Forest, SVM, and CNN models on training data
Step 5: Evaluate models using accuracy, F1-score, and inference time
Step 6: Compare the performance of the models

Comparative study of crop disease detection methods

Study	Model used	Dataset type	Accuracy (%)	Mobile deployment suitability
[35]	Deep CNN	Image-based	96.8	Low
[36]	CNN + Mobile App	Image-based	94.5	Medium
[37]	Enhanced CNN	Image-based	97.2	Medium
Proposed method	Lightweight CNN	Simulation-based	87.5	High

CNN model

Evaluation metric	Value
Loss	0.2279
Accuracy	0.875

RF classifier

Class	Precision	Recall	F1-score	Support
0 (No disease)	1.00	1.00	1.00	89
1 (disease)	1.00	1.00	1.00	111
Accuracy			1.00	200
Macro avg	1.00	1.00	1.00	200
Weighted avg	1.00	1.00	1.00	200

Data head

Soil health	Weather conditions	Plant health	Disease label
51	56	62	0
92	16	85	0
14	85	1	1
71	89	87	0
60	43	71	0

SVM

Class	Precision	Recall	F1-score	Support
0 (No disease)	1.00	1.00	1.00	89
1 (Disease)	1.00	1.00	1.00	111
Accuracy			1.00	200
Macro avg	1.00	1.00	1.00	200
Weighted avg	1.00	1.00	1.00	200

Articles in this issue

DOI: https://doi.org/10.2478/ijssis-2026-0047 | Journal eISSN: 1178-5608

Journal RSS Feed

Language: English

Submitted on: Jan 2, 2026

|

Published on: Jul 15, 2026

Published by: Macquarie University, Australia

In partnership with: Paradigm Publishing Services

Publication frequency: 1 issue per year

Keywords:

crop disease detection,

simulation-based data,

mobile devices,

machine learning

Related subjects:

Introductions and overviews,

Engineering, other

© 2026 Qi Jing, published by Macquarie University, Australia
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.

Volume 19 (2026): Issue 1 (January 2026)