Integrating Pattern Recognition and CNN-Based Models for Improved Bean Disease Detection and Agricultural Yield Enhancement

Farian S. Ishengoma; Joseph P. Telemala

doi:10.47852/bonviewAIA52024376

Authors

Farian S. Ishengoma Department of Informatics and Information Technology, Sokoine University of Agriculture, Tanzania https://orcid.org/0000-0002-8093-2259
Joseph P. Telemala Department of Informatics and Information Technology, Sokoine University of Agriculture, Tanzania https://orcid.org/0000-0003-3926-3425

DOI:

https://doi.org/10.47852/bonviewAIA52024376

Keywords:

beans, machine learning, convolutional neural network, pattern recognition, contour detection

Abstract

Early detection of plant diseases is crucial for minimizing crop losses and improving agricultural productivity. However, current manual detection methods are time-consuming and inaccurate. Although advanced techniques such as pattern recognition (PR) and image processing have demonstrated potential in automating disease identification, there is still a need for more accurate and efficient methods that can handle various image qualities such as complex backgrounds and complex leaf structures. This paper introduces a method that uses PR and convolutional neural network (CNN) models to detect and classify diseases in infected bean leaves based on shape features. Using contour detection, a computer vision-based PR method, we can effectively detect and distinguish different shapes in images. Two datasets were prepared: original bean leaf images subjected to background removal and enhanced images subjected to background removal, binary thresholding, and contour detection. These preprocessing steps reduced noise, improved color differentiation, and enhanced shape detection for more precise disease identification. Experimental results show that the models trained on the enhanced images showed up to an 8.16% improvement in accuracy, precision, and sensitivity compared to those trained on the original images. This study highlights the potential of integrating image processing techniques with CNN models for more accurate and efficient plant disease detection.

Received: 18 September 2024 | Revised: 9 June 2025 | Accepted: 11 July 2025

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.

Data Availability Statement

The data that support the findings of this study are openly available in AI-Lab-Makerere at https://github.com/AI-Lab-Makerere/ibean/.

Author Contribution Statement

Farian S. Ishengoma: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration, Funding acquisition. Joseph P. Telemala: Software, Validation, Formal analysis, Writing – review & editing.