A Phenotypic Approach for Guava fruit Disease Detection and Classification Using Computer Vision-Based Techniques
Download| Volume 6 Issue 1, 2025 | |
|---|---|
| Author(s): |
Saima Shoro Shah Abdul Latif University, Khairpur, shorosaima@gmail.com Hidayatullah Shaikh Shah Abdul Latif University, Khairpur, hidayat.shaikh@salu.edu.pk Ghulam Ali Mallah Shah Abdul Latif University, Khairpur, ghulam.ali@salu.edu.pk |
| Abstract | We developed an end-to-end, vision-based framework for automated phenotypic screening of four major guava fruit diseases: Anthracnose, Botryodiplodia Rot, Guava Fruit Canker, and Phytophthora Fruit Rot. Our pipeline integrates classical image processing (segmentation, noise reduction, normalization) with a purpose-built lightweight CNN architecture for real-time classification. Trained and validated on a curated dataset of 6,800 labeled field images, the model achieved 92.5% accuracy – outperforming RF (88.2%) and SVM (85.5%) baselines. Critically, it demonstrated exceptional robustness across precision (93.1±0.7), recall (91.8±1.2), and F1-score (92.4±0.9) metrics, minimizing false negatives critical for early detection. Notably, the system exhibits superior adaptability to variable lighting/field conditions versus existing SOTA approaches. This scalability positions it as a viable tool for precision agriculture, directly addressing key pain points: early pathology diagnosis (~48hrs faster than manual scouting), yield preservation (+15–22% in trials), and sustainable production via reduced fungicide use. |
| Keywords | Guava disease detection, digital image processing, convolutional neural network, precision agriculture, machine learning. |
| Year | 2025 |
| Volume | 6 |
| Issue | 1 |
| Type | Research paper, manuscript, article |
| Recognized by | Higher Education Commission of Pakistan, HEC | Category | Journal Name | ILMA Journal of Technology & Software Management | Publisher Name | ILMA University | Jel Classification | -- | DOI | - | ISSN no (E, Electronic) | 2790-590X | ISSN no (P, Print) | 2709-2240 | Country | Pakistan | City | Karachi | Institution Type | University | Journal Type | Open Access | Manuscript Processing | Blind Peer Reviewed | Format | Paper Link | https://ijtsm.ilmauniversity.edu.pk/arc/Vol6/i1/pdf2.pdf | Page | 13-19 |