YOLOv12 · Research Deployment

Citrus Greening
AI Detection

A minimal, research-driven AI system for early detection of Huanglongbing disease and nutrient deficiencies in citrus crops.

Developed by KEHEM IT in collaboration with Sylhet Agricultural University
Rounded box curve

System Workflow

1. Image Capture

Leaf images are captured using mobile devices or field cameras under natural lighting conditions.

2. Pre-processing

Images are resized to 416×416 and normalized before inference to ensure consistent detection quality.

3. AI Inference

YOLOv12s model analyzes visual features and localizes disease patterns in real time.

4. Result Output

Detected class and confidence score are returned for decision-making and treatment planning.

Dataset & Detection Classes

The dataset was developed as part of an academic research project funded by the University Grants Commission (UGC) and conducted at Sylhet Agricultural University. A total of 3,000 manually annotated citrus leaf images were collected under controlled and field conditions, with an equal distribution across three classes.

Healthy Leaves (1,000 images)

Healthy citrus leaves exhibit uniform dark-green coloration across the lamina and veins, indicating adequate nutrient uptake and physiological balance. These images serve as the control class and baseline reference during model training and evaluation.

Zinc Deficiency (1,000 images)

Zinc-deficient leaves display symmetrical interveinal chlorosis, reduced leaf size, and narrow leaf structure. This class is critical as zinc deficiency symptoms are often visually confused with citrus greening, leading to misdiagnosis in the field.

HLB-Infected (1,000 images)

HLB-infected leaves show irregular blotchy mottling, asymmetrical chlorosis, and vein corking in advanced stages. These visual patterns were carefully annotated to enable the model to distinguish disease symptoms from nutrient deficiencies.

Images were collected from Sylhet Agricultural University, Zakiganj, and Golapganj using a Nikon D5300 camera. Each leaf was manually annotated using LabelImg in YOLO format and verified multiple times to ensure dataset integrity.

Dataset Composition & Split

To ensure unbiased evaluation and strong generalization, the curated citrus leaf dataset was systematically divided into training, testing, and validation subsets following standard deep learning research practices.

Subset Percentage Number of Images Purpose
Training 70% 2,100 Model learning and parameter optimization
Testing 20% 600 Unseen data evaluation during development
Validation 10% 300 Final performance assessment and model selection

Model Training Pipeline

1. Data Acquisition

Leaf images collected from SAU orchards and nearby subdistricts under controlled and natural conditions.

2. Annotation

Manual bounding-box annotation using LabelImg following YOLO text format standards.

3. Pre-processing

Background normalization, image resizing (416×416), and dataset formatting.

4. Training

YOLOv12s fine-tuned for 20 epochs using optimized hyperparameters and disk caching.

5. Evaluation

Performance measured using precision, recall, F1-score, and mAP metrics.

YOLOv12s Architecture Overview

YOLOv12s is a single-stage object detection model optimized for real-time inference. Its attention-based backbone and multi-scale feature aggregation make it well suited for fine-grained agricultural disease detection tasks.

Backbone

Extracts spatial and semantic features such as leaf texture, vein structure, and color variation.

Neck

Aggregates multi-scale features, enabling accurate detection of leaves with varying sizes and orientations.

Head

Predicts bounding boxes, class probabilities, and confidence scores in a single forward pass.

Model Performance

Precision87.1%
Recall89.9%
mAP5096.0%
mAP50–9594.9%

Architecture: YOLOv12s

Image Size: 416 × 416

Epochs: 20

Optimizer: AdamW

IoU Threshold: 0.7

Technical Highlights

Real-time inference
High accuracy detection
Robust augmentation
Multi-class support

AI for Sustainable Citrus Farming

Research-backed technology designed for growers, researchers, and policymakers.