Chili Leaf Disease Dataset: Annotated Smartphone Images of Anthracnose, Cercospora Leaf Spot, Leaf Curl Disease, and Healthy Leaves in Bangladesh

Introduction
The Chili Leaf Disease Dataset contains 1544 images of chili leaves, captured using smartphone cameras in agricultural fields across Bangladesh. The images are divided into four classes: Anthracnose, Cercospora Leaf Spot, Leaf Curl Disease, and Healthy Leaves. This dataset is designed to assist in the development of machine learning models for automated disease detection in chili plants, supporting agricultural innovation and sustainable farming practices.

Dataset Overview
o Number of Images: 1544
o Classes: 4 — Anthracnose, Cercospora Leaf Spot, Leaf Curl Disease, Healthy Leaves
o Image Sources: Captured using smartphones with varying camera specifications: Redmi 12 (50 MP), Redmi 13 (108 MP), and Tecno Spark 8 Pro (48 MP)
o Geographical Region: Bangladesh
The dataset includes images with varying lighting, backgrounds, and angles to ensure diversity in field conditions. This allows the dataset to reflect real-world conditions that farmers might encounter when diagnosing.

Chili Leaf Diseases
Anthracnose: A fungal disease caused by Colletotrichum species, leading to dark, sunken lesions on leaves and fruits.
Cercospora Leaf Spot: Caused by Cercospora capsici, this disease results in dark, circular spots on leaves, causing premature leaf drop.
Leaf Curl Disease: Caused by viruses like ChiVMV and ToLCV, this disease causes leaves to curl, leading to stunted growth and reduced yield.
Healthy Leaves: Includes disease-free chili leaves, serving as a baseline for comparison with diseased leaves.

Data Collection
The images were captured across various chili fields in Bangladesh using the smartphones listed above. These smartphones were chosen for their accessibility and image quality, reflecting conditions under which farmers typically use smartphones for agricultural tasks. Images were taken from various angles and distances, with different lighting conditions, to simulate real-world scenarios.

Use Cases
Mobile Applications for Farmers: Develop smartphone apps enabling farmers to take pictures of their plants and receive instant diagnoses on disease presence.
Precision Agriculture: Assist farmers by providing early disease detection, reducing pesticide use, and improving crop management.
Agricultural Research: Support studies in plant pathology and machine learning for improved disease diagnosis and management systems.

Conclusion
The dataset is publicly available through Mendeley Data and comes in four different folders class-wise containing the raw JPG images and corresponding CSV metadata files.This dataset provides a valuable resource for developing automated systems that assist farmers in Bangladesh and other regions with disease detection and crop management. By leveraging machine learning, this dataset helps reduce reliance on manual inspection, improves crop health monitoring, and supports more sustainable agricultural practices.

### 引言
本辣椒叶片病害数据集（Chili Leaf Disease Dataset）共包含1544张辣椒叶片图像，均由智能手机相机在孟加拉国各地的农田中采集。图像共分为四类：炭疽病（Anthracnose）、尾孢叶斑病（Cercospora Leaf Spot）、卷叶病（Leaf Curl Disease）以及健康叶片（Healthy Leaves）。本数据集旨在助力开发用于辣椒植株病害自动检测的机器学习模型，为农业创新与可持续耕作实践提供支持。

### 数据集概览
- 图像总数：1544张
- 类别数：4类——炭疽病（Anthracnose）、尾孢叶斑病（Cercospora Leaf Spot）、卷叶病（Leaf Curl Disease）、健康叶片（Healthy Leaves）
- 图像来源：由多款不同相机配置的智能手机采集，具体机型及参数为：Redmi 12（5000万像素）、Redmi 13（1.08亿像素）以及Tecno Spark 8 Pro（4800万像素）
- 采集区域：孟加拉国
本数据集涵盖了不同光照、背景与拍摄角度的图像，以确保田间采集条件的多样性，从而能够真实反映农户在病害诊断过程中可能遇到的实际场景。

### 辣椒叶片病害详解
1. **炭疽病（Anthracnose）**：由刺盘孢属（Colletotrichum）真菌引发，会在叶片与果实表面形成深色凹陷病斑。
2. **尾孢叶斑病（Cercospora Leaf Spot）**：由辣椒尾孢菌（Cercospora capsici）引起，会在叶片上形成深色圆形病斑，导致叶片提前脱落。
3. **卷叶病（Leaf Curl Disease）**：由辣椒脉斑驳病毒（ChiVMV）、番茄曲叶病毒（ToLCV）等病毒引发，会导致叶片卷曲，造成植株生长迟缓、产量下降。
4. **健康叶片（Healthy Leaves）**：指无病害的辣椒叶片，作为对照基线用于与感病叶片进行对比分析。

### 数据采集流程
本数据集的图像均使用上述智能手机，在孟加拉国多个辣椒种植田中采集。选用这些机型的原因在于其普及率高且成像质量优良，贴合农户日常使用智能手机开展农业相关工作的实际场景。采集过程中，拍摄角度、拍摄距离与光照条件均有所差异，以模拟真实的田间诊断场景。

### 应用场景
1. **农户移动应用开发**：可用于开发智能手机应用，帮助农户通过拍摄植株照片即时获取病害诊断结果。
2. **精准农业应用**：通过早期病害检测辅助农户开展生产，减少农药使用量，优化作物管理方案。
3. **农业研究支撑**：可为植物病理学与机器学习相关研究提供支持，助力优化病害诊断与作物管理系统。

### 结语
本数据集可通过Mendeley Data公开获取，按类别分为四个文件夹，内含原始JPG图像与对应的CSV元数据文件。本数据集为开发面向孟加拉国及其他地区的病害检测与作物管理自动化系统提供了宝贵的资源。借助机器学习技术，本数据集有助于降低人工巡检的依赖程度，提升作物健康监测效率，并进一步推动可持续农业实践的发展。