SugarcaneLD-BD: A Sugarcane Leaf Diseases Dataset for Classification of Sugarcane Leaf Diseases Using Machine Learning

Sugarcane is susceptible to various leaf diseases that can negatively affect plant health and reduce yield. Early detection and proper management of these diseases are crucial for maintaining crop productivity.

SugarcaneLD-BD dataset was compiled to support research on low-cost, non-invasive, and rapid identification of sugarcane leaf diseases using machine learning.

It comprises 638 JPG images across five classes: four major sugarcane leaf diseases (a) Red Rot, (b) Red Leaf Spot, (c) Ring Spot, and (d) Eye Spot, and one class representing healthy leaves.

Images were collected from four locations in Bangladesh:
(a) Bangladesh Sugarcrop Research Institute (BSRI) research field in Pabna
(b) BSRI Gazipur Regional Station research field in Gazipur
(c) Farmers' field in Narsingdi
(d) Farmers' field in Natore

Data collection took place between September and October 2023 using two smartphone cameras. All images were labeled and verified by an expert plant pathologist and resized to 224 × 224 pixels.

Researchers and practitioners can use this dataset to develop and evaluate machine learning models for sugarcane leaf disease detection.

甘蔗易受多种叶部病害侵染，可对植株健康造成负面影响并导致产量下降。对这类病害进行早期检测与科学防控，对维持作物生产能力至关重要。

为支撑利用机器学习开展低成本、非侵入式且快速的甘蔗叶部病害识别研究，SugarcaneLD-BD数据集正式构建完成。

该数据集包含638张JPG格式图像，涵盖5个类别：4种主要甘蔗叶部病害（(a) 赤腐病、(b) 红斑病、(c) 环斑病、(d) 眼斑病），以及1个健康叶片类别。

图像采集自孟加拉国的4处场地：
(a) 帕布纳的孟加拉国甘蔗研究所（Bangladesh Sugarcrop Research Institute, BSRI）试验田
(b) 加齐布尔的BSRI加齐布尔区域站试验田
(c) 纳尔辛迪的农户种植地块
(d) 纳托尔的农户种植地块

数据采集工作于2023年9月至10月期间开展，采用两款智能手机摄像头完成拍摄。所有图像均经植物病理学专家标注与核验，并统一调整至224×224像素尺寸。

研究人员与产业从业者可利用该数据集，开发并评估用于甘蔗叶部病害检测的机器学习模型。