A comprehensive diverse Hog Plum (Spondias pinnata) leaf image dataset for precision agriculture and automated crop health assessment

The hog plum (Spondias pinnata) leaves are affected by diseases and insect infestations, which form a significant threat to sustainable crop production, resulting in low yields, low quality of fruit, and losses to agricultural economies such as Bangladesh. We present a full-scale Hog Plum Leaf Image Dataset that is intended to be designed in a multi-class manner to learn how to classify the level of leaf maturity, diseases, and insect attacks. The dataset contains high-resolution images of hog plum leaves taken in natural field conditions, with a wide variety of variations in the real world. There are 11 different classes. All the pictures were well marked with the help of the agricultural experts in order to have the correct annotation and class reliability.
Its dataset includes 5,425 original photographs that were taken during eight months (March 2025 to October 2025) in various locations within Bangladesh (Savar Bank-Town, Daffodil Smart City, and Bangladesh National Botanical Garden). Two Android smartphones, Google Pixel 7 and Samsung Galaxy A56 5G, were used to acquire images in varying environmental conditions. In order to improve the quality of the data and to decrease the class imbalance, a wide range of data augmentation methods was used, including rotation, flipping, zooming, shearing, noise addition, and brightness adjustment, which led to the overall count of 11,000 images, 1,000 of which represent the samples of each class.
#Original Dataset:
5,425 images in .jpg format, Categorized into 11 different classes.
Class Name Number of images (Original)
Anthracnose_Initial : 368
Anthracnose_Severe : 334
Deform_Shape : 454
Dried_Leaf : 505
Healthy_Leaf : 755
Insect_Infestation by White_Fly : 501
Insect_Infestation by Beetle : 538
Senescence_Leaf : 491
Shot_Hole : 490
Sooty_Mould : 500
Young_Leaf : 489

#Augmented Dataset:
11,000 images in .jpg format.
Each class has 1,000 images.

槟榔青（Spondias pinnata）叶片常受病害与虫害侵扰，这对作物可持续生产构成严重威胁，会导致产量下降、果品品质降低，并给孟加拉国等农业经济体带来经济损失。本研究构建了一套完整的槟榔青叶片图像数据集，采用多分类范式设计，用于学习叶片成熟度、病害及虫害侵袭程度的分类方法。该数据集包含在自然田间环境下拍摄的高分辨率槟榔青叶片图像，涵盖真实场景中的多种变异情况，共包含11个类别。所有图像均由农业专家完成标注，以确保标注准确性与类别可靠性。

本数据集的原始照片共5425张，拍摄于2025年3月至2025年10月的8个月间，采集地点涵盖孟加拉国境内的萨瓦尔滨河小镇（Savar Bank-Town）、水仙智能城（Daffodil Smart City）以及孟加拉国家植物园（Bangladesh National Botanical Garden）。图像采集使用了两款安卓智能手机：Google Pixel 7与Samsung Galaxy A56 5G，采集环境涵盖多种不同的现场条件。为提升数据质量并缓解类别不平衡问题，研究团队采用了旋转、翻转、缩放、剪切、添加噪声、亮度调整等多种数据增强方法，最终将数据集扩充至11000张图像，每个类别各包含1000张样本。

# 原始数据集：
共5425张 .jpg 格式图像，划分为11个类别，各原始图像数量如下：
炭疽病初期（Anthracnose_Initial）：368张
炭疽病严重期（Anthracnose_Severe）：334张
形态畸形（Deform_Shape）：454张
枯叶（Dried_Leaf）：505张
健康叶片（Healthy_Leaf）：755张
烟粉虱虫害（Insect_Infestation by White_Fly）：501张
甲虫虫害（Insect_Infestation by Beetle）：538张
衰老叶片（Senescence_Leaf）：491张
穿孔病（Shot_Hole）：490张
煤烟病（Sooty_Mould）：500张
幼叶（Young_Leaf）：489张

# 增强数据集：
共11000张 .jpg 格式图像，每个类别各包含1000张样本。