Qin2006/Strawberry_Dataset

Dataset Card for Strawberry Disease Multimodal Dataset

Dataset Description

This is a multimodal dataset for strawberry disease detection, which contains strawberry image data, corresponding environmental parameters (air temperature, air humidity, soil moisture) and strawberry variety information. It can be used to study the correlation between environmental factors and strawberry disease occurrence, as well as multimodal fusion disease detection algorithms.

Supported Tasks

• Object detection (YOLO format annotations provided)
• Image classification
• Multimodal learning (image + environmental parameters)
• Agricultural disease prediction research

Languages

English (documentation), Chinese (original data field names with English translations)

Dataset Structure

Directory Layout

Strawberry_Dataset/
├── images/
│   ├── train/              # Training set images (935 images)
│   └── val/                # Validation set images (235 images)
├── labels/
│   ├── train/              # Training set YOLO format annotations
│   └── val/                # Validation set YOLO format annotations
├── data.yaml               # YOLO training configuration file
├── environment_data.csv    # Multimodal environmental parameters for all images
└── README.md               # Chinese documentation

Data Splits

Split	Number of Images	Number of Annotations
train	935	930
val	235	231
Total	1170	1161

Data Fields

Image Data

• JPG format images of strawberry plants/fruits
• Corresponding YOLO format .txt annotation files with bounding boxes and class labels

Annotation Classes

Class ID	Category Name	Description
0	normal	Healthy strawberry
1	gray_mold	Gray mold (Botrytis cinerea) disease
2	powdery_mildew	Powdery mildew (Podosphaera aphanis) disease
3	black_spot	Black spot (Colletotrichum spp.) disease
4	overripe	Overripe strawberry fruit

Environmental Parameters (environment_data.csv)

The CSV file contains 1171 records with multimodal data for each image:

Field Name (Original)	Field Name (English)	Description
filename	filename	Corresponding image file name
空气温度	air_temperature	Ambient air temperature when shooting, unit: ℃
空气湿度	air_humidity	Ambient air humidity when shooting, unit: %
土壤湿度	soil_moisture	Soil moisture when shooting, unit: %
is_hongyan	is_hongyan	Hongyan strawberry variety: 1=Yes, 0=No
is_xiangye	is_xiangye	Xiangye strawberry variety: 1=Yes, 0=No

Usage

Load with Hugging Face Datasets

from datasets import load_dataset

# Load the full dataset
dataset = load_dataset("your_username/strawberry-disease-multimodal")

# Access splits
train_data = dataset["train"]
val_data = dataset["val"]

# Load environmental data
import pandas as pd
env_df = pd.read_csv("environment_data.csv")

YOLO Training Usage

The dataset is fully compatible with YOLO training frameworks:

# data.yaml already included in dataset
path: ./Strawberry_Dataset
train: images/train
val: images/val
nc: 5
names: ['normal', 'gray_mold', 'powdery_mildew', 'black_spot', 'overripe']

Multimodal Usage

Associate images with environmental parameters using the filename field:

# Get image filename
img_filename = train_data[0]["image"].filename.split("/")[-1]
# Get corresponding environmental data
env_data = env_df[env_df["filename"] == img_filename].iloc[0]

Dataset Curators

Agricultural AI research team

License

This dataset is available under the CC BY-NC-SA 4.0 license.