Data in Brief 52 (2024) 110046

Contents lists available at ScienceDirect

Data in Brief

journal homepage: www.elsevier.com/locate/dib

Data Article

Rice pest dataset supports the construction of

smart farming systems
Luyl-Da Quach∗, Quoc Khang Nguyen, Quynh Anh Nguyen,
Le Thi Thu Lan
FPT University, Can Tho campus, Cantho city, Vietnam

a r t i c l e i n f o a b s t r a c t

Article history: Rice holds a significant position in the global food sup-
Received 10 October 2023 ply chain, particularly in Asian, African, and Latin American
Revised 2 January 2024 countries. However, rice pests and diseases cause significant
Accepted 4 January 2024 damage to the supply and growth of the rice cultivation in-
Available online 10 January 2024
dustry. Therefore, this article provides a high-quality dataset
Dataset link: Rice Pest Dataset Supports
that has been reviewed by agricultural experts. The dataset
The Construction of Smart Farming Systems is well-suited to support the development of automation sys-
(Original data) tems and smart farming practices. It plays a vital role in fa-
cilitating the automatic construction, detection, and classifi-
Keywords:
cation of rice diseases. However, challenges arise due to the
Deep learning
diversity of the dataset collected from various sources, vary-
Machine learning
Image segmentation
ing in terms of disease types and sizes. This necessitates sup-
Computer vision port for upgrading and enhancing the dataset through var-
Rice disease ious operations in data processing, preprocessing, and sta-
tistical analysis. The dataset is provided completely free of
charge and has been rigorously evaluated by agricultural ex-
perts, making it a reliable resource for system development,
research, and communication needs.
© 2024 The Author(s). Published by Elsevier Inc.
This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/)

∗
Corresponding author.
E-mail address: [email protected] (L.-D. Quach).
Social media: @quachluylda (L.-D. Quach)

https://doi.org/10.1016/j.dib.2024.110046
2352-3409/© 2024 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/)
2 L.-D. Quach, Q.K. Nguyen and Q.A. Nguyen et al. / Data in Brief 52 (2024) 110046

Specifications Table

Subject Agricultural Sciences

Specific subject area Image Processing, Image Identification, Image classification, object detection,
computer vision, artificial intelligence, deep learning and reinforce learning
Data format Raw Image
Type of data Image
Data collection Dataset collected image data from the Internet through datasets such as
ImageNet [1], Microsoft COCO [2], images filtered from the IP102 dataset [3],
and images gathered from the internet.
Data source location All image data were labelled and evaluated by experts at the Cuu Long Delta
Rice Research Institute, Vietnam.
Data accessibility Repository name: Rice Pest Dataset Supports The Construction of Smart
Farming Systems
Data identification number: 10.5281/zenodo.8418217
Direct URL to data: https://zenodo.org/record/8418217
Instructions for accessing these data: Users can directly download data from
the URL, then decompress the data set and can use it.
Related research article Luyl-Da Quach, Khang Nguyen Quoc, Anh Nguyen Quynh and Hoang Tran
Ngoc, “Evaluation of the Efficiency of the Optimization Algorithms for Transfer
Learning on the Rice Leaf Disease Dataset” International Journal of Advanced
Computer Science and Applications(IJACSA), 13(10), 2022.
http://dx.doi.org/10.14569/IJACSA.2022.0131011

1. Value of the Data

• The dataset has a size of approximately 69.6 MB and comprises 3,156 images belonging to
10 classes. Each class contains various images related to rice pests and diseases, with the im-
age resolution standardized to 312×312 pixel. This dataset is diverse as it was curated from
various sources, resulting in differences in brightness, image resolution, and environmental
conditions. This diversity poses challenges in processing for pest and disease detection, clas-
sification, and identification.
• This dataset, which classifies different types of rice pests and diseases, can be used to develop
a system that supports the identification process, builds monitoring systems, and makes pre-
liminary steps toward creating a smart farming system. The dataset contributes significantly
to the development of systems for monitoring and classifying pests and diseases, such as
surveillance and classification systems.
• The dataset has the potential to enhance the effectiveness of detecting and mitigating rice
pests and diseases using smart agriculture systems. It also serves as a case study for develop-
ing eXplainable Artificial Intelligence algorithms, as seen in the classification of physiological
states in tomatoes [4].
• Although there are several widely published datasets on harmful rice pests, most of these
datasets are highly diverse. However, these datasets often contain repeated images, unclear
distinctions between different types of harmful pests, and a lack of official management by
agricultural research institutes. Therefore, the contribution of this research lies in combining
expert input, filtering out duplicate image data, and standardizing image normalization. This
aids in supporting subsequent studies by contributing and processing data related to larvae,
eggs, and adults.
• This dataset is expected to yield more positive outcomes through human collection and eval-
uation by experts in the field of agriculture.

2. Background

According to the International Rice Research Institute’s report, rice pests and diseases
can cause up to a 37% reduction in rice yield for farmers, with the actual impact on pro-
duction ranging from 24% to 41% depending on the specific agricultural conditions. This
 L.-D. Quach, Q.K. Nguyen and Q.A. Nguyen et al. / Data in Brief 52 (2024) 110046 3

highlights the significant global concern regarding the influence of pests and diseases on rice
production.
However, statistical data from scholarly sources indicates that from 2019 to the present, there
have been approximately 17,0 0 0 research studies related to the keyword "rice leaf disease de-
tection" and about 17,400 studies related to "rice pest detection" (content related to pests and
rice but not necessarily containing the complete search phrase). Some notable studies on rice
diseases are [5,6]. This demonstrates the scientific community’s keen interest in developing a
dataset to meet research needs in this field.

3. Data Description

The collected data includes images of harmful rice pests, primarily those affecting the leaves.
For this research, image data was curated from various publicly available datasets on different
types of rice pests, with a significant portion sourced from the IP102 dataset. In total, 3,156
images were gathered, representing 10 different rice pest species, which are Asiatic rice borer
(Chilo suppressalis), Brown planthopper, Paddy stem maggot (Hydrellia sasakii), Rice gall midge
(Orseolia oryzae), rice leafroller (Cnaphalocrocis medinalis), Rice leaf caterpillar, Rice leafhopper,
Rice water weevil, Small brown planthopper, and yellow rice borer. These data were organized
into ten folders, each in a zip file format, and the image format used was JPEG with a consistent
size of 312 × 312 pixels.
To make data uploading and downloading simple, images were divided into separate folders
by zip file. Fig. 1 illustrates sample data with 10 types of pests collected in the dataset.

4. Experimental Design, Materials and Methods

4.1. Field data collection

The data was collected and extracted from the IP102 dataset containing images of rice pests
by a research team affiliated with the College of Computer Science, Nankai University, Tianjin,
China, using sources from the Internet, ImageNet, and the Microsoft COCO dataset. During the
evaluation and selection of image data, the research team observed that there were many noisy
and feature-repetitive images. Consequently, the research team proceeded to gather and reeval-
uate the dataset of rice pest images. Subsequently, they sought the expertise of specialists in
rice diseases from the Cuu Long Delta Rice Research Institute, Vietnam, to assess the reliability
of the dataset.
The complexity of the datasets arises from the fact that the selected datasets are relatively
intricate, requiring a focus on data relevant to harmful insect pests in rice. For example:

– The IP102 dataset [1] comprises over 75,0 0 0 images belonging to 102 categories, classifying
insects hierarchically and affecting agriculture in general.
– The Microsoft COCO dataset [2] contains more than 330,0 0 0 images (with over 20 0,0 0 0 la-
beled images) featuring 81 object categories and 91 stuff categories.
– The ImageNet dataset [7] includes over 14 million images with more than 21,0 0 0 indexed
sets.
– The remaining data sources, consisting of researched and evaluated images, amount to over
10,0 0 0 images.

Therefore, the author’s selection process from the extensive datasets is illustrated in Fig. 2.
The results of the selection process are based on criteria outlined in Table 1.
4 L.-D. Quach, Q.K. Nguyen and Q.A. Nguyen et al. / Data in Brief 52 (2024) 110046

Fig. 1. Image data sample in dataset.

Fig. 2. Illustration of the image collection and processing process for the pest dataset.
 L.-D. Quach, Q.K. Nguyen and Q.A. Nguyen et al. / Data in Brief 52 (2024) 110046 5

Table 1
Statistics of original image data collected.

No. Class (Common Name) Scientific Name Total Images Adults Eggs Larvae

01 Asiatic rice borer Chilo suppressalis 498 155 24 319

02 Brown planthopper Nilaparvata lugens 346 335 7 4
03 Paddy stem maggot Chlorops oryzae Matsumura 89 86 1 2
04 Rice gall midge Orseolia oryzae Wood-Mason 217 217 0 0
05 Rice leaf roller Cnaphalocrocis medinalis 153 108 1 44
06 Rice leaf caterpillar Mythimna separata 716 108 1 44
07 Rice leaf hopper Nephotettix spp. 244 243 1 0
08 Rice water weevil Lissorhoptrus spp. 414 414 0 0
09 Small brown planthopper Laodelphax striatellus 243 243 0 0
10 Yellow rice borer Scirpophaga incertulas 236 213 18 5
Total 3.156 2.122 53 418

4.2. Data preprocessing

The original image data was processed, involving data cleaning, resizing, and evaluation of
reliability by agricultural experts, as illustrated in Fig.2. The expert evaluation process is carried
out as follows: one person labels the data, and another person reviews the data; this part of the
process is conducted independently. The data is considered valid when both individuals assign
matching labels. In the case of label discrepancies, the two experts reevaluate the issue with
the support of a system contributed by the author team, following a standardized procedure to
assist in machine learning and deep learning processes.
An overview of the original data before processing is shown in Table 1. The images in the
study have been re-checked for data duplication and have not undergone image transformations.
The illustrative images depicting the classification of various types of insects in each develop-
mental stage, such as larvae, eggs, and adults, are presented in Fig. 3.

Fig. 3. Illustrates the development process of each type of worm according to Adults, Eggs, and Larvae in the dataset.
6 L.-D. Quach, Q.K. Nguyen and Q.A. Nguyen et al. / Data in Brief 52 (2024) 110046

Limitations

Not applicable.

Ethics Statement

This study did not conduct experiments involving humans and animals.

Data Availability

Rice Pest Dataset Supports The Construction of Smart Farming Systems (Original data)

CRediT Author Statement

Luyl-Da Quach: Conceptualization, Methodology, Data curation, Writing – original draft, Vi-
sualization, Investigation, Writing – review & editing; Quoc Khang Nguyen: Conceptualization,
Methodology, Data curation, Writing – original draft, Visualization, Investigation, Writing – re-
view & editing; Quynh Anh Nguyen: Data curation, Visualization, Investigation; Le Thi Thu Lan:
Conceptualization, Methodology, Data curation, Investigation, Writing – review & editing.

Acknowledgements

The authors would like to thank the experts at Cuu Long Delta Rice Research Institute, Viet-
nam for supporting the data evaluation and labelling process during implementation. Email of
the person in charge: [email protected] – Dr. Chau La Hoang.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal rela-
tionships that could have appeared to influence the work reported in this paper.

References
[1] J. Deng, W. Dong, R. Socher, L.-J. Li, Kai Li, Li Fei-Fei, ImageNet: a large-scale hierarchical image database, in: 2009
IEEE Conference on Computer Vision and Pattern Recognition, IEEE, Miami, FL, Jun. 2009, pp. 248–255, doi:10.1109/
CVPR.2009.5206848.
[2] T.-Y. Lin, et al., Microsoft COCO: common objects in context, in: D. Fleet, T. Pajdla, B. Schiele, T. Tuytelaars (Eds.),
Computer Vision – ECCV 2014, Lecture Notes in Computer Science, 8693, Springer International Publishing, Cham,
2014, pp. 740–755, doi:10.1007/978- 3- 319- 10602- 1_48.
[3] X. Wu, C. Zhan, Y.-K. Lai, M.-M. Cheng, J. Yang, IP102: a large-scale benchmark dataset for insect pest recognition,
in: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, Long Beach, CA, USA, 2019,
pp. 8779–8788, doi:10.1109/CVPR.2019.00899.
[4] L.-D. Quach, K.N. Quoc, A.N. Quynh, N. Thai-Nghe, T.G. Nguyen, Explainable deep learning models with gradient-
weighted class activation mapping for smart agriculture, IEEE Access 11 (2023) 83752–83762, doi:10.1109/ACCESS.
2023.3296792.
[5] L.-D. Quach, K.N. Quoc, A.N. Quynh, H.T. Ngoc, Evaluation of the efficiency of the optimization algorithms for transfer
learning on the rice leaf disease dataset, IJACSA 13 (10) (2022), doi:10.14569/IJACSA.2022.0131011.
[6] L.-D. Quach, A.N. Quynh, K.N. Quoc, N.N. Thai, Using optimization algorithm to improve the accuracy of the CNN
model on the rice leaf disease dataset, in: C. So-In, N.D. Londhe, N. Bhatt, M. Kitsing (Eds.), Information Systems for
Intelligent Systems, Smart Innovation, Systems and Technologies, 324, Springer Nature Singapore, Singapore, 2023,
pp. 535–544, doi:10.1007/978- 981- 19- 7447- 2_47.
[7] Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, Li Fei-Fei, Imagenet: a large-scale hierarchical image database,
in: 2009 IEEE conference on computer vision and pattern recognition, Ieee, 2009, pp. 248–255.