Education, Science, Technology, Innovation and Life
Open Access
Sign In
  1. Home
  2. Journals
  3. Automation and Machine Learning
  4. Vol 7, Issue 1, 2026

Research on Farmland Pest Image Detection and Recognition Based on SE-YOLOv5

Download as PDF

DOI: 10.23977/autml.2026.070114 | Downloads: 4 | Views: 137

Author(s)

Wangxia Xia 1, Jianzhong Sun 1

Affiliation(s)

1 School of Information and Intelligent Engineering, University of Sanya, Sanya, 572022, Hainan, China

Corresponding Author

Wangxia Xia

ABSTRACT

Aiming at the problems of low efficiency and high missed detection rate in traditional manual farmland pest detection, as well as the defects of the original YOLOv5 algorithm in detecting small-target pests and pests in complex backgrounds, an improved YOLOv5 algorithm integrated with the SE attention mechanism (SE-YOLOv5) is proposed to realize accurate and real-time detection of farmland pests. Taking six common rice pests as the research objects, a pest dataset containing 4656 images was constructed, and the robustness of the dataset was improved through preprocessing steps such as label cleaning, size normalization and data augmentation. The SE attention mechanism was embedded between the C3 module and the SPPF layer of the YOLOv5s backbone network to strengthen the model's adaptive weight assignment for pest feature channels, highlight effective features and suppress redundant background features. The experimental results show that the mAP@50 of the SE-YOLOv5 algorithm reaches 97.0%, an increase of 2.0 percentage points compared with the original YOLOv5s. The inference speed for a single image is 6.0 ms, and the detection standard deviation and coefficient of variation are reduced to 0.031 and 0.032, respectively. The detection accuracy of small-target pests such as brown planthoppers is significantly improved. This algorithm performs excellently in detection accuracy, real-time performance and stability, providing an effective technical solution for the intelligent monitoring of farmland pests.

KEYWORDS

Farmland Pest Detection; YOLOv5; SE Attention Mechanism; Small Target Detection

CITE THIS PAPER

Wangxia Xia, Jianzhong Sun. Research on Farmland Pest Image Detection and Recognition Based on SE-YOLOv5. Automation and Machine Learning (2026). Vol. 7, No. 1, 112-118. DOI: http://dx.doi.org/10.23977/autml.2026.070114.

REFERENCES

[1] Wu Yunzhi, Hu Nan, Iftikhar Ahmad, et al. Fine-Grained Classification   of Plant Diseases and Pests Based on Transfer Learning[J]. Journal of Suzhou University, 2024, (9): 29-33.
[2] Hu J, Shen L, Albanie S, et al. Squeeze-and-Excitation Networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(08): 2011-2023.
[3] Sheng Wenshun, Yu Xiongfeng, Lin Jiayan, et al. Small-Scale Object Detection Algorithm Fusing Attention and Feature Pyramid[J]. Computer Engineering, 2024, (1): 242-250.
[4] Zheng Junjian, Lan Yubin, Xiong Wanjie, et al. Pest Identification Method for Pakchoi Based on Improved YOLOv5s Model[J]. Transactions of the Chinese Society of Agricultural Engineering, 2024, 40(13): 124-133.
[5] Qu Haicheng, Wang Meng, Chai Rui. Efficient Vehicle Detection in Remote Sensing Images Based on Bidirectional Multi-Scale Feature Fusion[J]. Computer Engineering and Applications, 2024, (12): 346-356.

Subscription
E-Mail Alert
Most popular papers
Call for Special Issue
Publication Ethics & OA Statement
Plagiarism Policy
Contact Us
Downloads: 5209
Visits: 261689

Sponsors, Associates, and Links

All published work is licensed under a Creative Commons Attribution 4.0 International License.

Copyright © 2016 - 2031 Clausius Scientific Press Inc. All Rights Reserved.