Related Experiment Video
Updated: Jan 13, 2026
05:16
Flying Insect Detection and Classification with Inexpensive Sensors
Published on: October 15, 2014
25.7K
An efficient feature pyramid network with adaptive LSTM for pest detection and classification in IoT
Rajasekaran Arunachalam1, Mohana Jaishankar2, Amit Arora3
1Department of Electronics and Communication Engineering, Saveetha School of Engineering, Saveetha Institute of Medical and Technical Sciences, Thandalam, Chennai, Tamil Nadu, 602105, India.
Scientific Reports
|January 6, 2026
Summary
This study introduces an automated Internet of Things (IoT) system for precise crop pest detection and classification. The new method significantly improves accuracy, crucial for protecting crops and the environment.
Area of Science:
- Agricultural Science
- Computer Science
- Environmental Science
Background:
- Crop pests cause significant global economic losses and environmental damage.
- Manual pest identification is time-consuming, requires expertise, and can lead to pesticide misuse.
- Automated solutions are needed for accurate pest detection and classification to improve crop yields and ecosystem health.
Purpose of the Study:
- To develop and implement an Internet of Things (IoT)-based technique for automated crop pest detection and classification.
- To enhance the precision and efficiency of pest identification compared to traditional manual methods.
- To address the challenges of economic loss and environmental damage caused by crop pests.
Main Methods:
- Collected pest images from a standard database, including IoT sensor-based images.
- Employed a novel Feature Pyramid Network with Multi-Attention Fusion Vision Transformer-based Adaptive Long Short Term Memory (FPN-MAFViT-ALSTM) framework for joint pest detection and classification.
- Optimized FPN-MAFViT-ALSTM parameters using Enhanced and Intelligent Gooseneck Barnacle Optimization with Randomized Exploration (EIGBO-RE) for improved performance.
Main Results:
- The FPN-MAFViT-ALSTM framework successfully detected and classified crop pests from IoT sensor-based images.
- Parameter optimization using EIGBO-RE enhanced the accuracy and efficiency of the pest detection and classification process.
- Experimental verification demonstrated the system's effectiveness under varying conditions.
Conclusions:
- The proposed IoT-based pest detection and classification technique offers a precise and automated solution.
- This approach can mitigate economic losses and environmental damage associated with crop pests.
- The FPN-MAFViT-ALSTM framework, optimized with EIGBO-RE, shows significant potential for agricultural applications.
