Related Experiment Video
Updated: May 17, 2025
08:47
Author Spotlight: UAV Remote Sensing for Efficient Invasive Plant Biomass Estimation
Published on: February 9, 2024
1.3K
Hybrid deep learning model for density and growth rate estimation on weed image dataset
Anand Muni Mishra1, Mukund Pratap Singh2, Prabhishek Singh2
1Dept of Information technology, GL Bajaj Institute of Technology and Management, Greater Noida, Uttar Pradesh, India.
Scientific Reports
|April 2, 2025
Summary
This study introduces a hybrid Convolutional Neural Network (HCNN) model for precise weed identification in agriculture. The advanced model achieved 98.95% accuracy, aiding farmers in managing invasive plants and improving crop yields.
Area of Science:
- Agricultural Science
- Computer Vision
- Machine Learning
Background:
- Crop production faces significant challenges globally, with weeds impacting 37% of crops.
- Weeds reduce crop purity through interbreeding and compete for resources.
- Effective weed management is crucial for food security and farmer livelihoods.
Purpose of the Study:
- To develop an accurate weed image segmentation model for estimating weed growth and density.
- To propose a hybrid Convolutional Neural Network (HCNN) model integrating SegNet and U-Net architectures.
- To enhance weed identification by modifying pooling layers and loss functions.
Main Methods:
- Utilized a dataset of 2100 weed images, including diverse weed types (broadleaf, monocot, dicot).
- Developed a hybrid Convolutional Neural Network (HCNN) model by combining SegNet and U-Net features.
- Implemented modified pooling layers and adjusted loss functions to improve weed leaf identification.
Main Results:
- The proposed HCNN model achieved a high accuracy of 98.95% in weed image segmentation.
- Segmentation masks effectively isolated weeds from crops, enabling growth and density estimation.
- The modifications improved the identification weight of weed leaves.
Conclusions:
- The developed HCNN model offers a highly accurate solution for weed identification and management.
- Accurate weed detection can significantly reduce financial losses in agriculture.
- This research provides a foundation for developing effective weed control strategies.
