Design and Development of Weed Detection Mechanism in Wheat using AI
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Deep learning constitutes an ongoing, modern technique for im-age processing and data analysis, with promising outcomes and enormous potential.As deep learning has been effectively ap-plied in different domains, it has likewise entered the area of agriculture.Weed control is important for high quality yields as it works the same as pest for crops.It is the need of the hour to limit the effects and price of non-essential herbicides in turn to reduce ill effects on the health of living beings as well as crops, to attain the goal of sustainable agriculture.In this paper, we have performed a comparison between the algorithms of machine learning and deep learning.SVM with non-linear kernel function and Random Forest is applied to obtain the classification accura-cy.We also proposed a novel weed detection mechanism that relies on deep fully convolutional neural network architecture which consists of an encoder-decoder structure followed by a pixel-wise classification layer, termed as SegNet.Alongside Se-gNet, another convolution network architecture, called U-Net is also applied.The proposed approach is experimentally validated on detection of Phalaris minor weed in wheat crops.The dataset contains RGB images taken during the month of December and January within the area of Punjab.The findings indicate that SegNet and U-Net provide higher accuracy and outperforms other commonly used techniques.
Fields of Study & Subjects
Smart Agriculture and AI
Advanced Neural Network Applications
Industrial Vision Systems and Defect Detection
Publication details
- Journal / source
- Advances in engineering research/Advances in Engineering Research
- Pages
- 117–130
- Publication date
- 30 March 2026
- Language
- EN
- Document type
- Book chapter