MUMBAI, India, Jan. 2 -- Intellectual Property India has published a patent application (202541123929 A) filed by Vignan's Foundation For Science, Technology And Research, Vadlamudi, Andhra Pradesh, on Dec. 9, 2025, for 'transforming weed detection: a vision transformer approach for soybean crops.'
Inventor(s) include Dr. Ch. Venkata Rami Reddy; Mrs. Parvathi Devi Budda; Dr. A. Ranganadha Reddy; Sanjay M; Deepashree P. Vaideeswar; and Mithisha Brilent Tavares.
The application for the patent was published on Jan. 2, under issue no. 01/2026.
According to the abstract released by the Intellectual Property India: "Weeds are unwanted plants that grow in agricultural crops, competing with them for resources such as sunlight, water, and nutrients, leading to significant yield losses. Therefore, early detection and precise identification of weeds are essential for effective weed management. This paper proposes the use of Vision Transformers for weed detection and classification in soybean crops. The proposed approach uses a deep learning framework and makes use of Vision Transformer's advantages to increase the accuracy of weed detection and classification. The dataset used in this study contains over 15,336 images of soybean crops, broadleaf, grass weeds and soil. The images were captured utilising a drone equipped with a high-quality camera. The dataset was divided into training, validation, and testing sets, with 81%, 9%, and 10% of the images in each set, respectively. The first step in the proposed method is to train a Vision Transformer model on the training set. The capacity of vision transformers to detect distant dependencies in pictures has lately increased their prominence in the computer vision sector. The trained model is then used to generate patches for each input image in the validation and testing sets. The proposed method is evaluated using accuracy. The experimental analysis of the proposed method demonstrates that the approach outperforms several state-of-the-art methods for weed detection and classification. The accuracy attained by the proposed approach is 98.83%, which is significantly higher than the accuracy of other methods, such as Convolutional Neural Networks and Support Vector Machines. In conclusion, the proposed method of using Vision Transformers for weed detection and classification in soybean crops is effective and efficient. The approach discussed has the potential to be applied to various types of crops, not just limited to soybean, and could enhance the precision and reliability of weed detection and classification. Consequently, this could lead to improved weed management practices and increased crop productivity."
Disclaimer: Curated by HT Syndication.
