Title: A deep learning based automated phenotyping for identification of overuse of synthetic fertilisers in Amaranthus crop

Authors: J. Dhakshayani; B. Surendiran; J. Jyothsna; A.S. Syed Shahul Hameed; Narendran Rajagopalan

Addresses: Department of Computer Science and Engineering, National Institute of Technology Puducherry, Karaikal, India ' Department of Computer Science and Engineering, National Institute of Technology Puducherry, Karaikal, India ' Department of Horticulture, College of Agriculture, Jawaharlal Nehru Krishi Vishwavidyalaya, Jabalpur, Madhya Pradesh, India ' Department of Computer Science and Engineering, National Institute of Technology Puducherry, Karaikal, India ' Department of Computer Science and Engineering, National Institute of Technology Puducherry, Karaikal, India

Abstract: Amaranth (Amaranthus spp.) is a significant leafy vegetable and cereal crop with high nutrient benefits that is widely consumed worldwide. To maximise its yield, farmers massively rely upon synthetic fertilisers to enhance the quality of the crop. However, this obsessive usage of inorganic fertiliser leads to severe ecosystem damage. For agricultural and ecological sustainability, it is essential to comprehend the process underlying this environmental degradation. This paper analyses the effect of inorganic fertilisers on the growth and yield of Amaranthus. By identifying Amaranthus's productivity and adaptability in different chemically treated soil conditions and automatically phenotyping its traits using image-based deep learning models, this study aims to determine the overuse of synthetic fertilisers. A comparative evaluation of different state-of-art CNN models was carried out, and the experimental result proves that DenseNet-121 could be a more appropriate learning algorithm for the proposed system with 84% accuracy. It is believed that the proposed deep learning based automated phenotyping framework could greatly assist farmers in understanding the actual requirement of soil, thus avoiding the residual impact of fertiliser abuse in the environment.

Keywords: Amaranthus; automated phenotyping; fertiliser overuse; deep learning; DenseNet-121.

DOI: 10.1504/IJCSE.2024.139712

International Journal of Computational Science and Engineering, 2024 Vol.27 No.4, pp.472 - 483

Received: 05 Mar 2023
Accepted: 25 Aug 2023
Published online: 05 Jul 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article