Detecting Phosphorus Levels in Paddy (Oryza sativa) Using Computer Vision: A Precision Agriculture Approach

Abstract

This study detects the phosphorus level in paddy, Oryza sativa, through computer vision to change how agriculture manages phosphorus. Many conventional phosphorus status assessment methods of crops are time-consuming, labor-intensive, and out of reach for smallholder farmers. To meet these challenges, the present study aims to develop a computer vision-based system that detects and monitors real-time phosphorus levels by image analysis of the plant leaves. This study involves two significant experiments. The former has investigated the various phosphorus levels on growth attributes such as plant height, root volume, and yield, indicating the optimal application rate of 184 mg of P per pot. Based on the data from the leaf images, phosphorus level classification was done using the CNN model in the second experiment. The CNN model showed 70% accuracy in the case of phosphorus deficiencies, which, though promising for scalability and further optimization, shows how valuable this study is to gain meaningful insights into the application of AI-based technology in nutrient management. This will contribute to more sustainable and efficient agricultural practices, reducing the environmental risk of improper fertilizer use.