Accurate water level monitoring in Alternate Wetting and Drying rice cultivation using attention-based ConvNeXt architecture

Abstract

The Alternate Wetting and Drying (AWD) method is a rice-growing water management technique promoted as a sustainable alternative to Continuous Flooding (CF). Climate change has placed the agricultural sector in a challenging position, particularly as global water resources become increasingly scarce, affecting rice production on irrigated lowlands. Rice, a staple food for over half of the world’s population, demands significantly more water than other major crops. In Bangladesh, the cultivation of dry season, irrigated Boro rice demands substantial water inputs. Traditional manual water level measurement methods are time-consuming and error-prone, while ultrasonic sensors offer more precise readings but may be affected by environmental factors such as temperature fluctuations, changes in humidity levels, varying light conditions, and accumulation of dust or debris To overcome these limitations, we propose an innovative approach leveraging computer vision, specifically an attention-based ConvNeXt architecture, to automate water height measurement. Our method achieves state-of-the-art performance with an $R^2$ score of 0.989, a Root Mean Squared Error (RMSE) of 0.523 cm, and a Mean Squared Error (MSE) of 0.277 $c{m}^2$, demonstrating superior accuracy and efficiency in managing AWD systems. This advancement represents a significant contribution to sustainable agriculture, enabling precise and automated water management in rice cultivation.