Development of an IoT Based Irrigation Control System using Convolutional Neural Network

No Thumbnail Available

Date

2023-06-22

Journal Title

Journal ISSN

Volume Title

Publisher

JOURNAL OF SCIENCE TECHNOLOGY AND EDUCATION

Abstract

The automation of irrigation activities has the potential to revolutionize traditional manual and static irrigation practices, leading to increased productivity with reduced human intervention. Manual irrigation practices often result in water wastage or inadequate water supply to specific crops, as different crops have varying water requirements (crop water need). Moreover, manual irrigation methods consume significant time and effort, especially when the farmland is located at a distance. This paper presents an IoT-based irrigation system that utilizes computer vision technology to capture and recognize crops in the irrigation field using a Convolutional Neural Network (CNN) model. The developed system continuously monitors and maintains the optimal soil moisture content for each specific crop, employing soil moisture and temperature sensors. The control unit of the system is implemented using the Raspberry Pi 3b+ platform. The performance of the developed system was evaluated using two key metrics: Accuracy and Response time. The CNN model achieved high accuracy, with a stabilized accuracy of 95 percent after 50 epochs of training and validation, using a dataset of 800 pictures. This indicates the system's capability to accurately identify crops in the field. The response time of the system was assessed through ten trials, resulting in an average response time of 14.3 seconds, which is considered satisfactory. The findings of this study demonstrate the effectiveness of the proposed IoT-based irrigation system in automating irrigation processes and optimizing water usage. By integrating crop recognition, soil moisture monitoring, and temperature sensing, the system ensures efficient irrigation practices, reducing water wastage and minimizing human effort. The successful implementation of the developed system paves the way for intelligent and dynamic irrigation systems, fostering higher agricultural productivity and sustainable water resource management

Description

This paper introduces a novel IoT-based smart irrigation system that integrates computer vision and sensor technologies to automate and optimize irrigation practices in agriculture. Traditional manual irrigation methods often lead to inefficiencies, including water wastage and inadequate supply for different crops, due to varying crop water needs and the time-intensive nature of manual processes. To address these challenges, the proposed system employs a Convolutional Neural Network (CNN) model for real-time crop recognition using images captured from the field. Coupled with soil moisture and temperature sensors, the system dynamically maintains optimal moisture levels for each crop. The system's control architecture is built on the Raspberry Pi 3b+ platform, enabling local processing and actuation based on the data collected. Experimental evaluations were conducted using a dataset of 800 crop images, achieving a CNN model accuracy of 95% after 50 training epochs. Additionally, the system demonstrated a satisfactory average response time of 14.3 seconds across ten trials. These results validate the system’s ability to accurately identify crops and respond promptly to irrigation needs. Overall, the paper highlights the potential of combining machine learning with IoT for precision agriculture. The developed solution significantly reduces human intervention, enhances water usage efficiency, and supports sustainable farming by enabling intelligent, responsive irrigation tailored to specific crop requirements.

Keywords

internet of things, CNN, Raspberry pi, Irrigation control

Citation

Adamu, M., Abdul-Malik, U. T., Maliki, D., & Paul T.S. (2023). Development of an IoT Based Irrigation Control System using Convolutional Neural Network. Journal of Science Technology and Education (JOSTE). 11(2), pp. 225-235. available at: www.atbuftejoste.com.

Endorsement

Review

Supplemented By

Referenced By