IRMA-International.org: Creator of Knowledge
Information Resources Management Association
Advancing the Concepts & Practices of Information Resources Management in Modern Organizations

A Non-Invasive Revolution in Soil and Root Water Content Monitoring by Integrating Deep Learning and GPR

A Non-Invasive Revolution in Soil and Root Water Content Monitoring by Integrating Deep Learning and GPR
View Sample PDF
Author(s): Mohammed Kahlaoui (National School of Applied Sciences, Morocco), Aboulkacem Karkri (National School of Applied Sciences, Morocco)and Mohammed Anisse Moutaouekkil (National School of Applied Sciences, Morocco)
Copyright: 2026
Pages: 30
Source title: Computational Intelligence and Optimization Methods for Sustainable Water Management
Source Author(s)/Editor(s): Yassine Ezaier (Hassan II University, Casablanca, Morocco), Rajae Gaamouche (Moroccan School of Engineering Sciences, Rabat, Morocco)and Mohamed Lahby (Hassan II University, Casablanca, Morocco)
DOI: 10.4018/979-8-3373-2700-6.ch010

Purchase

View A Non-Invasive Revolution in Soil and Root Water Content Monitoring by Integrating Deep Learning and GPR on the publisher's website for pricing and purchasing information.

Abstract

This chapter introduces a technique that combines simulations using Ground Penetrating Radar (GPR), advanced processing of signals and deep learning to determine water content in plant roots as well as nearby soil. Using dielectric mixing models, computer programs were used to create synthetic radargrams of root-soil interactions under various moisture levels. Radargrams were improved by using methods such as Singular Value Decomposition and Inverse Amplitude Decay. Convolutional Neural Networks (CNNs), including InceptionV3, MobileNetV2 and ResNet50V2, were adapted for using in classifying both root and soil water content. InceptionV3 got the top results among tested models for estimating plant roots. For easier soil moisture classification, MobileNetV2 was able to run quickly and efficiently. ResNet50V2 managed to perform with similar accuracy and level of stability for both types of tasks.

Related Content

Muhammad Naeem, Salman Memon, Anita Larik, Syed Rizwan Mehdi, Hasan Ahmed Faridi, Khalida Khan, Sana Zafar, Manoj Kumar. © 2026. 20 pages.
Imdad Ali Shah, N. Z. Jhanjhi. © 2026. 12 pages.
Hafsa Muzammal, Muhammad Zaman, Muhammad Safdar, Muhammad Adnan Shahid, Zuhaib Nishtar, Muhammad Bilal, Muntaha Munir, Mehar Muhammad Haseeb, Aamir Raza, Syed Intsar Hussain Shah, Usman Zafar, Nalain E. Muhammad, Hafiz Muhammad Bilawal Akram. © 2026. 30 pages.
Luminita Diaconu, Yassine Mouniane. © 2026. 32 pages.
Kumar J. Parmar, Tejas Chandulal Chauhan, T. Premavathi. © 2026. 32 pages.
Mahmoud Oudghiri, Mohamed El Bakkali, Yassine Mouniane, Nagla Abid, Samah Bouhassoun, Fatima-ezzahra Jaayefar, Fath Alah Elwahab, Issam El-Khadir, Ahmed Chriqui, Mohammed Ibriz. © 2026. 26 pages.
Issam El-Khadir, Yassine Mouniane, Ahmed Chriqui, Mohamed El Bakkali, Driss Hmouni. © 2026. 34 pages.
Body Bottom