Title: A Multi-modal Approach for Crop Health Mapping using low altitude Remote Sensing, Internet of Things (IoT) and Machine Learning

Biography:

Dr. Rafia Mumtaz received her Ph.D. degree from the University of Surrey, the UK in 2010. She was awarded a Scholarship for her Ph.D. Studies abroad by the National University of Sciences and Technology (NUST) from 2006 to 2010. She was awarded the Endowment Fund Scholarship for her MS studies, by NUST from 2004 to 2006. She joined NUST-School of Electrical Engineering and Computer Science (SEECS) in 2010  and currently serving as Tenured Associate Professor and the director of the Internet of Things (IoT) lab at NUST-SEECS. She served as Head of the Information Technology department from Feb 2018 to Feb 2021.

She has a commendable research record and was conferred the Best Researcher Award, 2018 by Rector NUST on 3rd December 2019. In addition, she has been awarded the Women of Wonder award by inspiremill and WoWpk for her tremendous contributions in the field of precision agriculture. Her name has been included in the directory of Productive Scientist of Pakistan by Pakistan Council for Science and Technology (PCST), 2017. She has been elevated to Senior IEEE member for her meritorious services in the field of engineering in 2017.

She has published high-quality research papers in reputed ISI indexed and impact factor journals throughout her research career, which demonstrates her excellent research aptitude and potential for future research in the areas of expertise. She has secured several competitive and prestigious national and international research grants as a principal investigator and trained various students as part of her received funding, which is a key for the sustainability of education programs in a developing country like Pakistan. Until now she has secured around PKR 43million of research grants from several national and international funding bodies. She has been awarded 2 x research patents by the Intellectual Property Right office (IPO), Pakistan.

She went beyond her national networking circle and collaborated with top research groups in leading institutes like the University of Leeds, UK, University of Glasgow, UK, Queen Mary University of London, UK, University of Malaga, Spain, Technical University of Kaiserslautern, Germany, and German center for Research of Artificial Intelligence (DFKI), Germany.

Her key research initiative includes the establishment of the Internet of Things (IoT), the lab at NUST-SEECS in 2018. Considering her contributions and accomplishments her trajectory and contributions are at par with academicians in top universities worldwide and are considered exceptional compared to her peers.

 

Abstract

 The agriculture sector holds paramount importance in Pakistan due to the agrarian 2 nature of the economy. Pakistan has its GDP based on agriculture, however, it relies on the manual monitoring of crops, which is a labor-intensive and ineffective method. In contrast to this, several cutting-edge technology-based solutions are used in developed countries to enhance crop yield with optimal resources. Pakistani industry, while slow to initially respond to AI and IoT culture, has recently exhibited a great potential to adapt and apply these technologies for solving local challenges. To this end, we have proposed an integrated approach for monitoring crop health using IoT, machine learning, and drone technology, where the IoT sensors provide real-time status of environmental parameters impacting the crop, and the drone platform provides the multi-spectral data used for generating Vegetation Indices (VIs) such as Normalized Difference vegetation Index (NDVI) for analyzing crop health. The NDVI provides the health status of the crop based on the chlorophyll content which offers limited information regarding the crop health. In order to obtain detailed crop information, health maps are generated by combining IoT and multi-spectral data. For this purpose, multi-modal data is collected in different development stages of the crop and labeled to perform pixel-wise classification. A number of machine and deep learning algorithms are applied to the collected data where a deep neural network with two hidden layers exhibited the highest accuracy (98.4%) relative to all other models. Further, IoT sensors data maps were generated which helped to correlate the real-time meteorological conditions with the affected areas delineated in the crop health maps. The benefit of the proposed approach is that it would minimize the crop ground survey by providing detailed and near real-time crop health information in the form of crop health maps.