In this issue, we highlight Mohammad El Sakka, a PhD student at IRIT, who is advancing satellite image analysis and crop classification using deep neural networks. We recap the AI4AGRI summer school in Brasov, Romania, which provided hands-on experience with AI and Earth observation technologies for sustainable agriculture. Upcoming events include an online presentation by Mohammad El Sakka on AI in Smart Agriculture on July 3, 2024. Additionally, we feature the latest AI4AGRI publications on deep learning applications in agriculture, including a study on soil roughness estimation and a review of Convolutional Neural Networks in crop imagery analysis.
Stay tuned for more updates and insights as we continue to explore the intersection of AI and agriculture.
Member Highlight
![](https://ai4agri.unitbv.ro/wp-content/uploads/2024/06/moelsakka_02-576x1024.jpeg)
![](https://ai4agri.unitbv.ro/wp-content/uploads/2024/06/moelsakka_03-1-768x1024.jpeg)
Mohammad El Sakka
PhD student – IRIT, UT3 (Toulouse, France) – SIG Team
Coming from a computer science and aerospace background, I started my PhD thesis at IRIT, University of Toulouse III, in machine learning in agriculture in 2023. My research as a Phd student focuses on applying artificial intelligence and remote sensing techniques to improve agriculture.
I recently participated in the AI4AGRI summer school (2024) in Brasov, Romania. During that event I had the wonderful opportunity of exchanging with senior professionals, experts, researchers, and other PhD students on different topics about smart agriculture. I learnt new skills in hyper/multi-spectral and SAR image processing, as well as UAV and satellite remote sensing.
It was also a great experience to be introduced and to manipulate tools that are used in-situ in smart farms today in addition to the theoretical lectures or labworks, all while discovering Romania’s culture.
AI4AGRI summer school allowed me to improve my skills on some topics that will be helpful for my thesis and future career.
News
AI4AGRI summer school – Artificial Intelligence for Earth Observation Data Analysis
📍 Transilvania University of Brașov
📅 Wed. 8 – Tue. 14 May 2024
The AI4AGRI Spring School 2024 in Brasov, Romania, successfully concluded on May 14. Hosted by the R&D Institute of Transilvania University of Brasov, the event gathered experts and participants from Europe and beyond to explore AI and Earth Observation (EO) technologies in agriculture. Supported by the AI4AGRI European project and EU funding, the week-long program featured lectures, practical sessions, outdoor activities and involved 60 trainees and 11 trainers. Highlights included workshops on hyperspectral image analysis, semantic segmentation, and object detection, using tools like PyTorch and MATLAB for real-world applications in crop and forest monitoring.
A notable session, “A Day in the Potato Field,” provided hands-on agricultural experience. The social event, ‘Junii Brașovului,’ showcased local culture. Mihai Ivanovici, the organizer, emphasized the importance of such initiatives in advancing digital transformation in agriculture. Attendees received certificates of attendance, potentially equivalent to ECTS credits. AI4AGRI expressed gratitude to all involved and announced plans for future educational initiatives to continue fostering innovation in agriculture.
![](https://ai4agri.unitbv.ro/wp-content/uploads/2024/06/AI4AGRI-Summer-school-2024-group-photo-small.jpg)
Upcoming Events
AI in Smart Agriculture (Online presentation)
🧑 Mohammad El Sakka
📅 Wed. 3 July. 2024 – 2 pm (Fr time) | 3 pm (Ro time)
📍Salle des thèses, IRIT – Toulouse (France)
https://univ-tlse2.zoom.us/j/98366074118?pwd=VFVOdVdacTdtSzVNelhvei9Wa3FOQT09
Agriculture is shifting toward smart farming, driven by advanced technologies like Artificial Intelligence. Convolutional Neural Networks act as the “eyes” of AI, allowing it to play a crucial role in this transformation by analyzing images from different sources, such as satellites, drones, and ground vehicles. In the upcoming presentation, Mohammad El Sakka will discuss in more detail how AI helps agriculture, from understanding the needs of plants in agricultural lands to predicting the yield of a field and more.
![](https://ai4agri.unitbv.ro/wp-content/uploads/2024/06/zoom.jpg)
Upcoming meetings
AI4AGRI Monthly meeting
📅 Tue. 25 June – 14:00 (Fr time) | 15:00 (Ro time)
https://univ-tlse2.zoom.us/j/98366074118?pwd=VFVOdVdacTdtSzVNelhvei9Wa3FOQT09
![](https://ai4agri.unitbv.ro/wp-content/uploads/2024/06/zoom.jpg)
Latest AI4AGRI Publications
Deep automatic soil roughness estimation from digital images (April 2024)
M Ivanovici, S Popa, K Marandskiy, C Florea
#Soil roughness #convolutional neural networks #VGG-11 #ResNet-18
https://www.tandfonline.com/doi/full/10.1080/22797254.2024.2342955
In their recent study, Ivanovici, Popa, Marandskiy, and Florea present a novel approach to estimating soil roughness using deep convolutional neural networks. Soil roughness, a critical factor for assessing soil water storage, infiltration, and overland flow, is pivotal for agricultural and soil moisture estimation models. The researchers introduce a framework that combines a specific data acquisition setup—using a red laser beam to project a line on the soil surface followed by digital color image capture—with deep learning models. Specifically, VGG-11 and ResNet-18 convolutional networks were trained in a supervised manner using soil roughness values obtained from a pinboard.
![](https://ai4agri.unitbv.ro/wp-content/uploads/2024/06/ivanovici_pub_fig.jpg)
Multisource Remote Sensing Data Visualization using Machine Learning (March 2024)
I. Cristina Plajer, A. Băicoianu, L. Majercsik, M. Ivanovici
#Multisource multispectral (MS) and hyperspectral (HS) images #neural network #normalization #remote sensing #standardization #visualization
https://ieeexplore.ieee.org/document/10458686
The study conducted by I. Cristina Plajer, A. Băicoianu, L. Majercsik, and M. Ivanovici focuses on the challenge of visualizing multisource multispectral (MS) and hyperspectral (HS) images for enhanced Earth observation. Traditional methods, which often utilize a limited spectrum, lead to significant information loss. This research leverages fully connected neural networks (FCNN) to process aggregated datasets from diverse sources, emphasizing the importance of preprocessing through standardization and normalization. The authors present numerous experiments to validate the effectiveness of AI-based techniques in providing accurate and visually appealing representations of remote sensing data.
![](https://ai4agri.unitbv.ro/wp-content/uploads/2024/06/plajer_pub_fig.gif)
The AI4AGRI project received funding from the European Union’s Horizon Europe research and innovation programme under the grant agreement no. 101079136.
Publishing manager: J. Mothe & S. Molina, UT3 & UT2, IRIT, France