Die an unserem Lehrstuhl durchgeführten Arbeiten befassen sich mit dem Software- und Systems Engineering im Bereich Digital Farming. Hierfür betrachten wir Systeme entlang der gesamten Wertschöpfungskette. Unser Standort in Rheinland-Pfalz bietet durch die vielseitige Abdeckung verschiedener Bereiche, wie Wein-, Gemüse- und Getreideanbau, aber auch der unmittelbaren Nähe zu Forschung und Industrie eine hervorragende Grundlage.

Interessierte Studierende auf der Suche nach Abschlussarbeiten und Projekten sind stets willkommen, mit uns Kontakt aufzunehmen, um Interessen und mögliche Themen zu besprechen.

Unsere Forschungsschwerpunkte:

• Software- und Systems Engineering im Bereich Digital Farming
• Anforderungsanalyse der verschiedenen Akteure im landwirtschaftlichen Ökosystem
• Verbesserung der Interoperabilität und Vernetzung zwischen Akteuren und Systemen
• Verbesserung der Nutzungsakzeptanz

Research Objectives:

• Investigate the current state-of-the-art methods in crop disease recognition and their limitations, particularly concerning data scarcity.

• Propose novel strategies for effectively utilizing limited data in training deep learning models for crop disease recognition.

• Design and implement a deep learning framework tailored for crop disease recognition, emphasizing techniques such as transfer learning, data augmentation, and semi-supervised learning.

• Evaluate the proposed framework on diverse crop disease datasets with varying degrees of data scarcity, comparing its performance against baseline methods.

• Analyze the effectiveness of different techniques employed in the proposed framework and provide insights into their contributions to model generalization and robustness.

• Explore potential applications and implications of the developed framework in real-world agricultural settings, considering factors such as scalability, computational efficiency, and practical usability.

Expected Contributions:

• Development of a novel deep learning framework for crop disease recognition with limited data, incorporating innovative techniques to address data scarcity challenges.

• Empirical evaluation of the proposed framework on diverse crop disease datasets, demonstrating its effectiveness and robustness compared to existing methods.

• Insights into the effectiveness of different strategies employed in handling limited data for crop disease recognition, providing guidance for future research and practical applications in agriculture.

Contact:

Vishal Sharbidar Mukunda

Introduction: 

The agricultural sector is increasingly integrating artificial intelligence (AI) technologies to increase efficiency, optimize use of resources to improve decision-making, and enhance productivity. Despite these advancements, the complexity and opacity of many AI systems pose a challenge for trust and acceptance among farmers. Explainable artificial intelligence (XAI) techniques aim to address these issues by providing transparency into AI decisions, making them understandable to users. While XAI applications are well-researched in other fields, their role in addressing agricultural challenges is still under-explored. Objective: This study aims to investigate how XAI techniques are applied to agricultural challenges through a systematic literature review (SLR). Objectives include identifying important agricultural problems that use XAI, categorizing the techniques used, and evaluate their effectiveness. The ultimate goal is to provide actionable insights for researchers and practitioners to improve the use and adoption of XAI in agriculture.

Research questions: 

RQ1: What are problems in dairy farming where XAI has been applied? 

RQ2: Which XAI techniques are most commonly used in dairy farming? 

RQ3: How effective are these XAI techniques at addressing specific agricultural challenges?

Contact: 

mengisti.berihu(at)rptu.de