An accurate plant & disease recognition system
The main objectives are as follows:
- Plant recognition
- Disease recognition
- Research objective
- Data analysis
- Quality control
Project’s primary goal is to establish a plant recognition system capable of accurately identifying different plant species and disease system recognition for the detection and diagnosis of known plant diseases, thereby facilitating timely and effective intervention.
Building upon the implementation and functioning of these systems described, in the future this project may be a tool to further investigate and expand knowledge in the context of unknown disease, by improving disease identification and management strategies. Rigorous data analysis will be pursued to identify patterns in disease prevalence across different geographical areas. This analysis will be crucial for determining whether these diseases are region-specific or migratory, and studying the relationship between the diseases, the local vegetation, and other present diseases.
Responsible organisation
A key element of TAAL’s long-term strategy involves the implementation of a stringent quality control methodology. This methodology will be designed to ensure consistent monitoring of crop health, thereby maintaining the highest quality standards throughout the agri-food supply chain.
Challenges and how they will be addressed
Possible challenges:
- Recognition Effectiveness: Ensuring the system accurately identifies plant species and detects diseases.
- Data Visualization: Developing interfaces that allow for easy interpretation of complex data and delivering significant insights to users promptly.
- User Accessibility: Making the system affordable to a wide range of personnel working in this field.
How they will be addressed:
The effectiveness of the recognition system will be ensured by combining high-quality initial image datasets with the most suitable AI algorithms and agronomic heuristics. We will commence with a focused and small model encompassing a select variety of plants and diseases, aiming to construct an efficient system. This foundational system will then be expanded to include additional plant species and diseases.
Tech components and data
Tech Components:
- Drones: Unmanned Aerial Vehicles (UAVs), equipped with navigation systems, capable of stable flight and precise maneuvering in various weather conditions.
- Sensors and Cameras: High-resolution cameras for capturing detailed images, possibly equipped with multispectral and hyperspectral imaging capabilities to capture data beyond the visible spectrum.
- Communication Systems: Systems for real-time data transmission between the drones and the cloud infrastructure. This system merges Internet of Things (IoT) principles with RESTful APIs that can facilitate fast information transfer.
- Cloud storage: Secure storage solutions for the large volumes of data collected. The database architecture will be of a non-relational nature, which allows the flexibility to store and organize multimedia content, such as images and documents, as traditional data types.
- Software and AI Algorithms: Advanced deep learning algorithms and CV libraries, such as OpenCV or TensorFlow, target, acquire, process, and analyze the imagery data. Agronomic heuristics specific to plant disease management will be incorporated into the system.
- Web User Interface: A user-friendly website that allows operators to monitor the system, visualize data, and receive alerts and reports.
Data:
- Imagery Data: High-resolution images of plants
- Geospatial Data: GPS and GIS data for mapping the locations of the plants and tracking the drones’ flight paths.
- Plant Species Databases: databases containing information on various plant species and their characteristics.
- Plant Disease Databases: Information on known plant diseases, and their symptoms, which is vital for training the AI to detect and identify issues.
Expected outcomes
1
Efficient and accurate plant recognition
2
Real-time plant health monitoring and assessment.
3
Enhanced agricultural productivity
Initially, the model will be limited, encompassing several plant species and diseases. This foundational dataset is essential for ensuring the efficiency and accuracy of our plant recognition capabilities at the early stages. Although starting on a smaller scale, the system will be engineered to be scalable. The intention is to have a robust system that will integrate additional plant species and diseases over time, thereby expanding SENSOR 2.0’s monitoring capacity. The system will be designed to offer real-time insights into the health status of plants, providing immediate feedback that is critical for the timely intervention and management of plant health issues.
Through these phases, we expect to see a tangible improvement in agricultural productivity. As the system grows more comprehensive, it will become an increasingly powerful tool for farmers and agriculturalists, promoting healthier crops and more efficient farming practices.
About TAAL
| Full name | TAAL s.r.l. |
| Country | Italy |
| Website | https://taal.it/en/ |
| Contact Persons | Massimiliano Giurelli |
| Contact’s email |
TAAL provides services, solutions, innovative products, business consultancy, and support for production processes.
Leveraging cutting-edge technologies, from Artificial Intelligence to Virtual and Augmented Reality, through to IoT solutions, we create tailor-made solutions with highly specialized know-how.
Stay Connected
We appreciate your feedback.
Join our newsletter
Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or Research Executive Agency. Neither the European Union nor the granting authority can be held responsible for them.