This building block represents the tangible, physical, and digital components that enable the solution. Its purpose is to provide the necessary tools, systems, and underlying support structures for data collection, processing, analysis, as well as for executing automated and human-led responses.

How it works in PyroSense:  

1. Sensor Network Deployment involves strategically placing IoT environmental sensors with long-lasting power (up to 10 years) and reliable communication methods (MQTT, LoRa, Sigfox, NBIoT) for data transmission from remote locations.

2. Drone Systems and 5G: This involves drones equipped with high-resolution cameras and Vision AI, relying on 5G networks to enable real-time video streaming.

3. Centralized Data Platform and Cloud Infrastructure for extensive data analysis and computational power for data fusion, and AI models.

4. Artificial Intelligence (AI) enhances system intelligence through algorithms for anomaly detection in sensor data, and in image/video analysis for smoke and flame detection.

5. Web Application and User Interface (UI) for human interaction, to access real-time information, view maps, receive alerts, etc.  

6. Wearable Technology for Firefighters: This includes biometric sensors, and software for data analysis to enhance responder safety through risk assessment and alerts.

This system translates PyroSense's advanced data into actionable intelligence, ensuring that the right people— from first responders to community members— are informed and prepared at the right time. Its primary goal is to share critical information about fire risks, incidents, and response efforts promptly and accurately. Effective communication is crucial in emergency management, as it facilitates coordination, reduces panic, and supports informed decision-making.

How it works in PyroSense:

1. Real-time Incident Communication: When an alarm is triggered by PyroSense, the web platform displays precise fire coordinates and the geolocation of all deployed resources on a comprehensive GIS map, facilitating better planning and coordination.
2. Emergency services receive detailed operational data, while communities are given simplified alerts and areas to avoid.
3. PyroSense integrates with automated alert systems like SMS and email, ensuring critical information reaches stakeholders even when they aren't monitoring the web app, maximizing reach and redundancy.

This initiative uses PyroSense data to drive public fire prevention campaigns, engage communities in understanding fire risks, and foster preparedness plans, while also enhancing educational materials on wildfire mitigation. 

Building local and regional expertise in DNA barcoding and metabarcoding is vital for sustainable biodiversity conservation. Supported by BBI-CBD funding, our training programs target conservation practitioners from Lebanon, Tunisia, Côte d'Ivoire, and Jordan, including those without prior molecular biology experience. These hands-on workshops cover sample collection, laboratory techniques, data analysis, and interpretation, empowering participants to independently apply molecular tools in their contexts. Capacity building democratizes access to innovative technologies, fosters regional collaboration, and ensures continuity beyond the project lifecycle.

Meaningful engagement with Lebanese Ministries of Environment and Agriculture, local NGOs, practitioners, and communities ensured that scientific insights informed policy and restoration practices. By communicating findings clearly and collaboratively, we helped integrate molecular data into the National Biodiversity Action Plan. Awareness campaigns targeted schools, universities, farmers, and land managers, raising understanding of the ecological roles animals play in forest regeneration. This knowledge mobilization builds local ownership, promotes evidence-based decision-making, and bridges science with societal needs for long-term ecosystem resilience.

Partnerships with the Smithsonian Institution, iBOL, and Saint Joseph University have been central to our project’s success. The Smithsonian provided advanced expertise in metabarcoding methodology and quality assurance, enabling rigorous application of DNA analysis. iBOL supports the expansion of barcoding efforts, particularly for insects, linking our regional data to global biodiversity initiatives. Saint Joseph University leads research implementation and capacity building, ensuring regional ownership and continuity. These collaborations combine global knowledge with local ecological and institutional context, enabling innovation and scalability.

The creation of a comprehensive, open-access DNA reference library of native plant and animal species was foundational to our solution. Recognizing that global databases lacked coverage for many Eastern Mediterranean species, we built the first Lebanese library encompassing plants, mammals, and now expanding to insects, birds, and fungi. This reference database improves the accuracy of DNA sequence matching and enables precise identification of species present in environmental samples. It also fills a critical regional data gap and facilitates ecological studies, biodiversity monitoring, and conservation planning. By publishing the library openly, we promote transparency, collaboration, and the potential for adaptation in similar biodiversity hotspots.

This building block provides hands-on training for researchers and conservationists in Benin and South Africa (offline) and globally (online via Zoom) on using Declas. The sessions cover:

• Software use: Uploading data, interpreting AI-generated results, and integrating findings into conservation strategies.
• Conservation advocacy: Raising awareness on vulture decline and AI’s role in scalable monitoring.

Trainees will learn to deploy Declas in field surveys, reducing reliance on manual counts while improving data accuracy. The hybrid approach ensures broad accessibility, empowering local teams with cost-effective technology.

Apart from common comprehension of wildlife, citizens can also get involved in the following advanced activities and become “citizen scientists”.

1. Verify the existence of wild animals through the AI recognition tool “Species Eye”;

2. Estimate the number of wild animals manually;

3. Select the species of wild animal;

4. Figure out the name of the wild animal through the AI recognition tool “Species Eye”;

5. Ensure the accuracy and consistency of data by marking the geographic location.

This mini program, through scanning or searching, provides the access for every citizen to infrared images of wildlife in his/her geographic location, which allows them to participate in the sharing of nature with wildlife in a simpler and more intuitive manner through this application.

The Mini Program, with AI intelligent recognition as the underlying technology and endangered species as the core targets for recognition, trains a large biodiversity recognition model capable of monitoring such systems as mountains, rivers, forests, fields, lakes, grasslands and sandy lands. This recognition model features a powerful “engine” – Species Eye. As the fundamental and general model of “Wildlife Friends”, it is primarily characterized by strong learning capability, that is, multimodal recognition based on no or a few samples. It indicates that it can rapidly recognize the location and species information of multiple wild animals based on a few samples. For instance, it may take thousands of photos to be captured and days’ training for a traditional model to recognize a new species, but Species Eye can complete such mission quickly through a small number of photos, as well as trainings and iterations. In addition, it boasts a high extent of tolerance. It is no longer limited to the training and prediction for specific species with a powerful capability of open vocabulary identification and zero-sample recognition, and is able to accurately recognize and locate species without trained data. For instance, traditional models can merely recognize species based on trained data, such as tiger and antelope; while this new model can discriminate snow leopard and fox at the same time even without trained data of those two wild animals before. Another advantage of “Species Eye” is lower costs. Common AI models rely heavily on high-performance accelerator cards, which cause a high cost in hardware environment and its maintenance.