Titulo Estágio
Using geographic and landscape data for forest fire early detection
Local do Estágio
DEI, CISUC
Enquadramento
The SenForFire project, part of the Interreg Sudoe 2021-2027 program, will demonstrate the viability of low-cost wireless sensor networks (WSNs) for their application in forest fire early warning and surveillance systems. Currently, meteorological data and satellite images (remote sensing) with low spatial and temporal resolution and low reliability (high rate of false positives) are used to assess the risk and detect forest fires in the Sudoe region. Additionally, meteorological and remote sensing equipment is costly. WSNs utilize electronic modules with sensors that measure meteorological and environmental parameters in real time, which are crucial for assessing forest fire risks. These modules are equipped with computing and wireless communication capabilities to interact among themselves and with the cloud. WSNs adapt to the characteristics of risk areas, are easy to deploy, and scalable. They serve as a proximity tool that complements remote sensing and facilitates co-responsible risk management by municipalities, local communities, and inhabitants of risk areas.
The project will conduct pilot activities in various Sudoe areas with different climates, orographies, and vegetation to achieve prevention and/or early detection objectives. It will also develop an action plan for the adoption of WSNs in municipalities for meteorological and environmental monitoring and an action plan for the training of professionals in WSNs for environmental risk management. WSNs integrate multiple technologies with varying levels of development in the Sudoe countries, making transnational cooperation essential for the realization of WSNs, pilot activities, and action plans. The achievements will benefit municipalities (innovative technology for effective environmental risk management), landowners and users (protection of crops, pastures, forests, and livestock), SMEs (products and services with high added value and qualified professionals), the young population (quality work and entrepreneurship), and the general population (protection of health, property, infrastructure, natural, and cultural heritage) in rural Sudoe.
Sensors such as thermal sensors, smoke detectors, optical sensors, and others can provide various types of data necessary for intelligent algorithms that detect the presence of fire. One of the challenges of SenForFire is to create an infrastructure for the Extract-Transform-Load (ETL) process: 1) ingest data from sensors, 2) filter, transform, and aggregate the data, and 3) load data into a database. Once data is loaded, the next challenge is to make it accessible flexibly, allowing researchers, firemen, rescuers, public officers, and algorithms to explore the data in the exact formats they need.
Objetivo
During this internship, the student will conceive and develop approaches to include multimodal data into the fire prevention process, namely geographic data that allows to include the landscape as a determining factor in the early detection of fire using intelligent models. The internship includes the following goals:
- Study the state of the art
- Study the available multimodal data for fire detection
- Study the available frameworks for intelligent model development
- Define, implement, and fine tune the learning architecture
- Propose and deploy test setup
Plano de Trabalhos - Semestre 1
- Literature review
- Identification and annotation of datasets for model development
- Identification and study of available frameworks for model development
- Define the architecture of the system
- Start implementing the proposed approach
- Write intermediate report
Plano de Trabalhos - Semestre 2
- Implement the proposed solution and fine tune models
- Test and evaluate performance
- Write final report
Condições
This work should take place in the context of a research project funded by FCT in a CISUC lab. There is the possibility of a 6-month scholarship of 990.98 euros per month.
Orientador
Catarina Silva, Cidália Fonte, Alberto Cardoso
catarina@dei.uc.pt 📩