Titulo Estágio
Intelligent System for Fire Detection
Áreas de especialidade
Sistemas Inteligentes
Local do Estágio
Laboratório de Redes Neuronais (LARN-CISUC)
Enquadramento
Within the scope of the Project FireLoc (Where’s the Fire? - Identification, positioning and monitoring forest fires with crowdsourced data) an application is under development that will allow the report of spotted fires by any volunteer citizen provided with a smartphone, including automatic geolocation of the place of observation, imagery of what is being observed (photograph taken with the smartphone), and data enabling to georeference the observed phenomenon - namely, the orientation (automatically picked up from the smartphone) and the approximate distance between the observer and the observed event.
The geolocation of the observed events is identified using the data uploaded by all volunteers contributing. As there are errors and uncertainty associated with the data collected, the positioning of the observed phenomena will also be affected by uncertainty. Therefore, methodologies need to be applied to: 1) assign a degree of confidence to the obtained locations and extent of the observed fires; 2) classify the photographs collected by the volunteers as showing flames or smoke; 3) develop a methodology to assign to the contributed data a degree of confidence based on the data provided, the ground conditions, the volunteers profile and their history of contributions.
This internship proposal aims to identify if the photographs collected by the volunteers are showing features related to fire, namely smoke or flames, as well as their extent in the photograph, so that this data can be used for contributions validation and assess the event severity.
Objetivo
This internship involves testing supervised classification algorithms to automatically identify if the photographs uploaded by the volunteers show smoke, flames or both. This requires the set-up of training and testing sets of photographs, and to evaluate the performance of the selected classification algorithms. The run times are also relevant for this application, as the classification results need to be obtained in a short period of time.
(i) Step 1: Collection of Training and Testing Data Sets: Sets of photographs will be collected for training and testing the classification algorithms;
(ii) Step 2: Testing of Classification Algorithms and Classification Validation: Several image classification approaches will be tested to identify the best methodology to perform the photographs classification;
(iii) Step 3: Feature Analytics: Perform data analytics to choose the best classification methodology and data fusion algorithms, to generate reliable and fast classifications of the photographs.
The second main goal is to implement the classification strategy so that it can provide results in a very short time from the photograph upload, so that the outputs of the classification may be used to classify the reliability and / or relevance of the contribution. The next steps include therefore:
(iv) Step 4: Classification methodology implementation: Develop the module for the photographs classification to be integrated into the FireLoc system.
(v) Step 5: System Validation: Validate the overall classification system performance.
Plano de Trabalhos - Semestre 1
• Literature Review;
• Analysis of the data collected by FireLoc;
• Select a preliminary collection of photographs for training the image classifiers that will be tested and for classification validation;
• Preliminary tests of image classification and validation of the results;
• Writing the intermediate report.
Plano de Trabalhos - Semestre 2
• Enrich the previously selected collection of photographs for training and testing the image classifiers.
•Based on the performance of the tested classifiers, develop and implement a methodology for the fast and accurate classification of the contributed photographs;
•Writing of scientific article;
•Writing the thesis.
Condições
This work will be carried out in the Laboratory of Neural Networks (LARN) in CISUC, where there will be a regular supervision and feedback on the behalf of the supervisor and co-supervisor.
Familiarity with machine learning and data mining algorithms and software tools are essential. Participating students will acquire valuable knowledge and experience with model building and data science by mining massive datasets, which skills are currently in high demand for various technology employers due to the relevance to various applications.
Observações
Grant under FireLOC Project will be available during the second semester internship depending on the 1st semester candidate evaluation. A 3-month scholarship of 745 euros per month is foreseen for this work, renewable for another 3 months.
Logistics @Laboratory of Neural Networks (LARN)
DEI-FCTUC
Orientador
Bernardete Ribeiro e Cidalia Fonte
bribeiro@dei.uc.pt 📩