Métodos de Clasificación con datos obtidos mediante LiDAR
Loading...
Identifiers
Publication date
Authors
Advisors
Tutors
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
O considerable aumento da aparición e severidade de incendios forestais nas últimas decadas,
supuxo o inicio de investigacións orientadas a evitar e reducir o seu impacto e aparecemento.
Neste marco, xorde o modelo Prometheus como un capaz de resumir de forma representativa
a distribución das masas forestais. Este proceso está respaldado pola tecnoloxía LiDAR, que é
capaz de obter puntos tridimensionais cunha gran precisión e facilidade para percorrer áreas
amplas.
Neste traballo proponse a revisión e comparación de algoritmos de clasificación aplicados
ao etiquetado de masas forestais seguindo o modelo Prometheus. Deste xeito, realizarase unha
introdución á aprendizaxe estatística e de forma máis concreta á clasificación. Tomando o anterior
como base, presentaranse distintos métodos aplicables ao problema de estudo, indicando o seu
marco teórico e características. Concretamente, estudarase o método de Bayes, KNN e Modelos
Lineais Xeneralizados para a súa aplicación ao anterior problema.
Na última parte deste traballo, introduciranse as métricas para medir o seu rendemento e
analizaranse os resultados obtidos tras a súa aplicación. Deste xeito será posible coñecer a súa
eficacia sobre nubes de puntos reais e comparar o seu rendemento.
In recent decades the increase of the appearance and severity of fires led to the outset of several researches for avoiding and reducing its impact. In this situation, the Prometheus model emerge as one capable of summarizing the distribution of the vegetation in forests in a representative way. This process is supported by the LiDAR technology, which is capable of obtaining threedimensional points with high precision and easiness for covering vast areas. The purpose of this project is the revision and comparison of vegetation classification algorithms following the Prometheus model. Primarily, an introduction to statistical learning and, particularly, to classification will be made. Taking into account this, several methods will be presented, specifying its theoretical context and main characteristics. Specifically, Bayes method, KNN and Generalized Additive Models will be studied for its application to this problem. In the last part of this project, some performance metrics will be introduced for measuring and analysing the results of the classifiers. All of this will make possible to evaluate its effectiveness in real cloud points and compare its performance.
In recent decades the increase of the appearance and severity of fires led to the outset of several researches for avoiding and reducing its impact. In this situation, the Prometheus model emerge as one capable of summarizing the distribution of the vegetation in forests in a representative way. This process is supported by the LiDAR technology, which is capable of obtaining threedimensional points with high precision and easiness for covering vast areas. The purpose of this project is the revision and comparison of vegetation classification algorithms following the Prometheus model. Primarily, an introduction to statistical learning and, particularly, to classification will be made. Taking into account this, several methods will be presented, specifying its theoretical context and main characteristics. Specifically, Bayes method, KNN and Generalized Additive Models will be studied for its application to this problem. In the last part of this project, some performance metrics will be introduced for measuring and analysing the results of the classifiers. All of this will make possible to evaluate its effectiveness in real cloud points and compare its performance.
Description
Traballo Fin de Grao en Matemáticas. Curso 2021-2022
Keywords
Bibliographic citation
Relation
Has part
Has version
Is based on
Is part of
Is referenced by
Is version of
Requires
Sponsors
Rights
Atribución-NoComercial-CompartirIgual 4.0 Internacional



