A carregar...
Publicação
Comparação de Bases de Dados Relacionais com Bases de Dados Temporais para armazenamento de rasters
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 5.48 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
Ao longos dos anos o setor da agricultura foi severamente afetado por diversas pragas,
que arrasaram plantações inteiras e chegaram mesmo a provocar períodos de fome ex-
trema. Desta forma, com o desenvolvimento desta dissertação vai ser possível auxiliar a
formulação de modelos capazes de prever essas mesmas pragas, e assim impedir tanto
prejuízos financeiros, como problemas sociais.
Para estes modelos serem eficientes, é necessário obter informação de elevada quali-
dade. Para tal, vão ser utilizados dados obtidos através de satélites. Por vezes, a informação
obtida pelos satélites vem incompleta devido à existência de nuvens em certas regiões,
logo é necessário recorrer-se também a modelos de previsão meteorológica para preencher
a informação que falta e, desta forma, se obter os dados meteorológicos necessários.
Assim, o objetivo desta dissertação é criar um sistema de informação capaz de ar-
mazenar todos esses dados meteorológicos, de preferência, de elevada qualidade para
posteriormente servirem de base para o cálculo de modelos de previsão de pragas.
De forma a desenvolver um sistema o mais eficiente possível, para o tipo de dados
referido, nesta dissertação foram criadas duas bases de dados distintas, uma relacional e
outra de séries temporais, com o intuito de se testar qual delas seria a mais útil ao sistema
que se pretende criar. Para isso, foram realizadas inúmeras consultas, a ambas as bases
de dados, tendo em conta diversos pormenores como, área da região analisada, intervalo
temporal que se pretende analisar, agregação de dados diversos e ainda ordenação de
resultados.
Over the years, the agriculture sector has been severely affected by pests, which have devastated entire plantations and even caused periods of extreme hunger. Thus, with the development of this dissertation, it will be possible to help formulating models capable of predicting these same pests, and thus prevent both financial losses and social problems. For these models to be efficient, it is necessary to obtain high quality information. For this, data obtained from satellites will be used. Sometimes, the information obtained by the satellites is incomplete due to the existence of clouds in certain regions, therefore it is necessary to also use weather forecasting models to fill in the missing information and, in this way, to obtain the necessary meteorological data. Thus, the objective of this dissertation is to create an information system capable of storing all these meteorological data, preferably of a high quality, to later serve as the basis for calculating pest prediction models. In order to develop a system as efficient as possible, for the type of data mentioned, in this dissertation two distinct databases were created, one relational and the other a time series, to test which one would be the most useful for the system. For this, numerous queries were carried out to both databases, taking into account various details such as the area of the analyzed region, the time interval to be analyzed, aggregation of various data and even ordering of results.
Over the years, the agriculture sector has been severely affected by pests, which have devastated entire plantations and even caused periods of extreme hunger. Thus, with the development of this dissertation, it will be possible to help formulating models capable of predicting these same pests, and thus prevent both financial losses and social problems. For these models to be efficient, it is necessary to obtain high quality information. For this, data obtained from satellites will be used. Sometimes, the information obtained by the satellites is incomplete due to the existence of clouds in certain regions, therefore it is necessary to also use weather forecasting models to fill in the missing information and, in this way, to obtain the necessary meteorological data. Thus, the objective of this dissertation is to create an information system capable of storing all these meteorological data, preferably of a high quality, to later serve as the basis for calculating pest prediction models. In order to develop a system as efficient as possible, for the type of data mentioned, in this dissertation two distinct databases were created, one relational and the other a time series, to test which one would be the most useful for the system. For this, numerous queries were carried out to both databases, taking into account various details such as the area of the analyzed region, the time interval to be analyzed, aggregation of various data and even ordering of results.
Descrição
Palavras-chave
Sistema de informação dados de satélite modelos de previsão de pragas modelos de previsão meteorológica base de dados Relacional base de dados de Séries Temporais