Combinação de dados de varredura a LASER aérea e terrestre: estudo de caso para uma área de floresta tropical
Data
2020-06-30
Autores
Silva, Lara dos Santos Lopes da
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Os avanços tecnológicos têm proporcionado novas técnicas de mapeamento. Uma técnica que vem sendo aplicada amplamente é a varredura a LASER, que utiliza o tempo de retorno de um pulso LASER emitido para determinar a distância entre objeto e emissor, técnica conhecida como LiDAR. A principal vantagem ao se utilizar esta tecnologia é a obtenção direta de coordenadas 3D de pontos no espaço objeto quando integrada a outros componentes. Os componentes integrados podem ser: Sistema Inercial (IMU), Sistema de Posicionamento (GNSS), e Sistema de Varredura. Em áreas de vegetação, os Sistemas de Varredura a LASER são vantajosos devido à penetração dos pulsos na folhagem e a otimização do tempo para se obter dados que possam ser utilizados para a extração das variáveis dendrométricas. Os pulsos provenientes de um Sistema de Varredura a LASER Aerotransportado (SVLA) podem não chegar ao nível do terreno, ou no caso de vegetação, há dificuldade de atingir os troncos das árvores, principalmente em regiões de florestas densas, como uma floresta tropical, por exemplo. Por outro lado, um Sistemas de Varredura a LASER Terrestre (SVLT) é capaz de registrar os elementos que não são representados usando o SVLA. O projeto em questão atuou no processamento de dados provenientes de um mapeamento aéreo de uma reserva de floresta tropical realizado por um SVLA, e dados de um mapeamento de uma porção da área da reserva através de um SVLT, de forma a produzir modelos digitais de terreno (MDT), modelos de altura de copa (CHM), e modelos resultantes da integração dos dados obtidos pelo SVLA e SVLT. A integração dos dados SVLA e SVLT foi realizada através dos referenciamentos planimétrico e altimétrico dos dados SVLT para o mesmo sistema de referência dos dados SVLA e posterior fusão de ambos os dados. Através de um modelo fusionado e de um CHM foi possível extrair algumas variáveis dendrométricas: altura das árvores (H), diâmetro de copa (DC), diâmetro à altura do peito (DAP), e área da copa (AC). Foram gerados MDTs com dados de SVLA para alturas de voo de 900 m e 2000 m, apresentando, respectivamente, acurácia de 16 cm e 17 cm, a um nível de confiança de 95%. A acurácia para a variável DAP extraída do modelo fusionado foi de aproximadamente 6 cm, enquanto a acurácia para a variável DC validada através de um modelo de regressão foi de aproximadamente 5 cm. A partir deste estudo foi possível analisar que os dados SVLA e SVLT se complementam e possibilitam a extração das informações de terreno e da vegetação para uma área de floresta tropical de maneira eficaz e precisa.
Technological advances have provided new mapping techniques. A technique that has been widely applied is the LASER scanning, which uses the return time of a LASER pulse emitted to determine the distance between object and emitter, a technique known as LiDAR. The main advantage when using this technology is to obtain 3D coordinates directly from points in the object space with other integrated components. The integrated components can be: Inertial System (IMU), Positioning System (GNSS), and Scanning System. In vegetation areas, LASER Scanning Systems are advantageous due to the penetration of the pulses in the canopy and the optimization of time to obtain data that can be used for the extraction of dendrometric variables. The pulses from an Airborne LASER Scanning System (SVLA) may not reach the ground level, or in the case of vegetation, it is difficult to reach the tree trunks, especially in regions of dense forests, such as a tropical forest, for example. On the other hand, a Terrestrial LASER Scanning System (SVLT) is capable of registering the elements that are not represented using the SVLA. The project in question acted in the processing of data from an aerial mapping of a tropical forest reserve carried out by an SVLA, and data from a mapping of a portion of the reserve area through an SVLT, in order to produce digital terrain models (MDT), canopy height models (CHM), and models resulting from the integration of data obtained by SVLA and SVLT. The integration of SVLA and SVLT data was performed through the planimetric and altimetric registration of SVLT data for the same reference system of SVLA data and subsequent merging of both data. Through a fused model and a CHM it was possible to extract some dendrometric variables: height of the trees (H), crown diameter (DC), diameter at breast height (DBH), and crown area (AC). MDTs with SVLA data were generated for flight heights of 900 m and 2000 m, presenting accuracy of 16 cm and 17 cm, respectively, at a 95% confidence level. The accuracy for the DBH variable extracted from the fused model was approximately 6 cm, while the accuracy for the DC variable validated through a regression model was approximately 5 cm. Based on this study, it was possible to analyze that the SVLA and SVLT data complement each other and make it possible to extract terrain and vegetation information for an area of tropical forest in an efficient and accurate manner.
Technological advances have provided new mapping techniques. A technique that has been widely applied is the LASER scanning, which uses the return time of a LASER pulse emitted to determine the distance between object and emitter, a technique known as LiDAR. The main advantage when using this technology is to obtain 3D coordinates directly from points in the object space with other integrated components. The integrated components can be: Inertial System (IMU), Positioning System (GNSS), and Scanning System. In vegetation areas, LASER Scanning Systems are advantageous due to the penetration of the pulses in the canopy and the optimization of time to obtain data that can be used for the extraction of dendrometric variables. The pulses from an Airborne LASER Scanning System (SVLA) may not reach the ground level, or in the case of vegetation, it is difficult to reach the tree trunks, especially in regions of dense forests, such as a tropical forest, for example. On the other hand, a Terrestrial LASER Scanning System (SVLT) is capable of registering the elements that are not represented using the SVLA. The project in question acted in the processing of data from an aerial mapping of a tropical forest reserve carried out by an SVLA, and data from a mapping of a portion of the reserve area through an SVLT, in order to produce digital terrain models (MDT), canopy height models (CHM), and models resulting from the integration of data obtained by SVLA and SVLT. The integration of SVLA and SVLT data was performed through the planimetric and altimetric registration of SVLT data for the same reference system of SVLA data and subsequent merging of both data. Through a fused model and a CHM it was possible to extract some dendrometric variables: height of the trees (H), crown diameter (DC), diameter at breast height (DBH), and crown area (AC). MDTs with SVLA data were generated for flight heights of 900 m and 2000 m, presenting accuracy of 16 cm and 17 cm, respectively, at a 95% confidence level. The accuracy for the DBH variable extracted from the fused model was approximately 6 cm, while the accuracy for the DC variable validated through a regression model was approximately 5 cm. Based on this study, it was possible to analyze that the SVLA and SVLT data complement each other and make it possible to extract terrain and vegetation information for an area of tropical forest in an efficient and accurate manner.
Descrição
Palavras-chave
LASER, SVLA, SVLT, 3D, Canopy Height Model, Nuvem de Pontos, Point Cloud