Caracterização espectral da folha de citrus contaminada por HLB baseado em espectrometria de campo
Carregando...
Data
2021-05-13
Autores
Porto, Letícia Rosim
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
O Brasil, como maior produtor e exportador de suco de laranja do mundo, tem enfrentado dificuldades em combater a doença Huanglongbing (HLB) em seus pomares de citros. São mais de 197 milhões de pés de laranja na região que mais produz no país, o cinturão citrícola Paulista e do Triângulo/Sudoeste Mineiro. Atualmente, o combate à doença é realizado por meio de vistorias realizadas por técnicos e se mostra como um método oneroso, demorado e ineficiente. O uso de sensores remotos acoplados em drones pode agregar muito numa detecção precoce de árvores doentes, com um custo-benefício atrativo e de forma ágil, evitando que a doença se dissemine nos pomares. Entretanto, a dificuldade de definir comprimentos de onda eficazes na diferenciação entre HLB e outras doenças ou deficiências nutricionais comuns vem sendo um desafio para o setor. A espectrorradiometria de campo pode contribuir nesse aspecto, com um estudo de observação associado a medidas de reflectância de indivíduos infectados em diferentes níveis de comprometimento. Nesse sentido, o presente estudo visa identificar comprimentos de onda e/ou regiões do espectro eletromagnético que sejam eficientes em distinguir folhas de citros infectadas por HLB e folhas saudáveis ou com sintoma de deficiência nutricional, como zinco. Para isso, foram coletadas medidas de radiometria de diversas folhas em dois pomares de citros localizados no interior do estado de São Paulo. A partir dessas medidas, foram calculados espectros de reflectância para cada classe das amostras (assintomática, HLB, saudável, deficiência de zinco, Hamlin, Natal e Pera). Após aplicar métodos de tratamento e análise de dados espectrorradiométricos, verificou-se que a região do rededge, próximo à 700 nm, é importante na diferenciação de folhas com HLB sintomáticas e folhas saudáveis, porém não é eficiente na diferenciação entre folhas com HLB assintomáticas e folhas saudáveis. Apesar da semelhança na condição de clorose, análises de espectrorradiometria demonstraram que folhas sintomáticas de HLB e folhas com deficiência de zinco diferem espectralmente. As análises também demonstraram que as folhas infectadas das variedades de laranja Hamlin, Natal e Pera são semelhantes espectralmente. Em resumo, o estudo indicou que um mesmo sensor multiespectral pode ser utilizado na detecção de citros com HLB para diferentes variedades, porém plantas que ainda são assintomáticas possivelmente não serão diferenciadas das não infectadas por HLB.
Brazil, as the world's largest producer and exporter of orange juice, has faced difficulties in fighting the Huanglongbing disease (HLB) in its citrus groves. There are more than 197 million orange trees in the region that produces the most in the country, citrus farms in São Paulo and West-Southwest Minas Gerais. Currently, the fight against the disease is carried out through inspections carried out by technicians and is shown to be an expensive, time-consuming and inefficient method. The use of remote sensors coupled to drones can improve early detection of sick trees, with an attractive cost-benefit and in an agile way, preventing the disease from spreading in orchards. However, the difficulty of defining effective wavelengths to differentiate between HLB and other common diseases or nutritional deficiencies has been a challenge for the industry. Field spectroradiometric measurements can contribute in this respect, with an observational study associated with measures of reflectance of infected individuals at different levels of impairment. In this sense, the present study aims to identify wavelengths and/or regions of the electromagnetic spectrum that are efficient in distinguishing citrus leaves infected with HLB and healthy leaves or those with a symptom of nutritional deficiency, such as zinc. For this purpose, radiometric measurements were collected from several leaves in two citrus orchards located in the interior of the state of São Paulo. From these measurements, reflectance spectra were calculated for each class of samples (asymptomatic, HLB, healthy, zinc deficiency, Hamlin, Natal and Pera). After applying treatment methods and analysis of spectroradiometric data, it was found that the rededge region, close to 700 nm, is important in differentiating symptomatic HLB leaves from healthy leaves, but it is not efficient in differentiating between HLB leaves asymptomatic and healthy leaves. In addition, despite the similarity in the chlorosis condition, spectroradiometric analyzes demonstrated that symptomatic HLB leaves and zincdeficient leaves differ spectrally. The analyzes also showed that the infected leaves of the Hamlin, Natal and Pera orange varieties are spectrally similar. In summary, this study indicated that the same multispectral sensor can be used to detect citrus with HLB for different varieties, but plants that are still asymptomatic will possibly not be differentiated from those not infected with HLB.
Brazil, as the world's largest producer and exporter of orange juice, has faced difficulties in fighting the Huanglongbing disease (HLB) in its citrus groves. There are more than 197 million orange trees in the region that produces the most in the country, citrus farms in São Paulo and West-Southwest Minas Gerais. Currently, the fight against the disease is carried out through inspections carried out by technicians and is shown to be an expensive, time-consuming and inefficient method. The use of remote sensors coupled to drones can improve early detection of sick trees, with an attractive cost-benefit and in an agile way, preventing the disease from spreading in orchards. However, the difficulty of defining effective wavelengths to differentiate between HLB and other common diseases or nutritional deficiencies has been a challenge for the industry. Field spectroradiometric measurements can contribute in this respect, with an observational study associated with measures of reflectance of infected individuals at different levels of impairment. In this sense, the present study aims to identify wavelengths and/or regions of the electromagnetic spectrum that are efficient in distinguishing citrus leaves infected with HLB and healthy leaves or those with a symptom of nutritional deficiency, such as zinc. For this purpose, radiometric measurements were collected from several leaves in two citrus orchards located in the interior of the state of São Paulo. From these measurements, reflectance spectra were calculated for each class of samples (asymptomatic, HLB, healthy, zinc deficiency, Hamlin, Natal and Pera). After applying treatment methods and analysis of spectroradiometric data, it was found that the rededge region, close to 700 nm, is important in differentiating symptomatic HLB leaves from healthy leaves, but it is not efficient in differentiating between HLB leaves asymptomatic and healthy leaves. In addition, despite the similarity in the chlorosis condition, spectroradiometric analyzes demonstrated that symptomatic HLB leaves and zincdeficient leaves differ spectrally. The analyzes also showed that the infected leaves of the Hamlin, Natal and Pera orange varieties are spectrally similar. In summary, this study indicated that the same multispectral sensor can be used to detect citrus with HLB for different varieties, but plants that are still asymptomatic will possibly not be differentiated from those not infected with HLB.
Descrição
Palavras-chave
Análise espectral, Greening, Sensoriamento remoto, Agricultura de precisão, Dados hiperespectrais, Huanglongbing (HLB), Spectral analysis, Remote sensing, Precision agriculture, Hyperspectral data