Identificación de síntomas de Huanglongbing en hojas de cítricos mediante técnicas de deep learning
Artificial vision systems allow automating tasks that require trained personnel to identify relevant characteristics of certain objects. This paper describes the development of a mobile application that uses deep learning techniques to identify symptoms of Huanglongbing and nutritional deficiencies...
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Articulo |
| Lenguaje: | Español |
| Publicado: |
2019
|
| Materias: | |
| Acceso en línea: | http://sedici.unlp.edu.ar/handle/10915/135067 https://publicaciones.sadio.org.ar/index.php/EJS/article/view/144 |
| Aporte de: |
| Sumario: | Artificial vision systems allow automating tasks that require trained personnel to identify relevant characteristics of certain objects. This paper describes the development of a mobile application that uses deep learning techniques to identify symptoms of Huanglongbing and nutritional deficiencies in citrus tree leaves. The transfer learning models Inception and MobileNet using Tensorflow and Python were evaluated. A mobile application was created for Android that managed to correctly classify 89% of the sheet images of an evaluation set using the MobileNet model. The application generated will improve the identification of symptoms in leaves of citrus trees during monitoring in citrus plantations. |
|---|