Please use this identifier to cite or link to this item:
http://hdl.handle.net/10609/81449
Title: | Diseases detection in citrus fruits using convolutional neural networks |
Author: | López Laso, Fernando |
Tutor: | Isern, David |
Others: | Universitat Oberta de Catalunya Ventura, Carles |
Abstract: | The Deep Learning is becoming the most promising technique for computer vision applications. In this work, deep learning techniques are used and a convoluted neuronal network is trained to help to predict fruit diseases. As this field is extremely wide, the work is focused in the case of one of the families that combines both being a very important economic market, and having many problems of diseases that cause loss of fruit value among others problems. Then, with the network trained and validated, an application is developed for Android mobile devices to be able to use it as a tool to help decision-making of professionals in the phytosanitary sector. To develop the network, the PyTorch library has been used. This library is programmed with the Python programming language, which is almost the standard de facto in the industry. In addition, the Android application is developed with Kotlin, and to be able to use the neuronal network developed in the first part of the thesis in the Android operative system, TensorFlow Lite library has been also used. |
Keywords: | citrus diseases mobile applications Android convolutional neural network |
Document type: | info:eu-repo/semantics/bachelorThesis |
Issue Date: | Jun-2018 |
Publication license: | http://www.gnu.org/licenses/gpl.html |
Appears in Collections: | Bachelor thesis, research projects, etc. |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
flopez215VideoPresentacio0618.mp4 | 49,13 MB | MP4 | View/Open | |
flopez215TFM0618memoria.pdf | Memoria del TFG | 14,44 MB | Adobe PDF | View/Open |
flopez215TFG0618presentación.pdf | Presentación del TFG | 3,98 MB | Adobe PDF | View/Open |
Share:
Items in repository are protected by copyright, with all rights reserved, unless otherwise indicated.