Spatiotemporal Data of Vegetation Images for Convolutional Neural Network: Okra Case Study

Rapid development of modern IoT applications requires the deep learning model to support the accuracy of their operation. In order to acquire the high accuracy to the deep learning model, the understanding of spatial and temporal dimension of their data acquisition is necessary for each different fo...

Full description

Saved in:
Bibliographic Details
Published in2021 IEEE 12th Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON) pp. 0504 - 0508
Main Authors Azizi, Barakatullah, Waraporn, Narongrit
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.12.2021
Subjects
Adaptation models
Convolutional Neural Network
Data acquisition
Data models
Deep learning
IoT
Okra
Spatial databases
Spatiotemporal Data
Spatiotemporal phenomena
Vegetation mapping
Online AccessGet full text

Cover

More Information
Summary:Rapid development of modern IoT applications requires the deep learning model to support the accuracy of their operation. In order to acquire the high accuracy to the deep learning model, the understanding of spatial and temporal dimension of their data acquisition is necessary for each different form of the learning processes. In this paper, we categorized four forms of spatiotemporal data acquisition according to its spatial growth and time length of data acquisition. We demonstrated the inert growth and long-term spatiotemporal data acquisition form from an IoT system. We used okra vegetation as a case study of the classification on Convolutional Neural Network, CNN. Okra plant images were collected for the growth of spatial data in the half-hour periodic acquisition. Two adaptive convolutional neural networks; GoogLeNet and AlexNet were experimented for classification models. The results showed their accuracy of 99.3% and 99.8% respectively.
DOI:10.1109/UEMCON53757.2021.9666664