Optimal dense convolutional network model for image classification in unmanned aerial vehicles based ad hoc networks
by I.S. Hephzi Punithavathi; S. Dhanasekaran; P. Duraipandy; E. Laxmi Lydia; M. Sivaram; K. Shankar
International Journal of Ad Hoc and Ubiquitous Computing (IJAHUC), Vol. 39, No. 1/2, 2022

Abstract: Unmanned aerial vehicles (UAVs) have the potential of generating an ad hoc communication network on the fly. Aerial image classification gains more importance in the remote sensing community and several studies have been carried out in recent days. This paper presents an optimal dense convolutional network (DenseNet) with bidirectional long short term memory (Bi-LSTM) based image classification model called optimal DenseNet (ODN)-BiLSTM for UAV based adhoc networks. DenseNet model is applied as a feature extractor, where the hyperparameters of DenseNet are tuned by the use of Adagrad optimiser. Secondly, the Bi-LSTM model is applied as a classifier, which classifies the aerial images captured by UAV. Detailed performance analysis of the proposed model takes place using UCM aerial dataset and the results are investigated under several dimensions. The ODN-BiLSTM model has provided effective image classification results with the maximum accuracy of 98.14% and minimum execution time of 80s.

Online publication date: Fri, 18-Feb-2022

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Ad Hoc and Ubiquitous Computing (IJAHUC):
Login with your Inderscience username and password:

    Username:        Password:         

Forgotten your password?


Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com