Environment Classification for Unmanned Aerial Vehicle Using Convolutional Neural Networks
Environment classification is one of the most critical tasks for Unmanned Aerial Vehicles (UAV). Since water accumulation may destabilize UAV, clouds must be detected and avoided. In a previous work presented by the authors, Superpixel Segmentation (SPS) descriptors with low computational cost are u...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2020-07-01
|
Series: | Applied Sciences |
Subjects: | |
Online Access: | https://www.mdpi.com/2076-3417/10/14/4991 |
_version_ | 1797561848787107840 |
---|---|
author | Carlos Villaseñor Alberto A. Gallegos Javier Gomez-Avila Gehová López-González Jorge D. Rios Nancy Arana-Daniel |
author_facet | Carlos Villaseñor Alberto A. Gallegos Javier Gomez-Avila Gehová López-González Jorge D. Rios Nancy Arana-Daniel |
author_sort | Carlos Villaseñor |
collection | DOAJ |
description | Environment classification is one of the most critical tasks for Unmanned Aerial Vehicles (UAV). Since water accumulation may destabilize UAV, clouds must be detected and avoided. In a previous work presented by the authors, Superpixel Segmentation (SPS) descriptors with low computational cost are used to classify ground, sky, and clouds. In this paper, an enhanced approach to classify the environment in those three classes is presented. The proposed scheme consists of a Convolutional Neural Network (CNN) trained with a dataset generated by both, an human expert and a Support Vector Machine (SVM) to capture context and precise localization. The advantage of using this approach is that the CNN classifies each pixel, instead of a cluster like in SPS, which improves the resolution of the classification, also, is less tedious for the human expert to generate a few training samples instead of the normal amount that it is required. This proposal is implemented for images obtained from video and photographic cameras mounted on a UAV facing in the same direction of the vehicle flight. Experimental results and comparison with other approaches are shown to demonstrate the effectiveness of the algorithm. |
first_indexed | 2024-03-10T18:20:25Z |
format | Article |
id | doaj.art-00ff0c1f50f3402bae054b99de30c478 |
institution | Directory Open Access Journal |
issn | 2076-3417 |
language | English |
last_indexed | 2024-03-10T18:20:25Z |
publishDate | 2020-07-01 |
publisher | MDPI AG |
record_format | Article |
series | Applied Sciences |
spelling | doaj.art-00ff0c1f50f3402bae054b99de30c4782023-11-20T07:22:24ZengMDPI AGApplied Sciences2076-34172020-07-011014499110.3390/app10144991Environment Classification for Unmanned Aerial Vehicle Using Convolutional Neural NetworksCarlos Villaseñor0Alberto A. Gallegos1Javier Gomez-Avila2Gehová López-González3Jorge D. Rios4Nancy Arana-Daniel5Department of Computer Science, University of Guadalajara, 1421 Marcelino García Barragán, Guadalajara 44430, Jalisco, MexicoDepartment of Artificial Intelligence, Hydra Technologies of Mexico, 6503 Vallarta Eje Poniente, Guadalajara 45010, Jalisco, MexicoDepartment of Computer Science, University of Guadalajara, 1421 Marcelino García Barragán, Guadalajara 44430, Jalisco, MexicoDepartment of Artificial Intelligence, Hydra Technologies of Mexico, 6503 Vallarta Eje Poniente, Guadalajara 45010, Jalisco, MexicoDepartment of Computer Science, University of Guadalajara, 1421 Marcelino García Barragán, Guadalajara 44430, Jalisco, MexicoDepartment of Computer Science, University of Guadalajara, 1421 Marcelino García Barragán, Guadalajara 44430, Jalisco, MexicoEnvironment classification is one of the most critical tasks for Unmanned Aerial Vehicles (UAV). Since water accumulation may destabilize UAV, clouds must be detected and avoided. In a previous work presented by the authors, Superpixel Segmentation (SPS) descriptors with low computational cost are used to classify ground, sky, and clouds. In this paper, an enhanced approach to classify the environment in those three classes is presented. The proposed scheme consists of a Convolutional Neural Network (CNN) trained with a dataset generated by both, an human expert and a Support Vector Machine (SVM) to capture context and precise localization. The advantage of using this approach is that the CNN classifies each pixel, instead of a cluster like in SPS, which improves the resolution of the classification, also, is less tedious for the human expert to generate a few training samples instead of the normal amount that it is required. This proposal is implemented for images obtained from video and photographic cameras mounted on a UAV facing in the same direction of the vehicle flight. Experimental results and comparison with other approaches are shown to demonstrate the effectiveness of the algorithm.https://www.mdpi.com/2076-3417/10/14/4991cloud detectionsuperpixel segmentationconvolutional neural networkssupport vector machines |
spellingShingle | Carlos Villaseñor Alberto A. Gallegos Javier Gomez-Avila Gehová López-González Jorge D. Rios Nancy Arana-Daniel Environment Classification for Unmanned Aerial Vehicle Using Convolutional Neural Networks Applied Sciences cloud detection superpixel segmentation convolutional neural networks support vector machines |
title | Environment Classification for Unmanned Aerial Vehicle Using Convolutional Neural Networks |
title_full | Environment Classification for Unmanned Aerial Vehicle Using Convolutional Neural Networks |
title_fullStr | Environment Classification for Unmanned Aerial Vehicle Using Convolutional Neural Networks |
title_full_unstemmed | Environment Classification for Unmanned Aerial Vehicle Using Convolutional Neural Networks |
title_short | Environment Classification for Unmanned Aerial Vehicle Using Convolutional Neural Networks |
title_sort | environment classification for unmanned aerial vehicle using convolutional neural networks |
topic | cloud detection superpixel segmentation convolutional neural networks support vector machines |
url | https://www.mdpi.com/2076-3417/10/14/4991 |
work_keys_str_mv | AT carlosvillasenor environmentclassificationforunmannedaerialvehicleusingconvolutionalneuralnetworks AT albertoagallegos environmentclassificationforunmannedaerialvehicleusingconvolutionalneuralnetworks AT javiergomezavila environmentclassificationforunmannedaerialvehicleusingconvolutionalneuralnetworks AT gehovalopezgonzalez environmentclassificationforunmannedaerialvehicleusingconvolutionalneuralnetworks AT jorgedrios environmentclassificationforunmannedaerialvehicleusingconvolutionalneuralnetworks AT nancyaranadaniel environmentclassificationforunmannedaerialvehicleusingconvolutionalneuralnetworks |