Air Quality Measurement Based on Double-Channel Convolutional Neural Network Ensemble Learning

Environmental air quality affects people's lives and has a profound guiding significance for the development of social activities. At present, environmental air quality measurement mainly adopts the method that setting air quality detectors at specific monitoring points in cities with fix-time...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE access 2019, Vol.7, p.145067-145081
Hauptverfasser: Wang, Zhenyu, Zheng, Wei, Song, Chunfeng, Zhang, Zhaoxiang, Lian, Jie, Yue, Shaolong, Ji, Senrong
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Environmental air quality affects people's lives and has a profound guiding significance for the development of social activities. At present, environmental air quality measurement mainly adopts the method that setting air quality detectors at specific monitoring points in cities with fix-time sampling and slow analysis, which is severely restricted by the time and location. To address this problem, recognizing air quality with mobile cameras is a natural idea. Some air quality measurement algorithms related to deep learning mostly adopt a single convolutional neural network to directly train the whole image, which will ignore the difference of each part of the image. In this paper, in order to learn the combined feature extracted from different parts of the environmental image, we propose the double-channel weighted convolutional network (DCWCN) ensemble learning algorithm. This mainly includes two aspects: ensemble learning of DCWCN and self-learning weighted feature fusion. Firstly, we construct a double-channel convolutional neural network, which uses each channel to train different parts of the environment images for feature extraction. Secondly, we propose a feature weights self-learning method, which weights and concatenates the extracted feature vectors to measure the air quality. Moreover, we build an environmental image dataset with random sampling time and locations to evaluate our method. The experiments show that our method can achieve over 87% accuracy on the newly built dataset. At the same time, through comparative experiments, we proved that the proposed method achieves considerable improvement in terms of performance compared with existing CNN based methods.
ISSN:2169-3536
2169-3536
DOI:10.1109/ACCESS.2019.2945805