Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Apr 21;22(9):3193.
doi: 10.3390/s22093193.

Weather Classification by Utilizing Synthetic Data

Saad Minhas  1 , Zeba Khanam  1 , Shoaib Ehsan  1 , Klaus McDonald-Maier  1 , Aura Hernández-Sabaté  2   3
Affiliations

Weather Classification by Utilizing Synthetic Data

Saad Minhas et al. Sensors (Basel). .

Abstract

Weather prediction from real-world images can be termed a complex task when targeting classification using neural networks. Moreover, the number of images throughout the available datasets can contain a huge amount of variance when comparing locations with the weather those images are representing. In this article, the capabilities of a custom built driver simulator are explored specifically to simulate a wide range of weather conditions. Moreover, the performance of a new synthetic dataset generated by the above simulator is also assessed. The results indicate that the use of synthetic datasets in conjunction with real-world datasets can increase the training efficiency of the CNNs by as much as 74%. The article paves a way forward to tackle the persistent problem of bias in vision-based datasets.

Keywords: advanced driver assistance systems; autonomous car; computer vision; dataset; deep learning; intelligent transportation systems; synthetic data; weather classification.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Figure 1
Figure 1
Simulator.
Figure 2
Figure 2
Virtual Interior.
Figure 3
Figure 3
Virtual Car.
Figure 4
Figure 4
Proposed Environment.
Figure 5
Figure 5
Synthetic Weather Dataset.
Figure 6
Figure 6
BDD (Berkeley Deep Dive) Dataset.
Figure 7
Figure 7
Pipeline: Step 1: Load the pretrained network, Step 2: Unfreeze the classification layers and add a softmax layer (4,1), Step 3: Train the weights of the classification layers with the synthetic dataset, Step 4: Test the network accuracy with a real time test dataset.
Figure 8
Figure 8
Accuracy variation over each epoch for (a) AlexNet, (b) VGG, and (c) GoogleLeNet models.
Figure 9
Figure 9
Accuracy variation over each epoch for Residual Networks (a) ResNet50 and (b) ResNet101.
See this image and copyright information in PMC

References

    1. Position Paper on Road Worthiness. [(accessed on 26 January 2022)]. Available online: https://knowledge-base.connectedautomateddriving.eu/wp-content/uploads/2....
    1. Cools M., Moons E., Wets G. Assessing the impact of weather on traffic intensity. Weather Clim. Soc. 2010;2:60–68. doi: 10.1175/2009WCAS1014.1. - DOI
    1. Achari V.P.S., Khanam Z., Singh A.K., Jindal A., Prakash A., Kumar N. I 2 UTS: An IoT based Intelligent Urban Traffic System; Proceedings of the 2021 IEEE 22nd International Conference on High Performance Switching and Routing (HPSR); Paris, France. 7–10 June 2021; pp. 1–6.
    1. Kilpeläinen M., Summala H. Effects of weather and weather forecasts on driver behaviour. Transp. Res. Part F Traffic Psychol. Behav. 2007;10:288–299. doi: 10.1016/j.trf.2006.11.002. - DOI
    1. Lu C., Lin D., Jia J., Tang C.K. Two-class weather classification; Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition; Columbus, OH, USA. 23–28 June 2014; pp. 3718–3725.

LinkOut - more resources