An automatic image recognition system for winter road surface condition classification

In winter, it is critical for cold regions to have a full understanding of the spatial variation of road surface conditions such that hot spots (e.g., black ice) can be identified for an effective mobilization of winter road maintenance operations. Acknowledging the limitations in present study, this paper proposes a systematic framework to estimate road surface temperature (RST) via the geographic information system (GIS). The proposed method uses a robust regression kriging method to take account for various geographical factors that may affect the variation of RST. A case study of highway segments in Alberta, Canada is used to demonstrate the feasibility and applicability of the method proposed herein. The findings of this study suggest that the geostatistical modelling framework proposed in this paper can accurately estimate RST with help of various covariates included in the model and further promote the possibility of continuous monitoring and visualization of road surface conditions.

Download Full-text

Connected Vehicle Solution for Winter Road Surface Condition Monitoring

Transportation Research Record Journal of the Transportation Research Board ◽

10.3141/2551-08 ◽

2016 ◽

Vol 2551 (1) ◽

pp. 62-72 ◽

Cited By ~ 9

Author(s):

Michael A. Linton ◽

Liping Fu

Keyword(s):

Condition Monitoring ◽

Surface Condition ◽

Road Surface ◽

Connected Vehicle ◽

Winter Road

Download Full-text

2C3-5　A Study of Winter Road Surface Condition Presentation by Steering Vibration

The Japanese Journal of Ergonomics ◽

10.5100/jje.55.2c3-5 ◽

2019 ◽

Vol 55 (Supplement) ◽

pp. 2C3-5-2C3-5

Author(s):

Kazuya ITOH ◽

Ryoma ITOH

Keyword(s):

Surface Condition ◽

Road Surface ◽

Winter Road

Download Full-text

Interactive and automatic image recognition system

Machine Vision and Applications ◽

10.1007/bf01212312 ◽

1988 ◽

Vol 1 (1) ◽

pp. 59-69 ◽

Cited By ~ 1

Author(s):

Fumiaki Tomita

Keyword(s):

Image Recognition ◽

Recognition System ◽

Automatic Image Recognition

Download Full-text

Real-time Winter Road Surface Condition Monitoring Using an Improved Residual CNN

Canadian Journal of Civil Engineering ◽

10.1139/cjce-2019-0367 ◽

2020 ◽

Cited By ~ 1

Author(s):

Guangyuan Pan ◽

Matthew Muresan ◽

Ruifan Yu ◽

Liping Fu

Keyword(s):

Real Time ◽

Surface Condition ◽

Model Performance ◽

Road Surface ◽

Specific Domain ◽

Ice Coverage ◽

Calibration Methods ◽

Winter Road ◽

The Relationship ◽

Fine Tune

This paper proposes a real-time winter road surface condition (RSC) monitoring solution that automatically generates descriptive RSC information in terms of snow/ice coverage by using images from fixed traffic/weather cameras. Several state-of-the-art pre-trained deep neural networks are customized and fine-tuned to address a specific domain, classifying the amount of snow coverage on a road surface. A thorough evaluation is conducted to identify and select the best model. This evaluation uses an extensive set of experiments to test the accuracy and generalization of each model and uses transfer-learning to fine-tune each of the pre-trained models on independent images from different traffic/weather cameras. The transferability of each model, the relationship between model performance and data size, and the system settings of each model are then examined. Lastly, three online weight calibration methods are proposed to automatically update the model in new environments. The result shows that re-training the model using images from a mixed set of cameras has the most promising results.

Download Full-text