Högskolan Dalarnas logga och länk till högskolans webbplats

Conventional road condition evaluation is an expensive and time-consuming task. Therefore data collection from indirect economical methods is desired by road monitoring agencies. Recently intelligent road condition monitoring has become popular. More studies have focused on automated paved road condition monitoring, and minimal research is available to date on automating gravel road condition assessment. Road roughness information gives an overall picture of the road but does not help in identifying the type of defect; therefore, it cannot be helpful in the more specific road maintenance plan. Road monitoring can be automated using data from conventional sensors, vehicles' onboard devices, and audio and video streams from cost-effective devices. This paper reviews classical and intelligent methods for road condition evaluation in general and, more specifically, reviews studies proposing automated solutions targeting gravel or unpaved roads.

Road condition evaluation is a critical part of gravel road maintenance. One of the assessed parameters is the amount of loose gravel, as this determines the driving quality and safety. Loose gravel can cause tires to slip and the driver to lose control. An expert assesses the road conditions subjectively by looking at images and notes. This method is labor-intensive and subject to error in judgment; therefore, its reliability is questionable. Road management agencies look for automated and objective measurement systems. In this study, acoustic data on gravel hitting the bottom of a car was used. The connection between the acoustics and the condition of loose gravel on gravel roads was assessed. Traditional supervised learning algorithms and convolution neural network (CNN) were applied, and their performances are compared for the classification of loose gravel acoustics. The advantage of using a pre-trained CNN is that it selects relevant features for training. In addition, pre-trained networks offer the advantage of not requiring days of training or colossal training data. In supervised learning, the accuracy of the ensemble bagged tree algorithm for gravel and non-gravel sound classification was found to be 97.5%, whereas, in the case of deep learning, pre-trained network GoogLeNet accuracy was 97.91% for classifying spectrogram images of the gravel sounds.

Road maintenance agencies subjectively assess loose gravel as one of the parameters for determininggravel road conditions. This study aims to evaluate the performance of deep learning-based pretrainednetworks in rating gravel road images according to classical methods as done by humanexperts. The dataset consists of images of gravel roads extracted from self-recorded videos andimages extracted from Google Street View. The images were labelled manually, referring to thestandard images as ground truth defined by the Road Maintenance Agency in Sweden (Trafikverket).The dataset was then partitioned in a ratio of 60:40 for training and testing. Various pre-trainedmodels for computer vision tasks, namely Resnet18, Resnet50, Alexnet, DenseNet121, DenseNet201,and VGG-16, were used in the present study. The last few layers of these models were replaced toaccommodate new image categories for our application. All the models performed well, with anaccuracy of over 92%. The results reveal that the pre-trained VGG-16 with transfer learning exhibitedthe best performance in terms of accuracy and F1-score compared to other proposed models.

This study investigates the combination of audio and image data to classify road conditions, particularly focusingon loose gravel scenarios. The dataset underwent binary categorisation, comprising audio segments capturinggravel sounds and corresponding images. Early feature fusion, utilising a pre-trained Very Deep ConvolutionalNetworks 19 (VGG19) and Principal component analysis (PCA), improved the accuracy of the Random Forestclassifier, surpassing other models in accuracy, precision, recall, and F1-score. Late fusion, involving decisionlevelprocessing with logical disjunction and conjunction gates (AND and OR) in combination with individualclassifiers for images and audio based on Densely Connected Convolutional Networks 121 (DenseNet121),demonstrated notable performance, especially with the OR gate, achieving 97 % accuracy. The late fusionmethod enhances adaptability by compensating for limitations in one modality with information from the other.Adapting maintenance based on identified road conditions minimises unnecessary environmental impact. Thismethod can help to identify loose gravel on gravel roads, substantially improving road safety and implementing aprecise maintenance strategy through a data-driven approach.