Assessment of Deep Convolutional Neural Networks for Road Surface Classification

TL;DR

This paper evaluates deep convolutional neural networks for classifying road surfaces using camera images to improve vehicle control by estimating road friction, highlighting challenges in dataset creation.

Contribution

It compares two CNN models for road surface classification and discusses challenges in training data collection and dataset construction.

Findings

CNN models can classify road surfaces effectively.

Dataset quality significantly impacts classification accuracy.

Challenges in data collection hinder model training.

Abstract

When parameterizing vehicle control algorithms for stability or trajectory control, the road-tire friction coefficient is an essential model parameter when it comes to control performance. One major impact on the friction coefficient is the condition of the road surface. A camera-based, forward-looking classification of the road-surface helps enabling an early parametrization of vehicle control algorithms. In this paper, we train and compare two different Deep Convolutional Neural Network models, regarding their application for road friction estimation and describe the challenges for training the classifier in terms of available training data and the construction of suitable datasets.