Slobodan Djukanović

Nikola Bulatovic, Slobodan Djukanovic

The paper addresses acoustic vehicle detection and speed estimation from single sensor measurements. We predict the vehicle's pass-by instant by minimizing clipped vehicle-to-microphone distance, which is predicted from the mel-spectrogram of input audio, in a supervised learning approach. In addition, mel-spectrogram-based features are used directly for vehicle speed estimation, without introducing any intermediate features. The results show that the proposed features can be used for accurate vehicle detection and speed estimation, with an average error of 7.87 km/h. If we formulate speed estimation as a classification problem, with a 10 km/h discretization interval, the proposed method attains the average accuracy of 48.7% for correct class prediction and 91.0% when an offset of one class is allowed. The proposed method is evaluated on a dataset of 304 urban-environment on-field recordings of ten different vehicles.

Nikola Bulatovic, Slobodan Djukanovic

We consider improving the performance of a recently proposed sound-based vehicle speed estimation method. In the original method, an intermediate feature, referred to as the modified attenuation (MA), has been proposed for both vehicle detection and speed estimation. The MA feature maximizes at the instant of the vehicle's closest point of approach, which represents a training label extracted from video recording of the vehicle's pass by. In this paper, we show that the original labeling approach is suboptimal and propose a method for label correction. The method is tested on the VS10 dataset, which contains 304 audio-video recordings of ten different vehicles. The results show that the proposed label correction method reduces average speed estimation error from 7.39 km/h to 6.92 km/h. If the speed is discretized into 10 km/h classes, the accuracy of correct class prediction is improved from 53.2% to 53.8%, whereas when tolerance of one class offset is allowed, accuracy is improved from 93.4% to 94.3%.

Slobodan Djukanović, Nikola Bulatović, Ivana Čavor

Accurate speed estimation of road vehicles is important for several reasons. One is speed limit enforcement, which represents a crucial tool in decreasing traffic accidents and fatalities. Compared with other research areas and domains, the number of available datasets for vehicle speed estimation is still very limited. We present a dataset of on-road audio-video recordings of single vehicles passing by a camera at known speeds, maintained stable by the on-board cruise control. The dataset contains thirteen vehicles, selected to be as diverse as possible in terms of manufacturer, production year, engine type, power and transmission, resulting in a total of $ 400 $ annotated audio-video recordings. The dataset is fully available and intended as a public benchmark to facilitate research in audio-video vehicle speed estimation. In addition to the dataset, we propose a cross-validation strategy which can be used in a machine learning model for vehicle speed estimation. Two approaches to training-validation split of the dataset are proposed.

Slobodan Djukanović, Yash Patel, Jiři Matas et al.

This paper addresses acoustic vehicle counting using one-channel audio. We predict the pass-by instants of vehicles from local minima of clipped vehicle-to-microphone distance. This distance is predicted from audio using a two-stage (coarse-fine) regression, with both stages realised via neural networks (NNs). Experiments show that the NN-based distance regression outperforms by far the previously proposed support vector regression. The $ 95\% $ confidence interval for the mean of vehicle counting error is within $[0.28\%, -0.55\%]$. Besides the minima-based counting, we propose a deep learning counting that operates on the predicted distance without detecting local minima. Although outperformed in accuracy by the former approach, deep counting has a significant advantage in that it does not depend on minima detection parameters. Results also show that removing low frequencies in features improves the counting performance.