Convolutional Neural Networks for Detecting Bridge Crossing Events With Ground-Based Interferometric Radar Data

Abstract

Abstract. This study focuses on detecting vehicle crossings (events) with ground-based interferometric radar (GBR) time series data recorded at bridges in the course of critical infrastructure monitoring. To address the challenging event detection and time series classification task, we rely on a deep learning (DL) architecture. The GBR-displacement data originates from real-world measurements at two German bridges under normal traffic conditions. As preprocessing, we only apply a low-pass filter. We develop and evaluate a one-dimensional convolutional neural network (CNN) to achieve a solely data-driven event detection. As a baseline machine learning approach, we use a Random Forest (RF) with a selected feature-based input. Both models’ performance is evaluated on two datasets by focusing on identifying events and pure bridge oscillations. Generally, the event classification results are promising, and the CNN outperforms the RF with an overall accuracy of 94.7% on the test subset. By relying on an entirely unknown second dataset, we focus on the models’ performances regarding the distinction between events and decays. On this dataset, the CNN meets this challenge successfully, while the feature-based RF classifies the majority of non-event decays as events. To sum up, the presented results reveal the potential of a data-driven DL approach concerning the detection of bridge crossing events in GBR-based displacement time series data. Based on such an event detection, a prospective assessment of bridge conditions seems feasible as an extension to previous structural health monitoring approaches.

Real-Time Vehicle Detection for Traffic Monitoring: A Deep Learning Approach

“…As in (1) , we process individual frames for video inputs. • Step 2 -Image Pre-processing: We apply techniques like resizing to 416x416 pixels as used in (2) , normalization to scale pixel values between 0-1 (3) , and Gaussian blurring to reduce noise. (4) This improves consistency across varying conditions.…”

“…(1,2) However, these methods using handcrafted features struggle with realworld variations in lighting, weather, occlusion and clutter. (3,4) Recently, deep learning has catalyzed immense progress in object detection across domains. (5,6) Convolutional neural networks (CNNs) now dominate, significantly outperforming prior techniques.…”

“…R-CNN introduced a region-based two-stage CNN approach using selective search and SVMs. (3) Though highly accurate, R-CNN was extremely slow. Subsequent works like Fast R-CNN (3) and Faster R-CNN (4) optimized it through shared convolutions and region proposal networks.…”

“…(3) Though highly accurate, R-CNN was extremely slow. Subsequent works like Fast R-CNN (3) and Faster R-CNN (4) optimized it through shared convolutions and region proposal networks. Single shot detectors like YOLO (5) and SSD (6) enabled real-time processing by eliminating region proposals.…”

Fatigue Analysis on Four Months of Data on a Steel Railway Bridge: Event Detection and Train Features’ Effect on Fatigue Damage

“…In a study of Arnold et al [3], they detected and classified vehicle crossings (events) on bridges with ground-based interferometric radar (GBR) data and machine learning (ML) approaches. In another work of the same authors [4], they used a deep learning (DL) method and developed a one-dimensional convolutional neural network (CNN) to achieve a solely data-driven event detection. In this contribution, another technique is suggested to detect the events by defining a threshold and comparing sliding windows during the strain measurement.…”