1 June 2020 in Automation in Construction
27 June 2022 in IEEE Transactions on Intelligent Transportation Systems
Recent technological advancements in computer vision algorithms and data acquisition devices have greatly facilitated the research and applications of deep learning-based traffic object recognition from Light Detection and Ranging (LiDAR) data. The majority of existing methodologies applied deep learning (DL)-based techniques, especially Convolutional Neural Networks (CNNs), for vehicle detection and tracking on autonomous driving datasets. Nevertheless, fewer studies were focused on DL-based vehicle detection using roadside LiDAR data, partially due to the lack of publicly available roadside LiDAR datasets for network training and testing. This paper develops a novel framework based on CNNs and LiDAR data for automated vehicle detection. It leverages the domain knowledge of CNNs trained on large-scale autonomous driving datasets for vehicle detection from roadside LiDAR data. In the experimental study, roadside LiDAR data were collected at a road intersection in Reno, Nevada, U.S. Meanwhile, a CNN architecture was proposed to detect vehicles from LiDAR data through 3D bounding boxes. The proposed CNN was modified from the established PointPillars network by adding dense connections to the convolutional layers to achieve more comprehensive feature extraction. Three CNNs, including the proposed CNN, PointPillars, and YOLOv4, were trained and tested on PandaSet, a publicly available large-scale autonomous driving LiDAR dataset. Subsequently, the trained CNNs were reused for vehicle detection from the captured roadside LiDAR data. The experimental results demonstrated that the proposed CNN outperformed the others in the testing metrics. All three networks showed good performance on vehicle detection from the captured roadside LiDAR data.
1 March 2020 in Journal of Computing in Civil Engineering
Image-based automated crack detection techniques have been developed and applied extensively during the past decade. However, issues existing in this type of methodology have not yet been fully addressed. Early studies adopted image intensity data for crack detection, assuming the crack pixels are darker than their surrounding pixels. However, when the intensity image suffers from shadows, blemishes, or poor contrast between cracks and surrounding surface, this assumption is compromised. Another issue exists in the image preprocessing stage. Traditional methods, such as median filtering and surface fitting, are used for background correction and crack enhancement purposes. Because most of these preprocessing techniques require subjective parameters such as kernel size of the filter, their performance can be influenced by different images and operating personnel. Moreover, conventional crack detection methods such as edge detection usually do not provide an enclosed crack boundary. This leads to difficulties in extracting individual cracks and estimating the associated cracking properties. To address these issues, an automated crack-detection method that utilizes frequency domain filtering and contouring analysis on three dimensional (3D) laser image range data is proposed. In the image preprocessing stage, frequency domain filtering techniques are applied to remove surface variations (or oscillations), image noises, and grooved patterns. In the proposed methodology, the relationship between the width of a crack and the cutoff frequency of the image filters is derived. The filter parameters are designed based on the physical characteristics (e.g., crack width, groove spacing, and orientation), thus having less subjectivity. A systematic crack detection methodology based on contouring analysis and a set of logics and criteria is proposed to extract enclosed crack boundary and provide accurate detection results under noise contamination and surface variations. Experimental study of crack detection on bridge deck surfaces is performed. The crack detection results are examined through a precision-recall analysis, which successfully demonstrates the effectiveness and robustness of this methodology.
14 May 2023 in Journal of Intelligent Transportation Systems
In recent years, rapid advancements in the Autonomous Vehicles (AVs) industry have greatly motivated the research and development in pedestrian trajectory prediction and risk assessment. One of the critical requirements for AVs is to predict the future trajectories of pedestrians and provide collision warnings in an accurate and prompt manner. Nevertheless, accurate prediction of pedestrian trajectories remains a technical challenge, mainly caused by the heterogeneity of pedestrian crossing behavior and uncertainties in vehicle-pedestrian interactions. This paper proposes a deep learning-based method for pedestrian trajectory prediction and risk assessment, using trajectory data extracted from roadside LiDAR data and corresponding signal phasing information at MLK and Georgia Avenue in Chattanooga, TN. Meanwhile, a set of criteria referred to as the risk factor is established to quantitatively evaluate the risk of the pedestrian crossing behavior, which also serves as a learnable feature. A Long Short-Term Memory (LSTM) network is proposed, which takes the following data as the input: the pedestrian trajectory data, signal phasing data, and risk factors from the past 10 steps. Meanwhile, the network predicts the pedestrian trajectory and risk factor at the future time step. In the experimental study, the root-mean-square errors between the predicted and ground truth x and y coordinates are 0.225 meters and 0.377 meters, respectively, and the F1 score value for the risk factor is 99.6%, demonstrating the efficacy of the proposed LSTM-based methodology on pedestrian trajectory prediction and risk assessment.
16 January 2024 in IEEE Transactions on Intelligent Transportation Systems
Since the past decade, Light Detection and Ranging (LiDAR) data have been extensively adopted for traffic object recognition tasks. Existing methodologies often assume LiDAR data are acquired under normal weather conditions. Nevertheless, many researchers have observed that the LiDAR data captured under inclement weather are often contaminated with noises such as fog and snow, which may deteriorate the data quality and lead to false detections in traffic object recognition. This paper proposes a neighborhood-based noise removal methodology to eliminate snow noises from LiDAR data. It identifies a point of interest from a specific laser channel as an outlier, if the number of neighboring points in the same channel within a dynamic search radius is fewer than a threshold. Unlike existing methods that filter the entire LiDAR point cloud, the proposed methodology processes LiDAR data channel-by-channel, which helps reduce the data dimensionality and decouple the snow effects along the vertical axis of the 3D point cloud, leading to more effective and efficient outlier detection. Furthermore, by dynamically changing the search radius based on the point-to-sensor distance rather than adopting a fixed search radius, the proposed methodology can account for the reduced point density at far distances caused by the non-uniformity of LiDAR data. In the experimental study, the proposed methodology is compared against some existing LiDAR de-noising approaches, including two state-of-the-art methods, and demonstrates superior performance in both accuracy (i.e., F1 score $=$ 98.3%) and efficiency.
31 December 2024 in IEEE Transactions on Intelligent Transportation Systems
In vehicle trajectory prediction, traditional methods like Kalman filtering often rely heavily on user expertise and prior knowledge, while newer deep learning approaches, such as Long Short-Term Memory (LSTM) networks, also face challenges related to human intervention and subjective hyperparameter selection. This study proposes a systematic approach for Light Detection and Ranging (LiDAR)-based vehicle trajectory prediction, leveraging LSTM networks to predict vehicle trajectories and employing Bayesian optimization to automatically search for optimal hyperparameter values related to both the training scheme and LSTM architectures. In the experimental study, a custom vehicle trajectory dataset extracted from roadside LiDAR data, along with the V2X-Seq-TFD dataset, was utilized for network training and testing. The optimal LSTM network obtained through Bayesian optimization was compared against two benchmark models: a handcrafted LSTM network and a Kalman filter with a 2D constant velocity motion model. The results demonstrate that the proposed deep learning-based framework, with robust hyperparameter selection through Bayesian optimization, yields more accurate and consistent prediction performance than the benchmark models.