Currently, research on road traffic safety is mostly focused on traffic safety evaluations based on statistical indices for accidents. There is still a need for in-depth investigation on preaccident identification of safety risks. In this study, the correlations between high-incidence locations for aberrant driving behaviors and locations of road traffic accidents are analyzed based on vehicle OBD data. A road traffic safety risk estimation index system with road traffic safety entropy (RTSE) as the primary index and rapid acceleration frequency, rapid deceleration frequency, rapid turning frequency, speeding frequency, and high-speed neutral coasting frequency as secondary indices is established. A calculation method of RTSE is proposed based on an improved entropy weight method. This method involves three aspects, namely, optimization of the base of the logarithm, processing of zero-value secondary indices, and piecewise calculation of the weight of each index. Additionally, a safety risk level determination method based on two-step clustering (density and k-means clustering) is also proposed, which prevents isolated data points from affecting safety risk classification. A risk classification threshold calculation method is formulated based on k-mean clustering. The results show that high-incidence locations for aberrant driving behaviors are consistent with the locations of traffic accidents. The proposed methods are validated through a case study on four roads in Chongqing with a total length of approximately 38 km. The results show that the road traffic safety trends characterized by road safety entropy and traffic accidents are consistent.

According to the scarcity of analysis on temporal and spatial shared impacts of parking behavior of comprehensive development blocks as well as that on differences referring to traffic location-based factors in parking demand forecasting of urban comprehensive development blocks, this article uncovers the mechanism of shared parking behavior, discusses and designs the process and method of analysis of shared parking behavior. Then, formulas both for the influence coefficient of shared parking behavior and that of traffic location-based factors are worked out. Based on the work above, a new forecasting model of unit parking demand generation for comprehensive development blocks is proposed. In the end, the validity of the model is verified by a case study.

For accurate and effective automatic vehicle identification, morphological detection and deep convolutional networks were combined to propose a method for locating and identifying vehicle models from unmanned aerial vehicle (UAV) videos. First, the region of interest of the video frame image was sketched and grey-scale processing was performed; sub-pixel-level skeleton images were generated based on the Canny edge detection results of the region of interest; then, the image skeletons were decomposed and reconstructed. Second, a combination of morphological operations and connected domain morphological features were applied for vehicle target recognition, and a deep learning image benchmark library containing 244,520 UAV video vehicle samples was constructed. Third, we improved the AlexNet model by adding convolutional layers, pooling layers, and adjusting network parameters, which we named AlexNet*. Finally, a vehicle recognition method was established based on a candidate target extraction algorithm with AlexNet*. The validation analysis revealed that AlexNet* achieved a mean $F_1$ of 85.51% for image classification, outperforming AlexNet (82.54%), LeNet (63.88%), CaffeNet (46.64%), VGG16 (16.67%), and GoogLeNet (14.38%). The mean values of $P_{cor}$, $P_{re}$, and $P_{miss}$ for cars and buses reached 94.63%, 6.87%, and 4.40%, respectively, proving that this method can effectively identify UAV video targets.

Video surveillance has become an important measure of urban traffic monitoring and control. However, due to the complex and diverse video scenes, traffic data extraction from original videos is a sophisticated and difficult task, and corresponding algorithms are of high complexity and calculation cost. To reduce the algorithm complexity and subsequent computation cost, this study proposed an autoencoder model which effectively reduces the video dimension by optimizing structural parameters; thus several traffic recognition models can conduct image processing work based on dimension-reduced videos. Firstly, an optimal autoencoder model with five hidden layers was constructed. Then, it was combined with a linear classifier, support vector machine, deep neural network, DNN linear classification method, and the k-means clustering method; thus, five traffic state recognition models were constructed: -Linear, -SVM, -DNN, -DNN_Linear, and -k-means. Train and test results show that the accuracy rate and recall rate of -linear, -SVM, -DNN, and -DNN_Linear were 94.5%–97.1%, and the F1 score was 94.4%–97.1%; besides, the accuracy rate, recall rate, and F1 score of -k-means were all approximately 95%, which suggests that the combination of the autoencoder and common classification or clustering methods achieve good recognition performance. Comparison was also implemented among the five models proposed above and four CNN-based models such as AlexNet, LeNet, GoogLeNet, and VGG16, which shows that the five proposed modes achieve F1 scores of 94.4%–97.1%, while the four CNN-based models achieve F1 scores of 16.7%–94%, indicating that the proposed light weight design methods outperform more complex CNN-based models in traffic state recognition.

To accurately prevent and warn of traffic accidents, this article proposes a method for predicting urban road traffic safety risks based on vehicle driving behaviour data and information entropy theory. This method uses data from radar video-integrated sensors to calibrate the thresholds for identifying unsafe driving behaviour, introduces recognition principles and algorithms, and analyses spatiotemporal distribution patterns. By incorporating entropy theory, an evaluation system with traffic safety entropy as the primary indicator and the unsafe driving behaviour rate as the secondary indicator is established. Clustering algorithms determine the classification number and threshold of traffic safety entropy, constructing a tunnel traffic safety risk assessment model, which is validated with road accident data. Using 13 days of data from the left lane of Qingdao Jiaozhou Bay Tunnel, the model divides traffic operation risk into high and low categories based on K-means clustering results of accident and safety entropy data. The study finds that when the safety entropy classification threshold is 0.0507, the classification accuracy is the highest at 92%. These results provide technical support for identifying road traffic safety risk points and preventing accidents.