Vehicle Type Classification using Hybrid Features and a Deep Neural Network

Currently, considerable research has been done in vehicle type classification, especially due to the success of deep learning in many image classification problems. In this research, a system incorporating hybrid features is proposed to improve the performance of vehicle type classification. The feature vectors are extracted from the pre-processed images using Gabor features, a histogram of oriented gradients and a local optimal oriented pattern. The hybrid set of features contains complementary information that could help discriminate between the classes better, further, an ant colony optimizer is utilized to reduce the dimension of the extracted feature vectors. Finally, a deep neural network is used to classify the types of vehicles in the images. The proposed approach was tested on the MIO vision traffic camera dataset and another more challenging real-world dataset consisting of videos of multiple lanes of a toll plaza. The proposed model showed an improvement in accuracy ranging from 0.28% to 8.68% in the MIO TCD dataset when compared to well-known neural network architectures.

CLASSIFICATION OF VEHICLE TYPES USING BACKPROPAGATION NEURAL NETWORKS WITH METRIC AND ECENTRICITY PARAMETERS

One of the efforts to break down traffic jams is to establish special lanes that can be passed by two, four or more wheeled vehicles. By being able to recognize the type of vehicle can reduce congestion. Citran based vehicle classification helps in providing information about the vehicle type. This study aims to classify the type of vehicle using a backpropagation neural network algorithm. The vehicle image can be recognized based on its shape, then the backpropagation neural network algorithm will be supported by metric and eccentricity parameters to perform feature extraction. Then from the results of feature extraction with metric parameters and eccentricity, the object will be classified using a backpropagation neural network algorithm. The test results show an accuracy of 87.5%. This shows the algorithm can perform classification well.

Spatial Seemingly Unrelated Regression Models for Freight Trip Generation by Vehicle Type: Application to the Chennai Metropolitan Area in India

During the last two decades, there has been substantial interest in developing freight trip generation (FTG) models. Most studies consider only truck trips or convert all freight trips into equivalent truck trips. Freight in several large cities is increasingly being moved by smaller vehicles. This calls for modeling FTG by vehicle type. The present research identifies and compares establishment characteristics affecting FTG by different vehicle types. In this context, spatial correlations among nearby establishments and the error-term correlations between independent models by vehicle type become relevant. Based on the Lagrange-Multiplier (LM) tests, we develop non-spatial seemingly unrelated regression (SUR) models for freight trip production (FTP) and spatial SUR models with a spatial lag in the dependent variable to account for both spatial and error-term correlations for freight trip attraction (FTA). The results show that establishment type and size affect FTG by different vehicle types.

Acceleration and Deceleration Rates of Various Vehicle Categories at Signalized Intersections in Mixed Traffic Conditions

The acceleration and deceleration rates vary from one vehicle type to another. The same vehicle type also exhibits variations in acceleration and deceleration rates due to vast variation in their dynamic and physical characteristics, ratio between weight and power, driver behaviour during acceleration and deceleration manoeuvres. Accurate estimation of acceleration and deceleration rates is very important for proper signal design to ensure minimum control delay for vehicles, which are passing through the intersection. The present study measures acceleration and deceleration rates for four vehicle categories: Two-wheeler, Three-wheeler, Car, and Light Commercial Vehicle (LCV), by using Open Street Map (OSM) tracker mobile application. The acceleration and deceleration rates were measured at 24 signalized intersection approaches in Hyderabad and Warangal cities. The study also developed acceleration and deceleration models for each vehicle type and the developed models were validated based on field data. The results showed that the predicted acceleration and deceleration models showed close relation with those measured in the field. The developed models are useful in predicting average acceleration and deceleration rate for different vehicle types under mixed and poor lane disciplined traffic conditions.

Integrative surveillance of cattle welfare at the abattoir level: Risk factors associated with liver condemnation, severe hoof disorders, carcase bruising and high muscle pH

Given the multi-dimensionality of animal welfare, any monitoring system for slaughter animals should comprise an integrative vision that facilitates animal welfare and food safety assessment. Thus, the aim of this study was to investigate risk factors as possible causes for liver condemnations, hoof disorders, bruise prevalence, and the quality of beef carcases under commercial operating conditions in Mexico. Data were recorded for 143 journeys encompassing 1,040 commercial cattle, originating from feedlots, free-range, and dairy production systems. Details on journey distance, vehicle type, cattle type, and animals' origin were gathered from abattoir reports. We found that carcase bruising (41%) and hoof disorders (43.9%) had the highest prevalence, regardless of the production system. Variables such as cattle type and production system influenced liver condemnations; old bulls extensively raised were more prone to present parasitosis such as Fasciola hepatica. Transportation conditions (journey distance, vehicle type) and cattle type might have influenced the development of hoof disorders in the evaluated animals. Multivariable logistic regression showed that animals' origin was a potential risk factor for severe bruising and high muscle pH, with cull dairy cows getting the most serious damage. In general, cattle transport conditions were factors that showed interactions with three of the evaluated indicators (severe hoof injuries, carcase bruising, meat pH24h). Our study shows the need to implement integrative surveillance to identify risk factors according to the production system from which the animals originate. With this information it is possible to develop strategies to mitigate specific cattle welfare problems.

AUTOMATED AND AUTONOMOUS VEHICLES - SAFETY, APPROVAL, SOCIAL BENEFITS AND FEARS OF INTRODUCING AUTOMATIC DRIVING SYSTEMS

The paper discusses the details of the work of automation systems for motor vehicles and their methods of testing. The requirements of the new UN Regulation No. 157 were presented as a tool for conducting the type-approval of automated vehicles in the field of the Automated Lane Keeping System (ALKS). The most important requirements and methods of testing ALKS systems for the vehicle type-approval are described. Evaluation of the advantages and disadvantages as well as the effects and social concerns of the implementation of such systems in relation to the road safety, has been carried out.

Bangladeshi Native Vehicle Classification Based on Transfer Learning with Deep Convolutional Neural Network

Vehicle type classification plays an essential role in developing an intelligent transportation system (ITS). Based on the modern accomplishments of deep learning (DL) on image classification, we proposed a model based on transfer learning, incorporating data augmentation, for the recognition and classification of Bangladeshi native vehicle types. An extensive dataset of Bangladeshi native vehicles, encompassing 10,440 images, was developed. Here, the images are categorized into 13 common vehicle classes in Bangladesh. The method utilized was a residual network (ResNet-50)-based model, with extra classification blocks added to improve performance. Here, vehicle type features were automatically extracted and categorized. While conducting the analysis, a variety of metrics was used for the evaluation, including accuracy, precision, recall, and F1 − Score. In spite of the changing physical properties of the vehicles, the proposed model achieved progressive accuracy. Our proposed method surpasses the existing baseline method as well as two pre-trained DL approaches, AlexNet and VGG-16. Based on result comparisons, we have seen that, in the classification of Bangladeshi native vehicle types, our suggested ResNet-50 pre-trained model achieves an accuracy of 98.00%.