Traffic Flow Estimation based on Deep Learning for Emergency Traffic Management using CCTV Images (original) (raw)
Related papers
2020
Deep Learning based networks especially Convolutional Neural Network (CNN) models are widely used in vehicle detection, classification and counting system. On the other hand, transfer learning is a process of re-using a trained model to solve a problem similar to the one it was trained. Two ways of implementing transfer learning are direct usage of a model as a classifier and usage of a pre-trained model as a weight initialization for training with a new dataset. With recent development in the field of deep learning, many CNN models and architectures are available which makes the selection of a suitable model for performing vehicle detection, classification and counting a big challenge. Besides that, a tracking method is also required to track the vehicles in the video sequences so that the counting can be done as accurate as possible. In this project three types of CNN models i.e. SSD Inception, Faster R-CNN ResNet and Yolo DarkNet were tested on 10 traffic video samples using tran...
OBJECT DETECTION IN TRAFFIC SCENARIOS -A COMPARISON OF TRADITIONAL AND DEEP LEARNING APPROACHES
In the area of computer vision, research on object detection algorithms has grown rapidly as it is the fundamental step for automation, specifically for self-driving vehicles. This work presents a comparison of traditional and deep learning approaches for the task of object detection in traffic scenarios. The handcrafted feature descriptor like Histogram of oriented Gradients (HOG) with a linear Support Vector Machine (SVM) classifier is compared with deep learning approaches like Single Shot Detector (SSD) and You Only Look Once (YOLO), in terms of mean Average Precision (mAP) and processing speed. SSD algorithm is implemented with different backbone architectures like VGG16, MobileNetV2 and ResNeXt50, similarly YOLO algorithm with MobileNetV1 and ResNet50, to compare the performance of the approaches. The training and inference is performed on PASCAL VOC 2007 and 2012 training, and PASCAL VOC 2007 test data respectively. We consider five classes relevant for traffic scenarios, namely, bicycle, bus, car, motorbike and person for the calculation of mAP. Both qualitative and quantitative results are presented for comparison. For the task of object detection, the deep learning approaches outperform the traditional approach both in accuracy and speed. This is achieved at the cost of requiring large amount of data, high computation power and time to train a deep learning approach.
International journal of simulation: systems, science & technology, 2019
Intelligent Transportation System (ITS) is one of the attributes that describe smart cities. One of its functions is detection and classification of vehicles that pass through roadways. With this information, traffic management sectors can plan and implement road rules for the betterment of the traffic flow. Vision-based approaches and other methods, however, work only in ideal environment which make researchers find new ways on how limitations like occlusions, nighttime and camera angle can be solved. This paper demonstrates using a deep learning method to accurately detect and classify vehicles on urban roadways in a certain city. Additionally, a vehicle classifier was built and tested using a machine learning framework known as TensorFlow. Faster R-CNN model, with captured CCTV-video as dataset, was used to train the vehicle classifier. The performance of the newlytrained classifier has been evaluated using different classification metrics. Results show that using the proposed method, 93% accuracy and 78% F1-score in detecting and classifying vehicles were achieved based on labeled data. However, researchers also took note of the detection errors that showed during testing. Configurations in some steps has been provided to minimize such misclassifications. It was also recommended that the method be integrated as vital part of Intelligent Transportation Systems (ITS) in terms of vehicle detection and classification for future smart cities.
2021
Accurate traffic data collection is crucial to the relevant authorities in ensuring the planning, design, and management of the road network can be done appropriately. Traditionally, traffic data collection was done manually by having human observers at the site to count the vehicle as it passes the observation point. This approach is cost-effective; however, the accuracy can’t be verified and may cause danger to the observers. Another common approach is utilizing sensors that need to be installed underneath the road surface to collect traffic data. The accuracy of the data reading from the sensor is highly dependent on the sensor installation, calibration, and reliability which usually deteriorated over time. For these reasons, vision-based approaches have become more popular in traffic flow estimation tasks. Nevertheless, conventional image processing techniques which utilize background subtraction-based approach may face problems in complex highway environment where the number of...
Vehicle Counting using Deep Learning Models: A Comparative Study
International Journal of Advanced Computer Science and Applications, 2020
Recently, there has been a shift to deep learning architectures for better application in vehicle traffic control systems. One popular deep learning library used for detecting vehicle is TensorFlow. In TensorFlow, the pre-trained model is very efficient and can be transferred easily to solve other similar problems. However, due to inconsistency between the original dataset used in the pre-trained model and the target dataset for testing, this can lead to low-accuracy detection and hinder vehicle counting performance. One major obstacle in retraining deep learning architectures is that the network requires a large corpus training dataset to secure good results. Therefore, we propose to perform data annotation and transfer learning from an existing model to construct a new model for vehicle detection and counting in the real world urban traffic scenes. Then, the new model is compared with the experimental data to verify the validity of the new model. Besides, this paper reports some experimental results, comprising a set of innovative tests to identify the best detection algorithm and system performance. Furthermore, a simple vehicle tracking method is proposed to aid the vehicle counting process in challenging illumination and traffic conditions. The results showed a significant improvement of the proposed system with the average vehicle counting of 80.90%.
International Journal of Scientific Research in Computer Science, Engineering and Information Technology, 2021
In the current scenario on the increasing number of motor vehicles day by day, so traffic regulation faces many challenges on intelligent road surveillance and governance, this is one of the important research areas in the artificial intelligence or deep learning. Among various technologies, computer vision and machine learning algorithms have the most efficient, as a huge vehicles video or image data on road is available for study. In this paper, we proposed computer vision-based an efficient approach to vehicle detection, recognition and Tracking. We merge with one-stage (YOLOv4) and two-stage (R-FCN) detectors methods to improve vehicle detection accuracy and speed results. Two-stage object detection methods provide high localization and object recognition precision, even as one-stage detectors achieve high inference and test speed. Deep-SORT tracker method applied for detects bounding boxes to estimate trajectories. We analyze the performance of the Mask RCNN benchmark, YOLOv3 and Proposed YOLOv4 + R-FCN on the UA-DETRAC dataset and study with certain parameters like Mean Average Precisions (mAP), Precision recall.
Vehicle Traffic Analysis using CNN Algorithm
IRJET, 2022
The goal is to build a traffic light system that changes based on how many people are in the area. When there is a lot of traffic at an intersection, the signal time automatically changes. Many major cities around the world have a lot of traffic, which makes it hard to get to work every day. Traditional traffic signal systems are based on the idea that each side of the intersection has a set amount of time. They can't be changed to account for more traffic. People can't change the times of the intersections that have been set up for them. There may be more traffic on one intersection, which could make it more difficult for the typical green period to end. After processing and translating the traffic signal object detection into a simulator, a threshold is set and a contour is drawn many cars are in the area. After , we can figure out which side has the most cars based on the signals sent to each side. Paper provides a solution based on camera feed at crossing for each lane process the data through and allocates the "green" time according to its traffic flow density using YOLO v3 and also takes care of starvation issue that might arise of the solution. As a result ,the flow of traffic on each lane is automatically optimized and the congestion that used to happen unnecessarily is eliminated earlier and results show significant benefits in reducing traffic waiting time
Sensors, 2019
Automatic traffic flow classification is useful to reveal road congestions and accidents. Nowadays, roads and highways are equipped with a huge amount of surveillance cameras, which can be used for real-time vehicle identification, and thus providing traffic flow estimation. This research provides a comparative analysis of state-of-the-art object detectors, visual features, and classification models useful to implement traffic state estimations. More specifically, three different object detectors are compared to identify vehicles. Four machine learning techniques are successively employed to explore five visual features for classification aims. These classic machine learning approaches are compared with the deep learning techniques. This research demonstrates that, when methods and resources are properly implemented and tested, results are very encouraging for both methods, but the deep learning method is the most accurately performing one reaching an accuracy of 99.9% for binary tr...
Procedia Computer Science, 2018
In this paper, we present an intelligent traffic congestion detection method using image classification approach on CCTV camera image feeds. We use a deep learning architecture, convolutional neural network (CNN) which is currently the state-of-the art for image processing method. We only do minimal image preprocessing steps on the small size image, where the conventional methods require a high quality, handcrafted features need to do manual calculation. The CNN model is trained to do binary classification about road traffic condition using 1000 CCTV monitoring image feeds with balance distribution. The result shows that a CNN with simple, basic architecture that trained on small grayscale images has an average classification accuracy of 89.50%.
An Advanced Deep Learning Approach for Multi-Object Counting in Urban Vehicular Environments
Future Internet
Object counting is an active research area that gained more attention in the past few years. In smart cities, vehicle counting plays a crucial role in urban planning and management of the Intelligent Transportation Systems (ITS). Several approaches have been proposed in the literature to address this problem. However, the resulting detection accuracy is still not adequate. This paper proposes an efficient approach that uses deep learning concepts and correlation filters for multi-object counting and tracking. The performance of the proposed system is evaluated using a dataset consisting of 16 videos with different features to examine the impact of object density, image quality, angle of view, and speed of motion towards system accuracy. Performance evaluation exhibits promising results in normal traffic scenarios and adverse weather conditions. Moreover, the proposed approach outperforms the performance of two recent approaches from the literature.