Abstract
Traffic congestion detection plays an important role for road management. However, it is difficult to automatically report traffic congestion when it occurs in large-scale road network. One of key challenges for rapidly and precisely identifying early congestion is huge variations in appearance caused by illumination, weather, camera settings and other traffic conditions. To address it, we proposed a traffic-oriented model to classify congestion from large dataset of ultra-low frame rate video captured from traffic surveillance system. The proposed deeply supervised traffic congestion detector has two modules: attention proposal module and deeply supervised inception network. Specifically, within the shallow layers, the binary edge/corner density features are used in attention proposal module to generate the rang of interest (ROI) mask automatically. This strategy keeps the training process focusing on the congestion features without disturbances. Following the attention proposal module, a very deep structure based on the inception network was used together to effectively extract rich and discriminative features then detect traffic congestion. The approach was tested on a self-established dataset based on empirical data, which contains images captured from 14470 surveillance cameras for monitoring 5,215 km of freeway in Shaanxi province, China. The experimental results show that the accuracy of the proposed method could reach 95.77% considering various disturbances, conditions and other limitations, which is improved than unsupervised networks.
Highlights
Traffic congestion detection plays an important part for road management
In our previous work [13], [14], we explored the use of deep convolution neural network (CNN) for congestion detecting
The traffic congestion dataset is established based on an existing surveillance system, which includes 14470 cameras used for freeway monitoring by the Traffic Management Bureau in Shaanxi Province, China
Summary
Traffic congestion detection plays an important part for road management. Distinction of traffic condition facilitate better decision making in accident rescuing, extended congestion preventing and traffic efficiency improving. Reference [21] proposed a method to count entering cars using virtual line, the traffic condition can be identified based on the number of vehicles. Features, such as LBP (local binary pattern), Harr-like features and middle-level image features (sparse encoding), to compare the similarity of two adjacent frames, images with high similarity are identified as congestion Another way [28] is similar to the vehicle density based methods using successive video frames, but the vehicle and road are detected by the sematic segmentation technique with CNN method, which is more efficiently than the method based on the frame subtraction. References [29], [30] classify the traffic conditions based on some low-level image features (such as color, edge, texture and key-point features) and middle-level features (such as codebook descriptors) extracted from one image.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
