Traffic monitoring is of major importance for enforcing traffic management policies.To accomplish this task,the detection of vehicle can be achieved by exploiting image analysis techniques.In this paper,a solution is ...Traffic monitoring is of major importance for enforcing traffic management policies.To accomplish this task,the detection of vehicle can be achieved by exploiting image analysis techniques.In this paper,a solution is presented to obtain various traffic parameters through vehicular video detection system(VVDS).VVDS exploits the algorithm based on virtual loops to detect moving vehicle in real time.This algorithm uses the background differencing method,and vehicles can be detected through luminance difference of pixels between background image and current image.Furthermore a novel technology named as spatio-temporal image sequences analysis is applied to background differencing to improve detection accuracy.Then a hardware implementation of a digital signal processing (DSP) based board is described in detail and the board can simultaneously process four-channel video from different cameras. The benefit of usage of DSP is that images of a roadway can be processed at frame rate due to DSP′s high performance.In the end,VVDS is tested on real-world scenes and experiment results show that the system is both fast and robust to the surveillance of transportation.展开更多
A new training symbol weighted by pseudo-noise(PN) sequence is designed and an efficient timing and fre-quency offset estimation scheme for orthogonal frequency division multiplexing(OFDM) systems is proposed. The tim...A new training symbol weighted by pseudo-noise(PN) sequence is designed and an efficient timing and fre-quency offset estimation scheme for orthogonal frequency division multiplexing(OFDM) systems is proposed. The timing synchronization is accomplished by using the piecewise symmetric conjugate of the primitive training symbol and the good autocorrelation of PN weighted factor. The frequency synchronization is finished by utilizing the training symbol whose PN weighted factor is removed after the timing synchronization. Compared with conventional schemes, the pro-posed scheme can achieve a smaller mean square error and provide a wider frequency acquisition range.展开更多
Detecting small objects is a challenging task.We focus on a special case:the detection and classification of traffic signals in street views.We present a novel framework that utilizes a visual attention model to make ...Detecting small objects is a challenging task.We focus on a special case:the detection and classification of traffic signals in street views.We present a novel framework that utilizes a visual attention model to make detection more efficient,without loss of accuracy,and which generalizes.The attention model is designed to generate a small set of candidate regions at a suitable scale so that small targets can be better located and classified.In order to evaluate our method in the context of traffic signal detection,we have built a traffic light benchmark with over 15,000 traffic light instances,based on Tencent street view panoramas.We have tested our method both on the dataset we have built and the Tsinghua–Tencent 100K(TT100K)traffic sign benchmark.Experiments show that our method has superior detection performance and is quicker than the general faster RCNN object detection framework on both datasets.It is competitive with state-of-theart specialist traffic sign detectors on TT100K,but is an order of magnitude faster.To show generality,we tested it on the LISA dataset without tuning,and obtained an average precision in excess of 90%.展开更多
文摘Traffic monitoring is of major importance for enforcing traffic management policies.To accomplish this task,the detection of vehicle can be achieved by exploiting image analysis techniques.In this paper,a solution is presented to obtain various traffic parameters through vehicular video detection system(VVDS).VVDS exploits the algorithm based on virtual loops to detect moving vehicle in real time.This algorithm uses the background differencing method,and vehicles can be detected through luminance difference of pixels between background image and current image.Furthermore a novel technology named as spatio-temporal image sequences analysis is applied to background differencing to improve detection accuracy.Then a hardware implementation of a digital signal processing (DSP) based board is described in detail and the board can simultaneously process four-channel video from different cameras. The benefit of usage of DSP is that images of a roadway can be processed at frame rate due to DSP′s high performance.In the end,VVDS is tested on real-world scenes and experiment results show that the system is both fast and robust to the surveillance of transportation.
基金Supported by the National High Technology Research and Development Program of China ( "863" Program, No.2006AA01Z270)Natural Science Foun-dation of Shaanxi Province (No. 2007F07)+1 种基金Natural Science Foundation of Guangdong Province (No. U0635003)National "111" Program of Intro-ducing Talents of Discipline to Universities (No. B08038)
文摘A new training symbol weighted by pseudo-noise(PN) sequence is designed and an efficient timing and fre-quency offset estimation scheme for orthogonal frequency division multiplexing(OFDM) systems is proposed. The timing synchronization is accomplished by using the piecewise symmetric conjugate of the primitive training symbol and the good autocorrelation of PN weighted factor. The frequency synchronization is finished by utilizing the training symbol whose PN weighted factor is removed after the timing synchronization. Compared with conventional schemes, the pro-posed scheme can achieve a smaller mean square error and provide a wider frequency acquisition range.
基金supported by the National Natural Science Foundation of China (No.61772298)Research Grant of Beijing Higher Institution Engineering Research Centerthe Tsinghua–Tencent Joint Laboratory for Internet Innovation Technology
文摘Detecting small objects is a challenging task.We focus on a special case:the detection and classification of traffic signals in street views.We present a novel framework that utilizes a visual attention model to make detection more efficient,without loss of accuracy,and which generalizes.The attention model is designed to generate a small set of candidate regions at a suitable scale so that small targets can be better located and classified.In order to evaluate our method in the context of traffic signal detection,we have built a traffic light benchmark with over 15,000 traffic light instances,based on Tencent street view panoramas.We have tested our method both on the dataset we have built and the Tsinghua–Tencent 100K(TT100K)traffic sign benchmark.Experiments show that our method has superior detection performance and is quicker than the general faster RCNN object detection framework on both datasets.It is competitive with state-of-theart specialist traffic sign detectors on TT100K,but is an order of magnitude faster.To show generality,we tested it on the LISA dataset without tuning,and obtained an average precision in excess of 90%.
