Vehicle Flow Detection Using Fast Region Matching with Adaptive Gaussian Mixture Background Model

2010 ◽  
Vol 108-111 ◽  
pp. 1272-1277
Author(s):  
Hui Ding

Video surveillance play an important role in many ITS. In this paper, we present a fast and reliable algorithm for detecting traffic flow count. The core of the algorithm relies on Gaussian mixture background model combined fast cross-correlation region-based techniques in moving object matching. By working with adaptive Gaussian mixture model, obtained the moving vehicle as foreground. Then, fast local correlation, referred to as single matching phase, is achieved by using recursive computation schemes, which enabled us to minimize the amount of calculations required at every new pixel. We have tested our match algorithm in a large set of experiments with video clips and achieved good matching results.

2015 ◽  
Vol 734 ◽  
pp. 463-467 ◽  
Author(s):  
Pan Pan Zhang ◽  
Chun Yang Mu ◽  
Xing Ma ◽  
Fu Lu Xu

Detection of moving object is a hot topic in computer vision. Traditionally, it is detected for every pixel in whole image by Gaussian mixture background model, which may waste more time and space. In order to improving the computational efficiency, an advanced Gaussian mixture model based on Region of Interest was proposed. Firstly, the solution finds out the most probably region where the target may turn up. And then Gaussian mixture background model is built in this area. Finally, morphological filter algorithm is used for improving integrity of the detected targets. Results show that the improved method could have a more perfect detection but no more time increasing than typical method.


2018 ◽  
Vol 7 (4) ◽  
pp. 2678
Author(s):  
Budi Setiyono ◽  
Dwi Ratna Sulistyaningrum ◽  
Soetrisno . ◽  
Hasanuddin Al-Habib

Intelligent Transportation System (ITS) is a concept to manage transportation based on technology development. Video from surveillance cameras can be used for monitoring the number of vehicles and speed using digital image processing. Shadows on the vehicle is one of the noise that must be removed in order to obtain better accuracy. Shadow is caused by the reflection of objects exposed to the light. In this study, we combined two methods to eliminate shadows on moving vehicle, the subregion illumination transfer method and the background-based Gaussian mixture model. Foreground image is used for sub-Region Illumination Transfer and gamma decoding processes is used to detect the presence of shadows The detected shadow is removed by replacing it with the background in that position. Experiments are done by making simulated video of moving objects without shadows and objects that have a shadow. By using the proposed method, the shadow will be omitted, and the results are compared with the object without the shadow. The experimental results are: mean value of PSNR for objects moving closer to the camera with a light intensity of 0.8 is 53.47. While on the moving object with a small shadow area, we obtained an average PSNR of 51.87927dB.  


2015 ◽  
Vol 734 ◽  
pp. 203-206
Author(s):  
En Zeng Dong ◽  
Sheng Xu Yan ◽  
Kui Xiang Wei

In order to enhance the rapidity and the accuracy of moving target detection and tracking, and improve the speed of the algorithm on the DSP (digital signal processor), an active visual tracking system was designed based on the gaussian mixture background model and Meanshift algorithm on DM6437. The system use the VLIB library developed by TI, and through the method of gaussian mixture background model to detect the moving objects and use the Meanshift tracking algorithm based on color features to track the target in RGB space. Finally, the system is tested on the hardware platform, and the system is verified to be quickness and accuracy.


2021 ◽  
Author(s):  
Guohua Gao ◽  
Jeroen Vink ◽  
Fredrik Saaf ◽  
Terence Wells

Abstract When formulating history matching within the Bayesian framework, we may quantify the uncertainty of model parameters and production forecasts using conditional realizations sampled from the posterior probability density function (PDF). It is quite challenging to sample such a posterior PDF. Some methods e.g., Markov chain Monte Carlo (MCMC), are very expensive (e.g., MCMC) while others are cheaper but may generate biased samples. In this paper, we propose an unconstrained Gaussian Mixture Model (GMM) fitting method to approximate the posterior PDF and investigate new strategies to further enhance its performance. To reduce the CPU time of handling bound constraints, we reformulate the GMM fitting formulation such that an unconstrained optimization algorithm can be applied to find the optimal solution of unknown GMM parameters. To obtain a sufficiently accurate GMM approximation with the lowest number of Gaussian components, we generate random initial guesses, remove components with very small or very large mixture weights after each GMM fitting iteration and prevent their reappearance using a dedicated filter. To prevent overfitting, we only add a new Gaussian component if the quality of the GMM approximation on a (large) set of blind-test data sufficiently improves. The unconstrained GMM fitting method with the new strategies proposed in this paper is validated using nonlinear toy problems and then applied to a synthetic history matching example. It can construct a GMM approximation of the posterior PDF that is comparable to the MCMC method, and it is significantly more efficient than the constrained GMM fitting formulation, e.g., reducing the CPU time by a factor of 800 to 7300 for problems we tested, which makes it quite attractive for large scale history matching problems.


Author(s):  
Swaroop Dinakar ◽  
Jeffrey W. Muttart ◽  
Teena Garrison ◽  
Suntasy Gernhard ◽  
Jim Marr

Rear-end crashes contribute to a large percentage of fatal collisions in the United States. However, every rear-end collision cannot be classified as a single type of crash. Some crashes may be caused due to human error while some crashes may be attributed to a human inability to recognize closing speed well. Observers were shown two 4-second video clips of a commercial vehicle closing on a slow-moving vehicle on an unlit highway. The lead vehicle was depicted at distances of 91m (300 ft), 128m (420 ft) and 152m (500 ft). Closing speeds of 40 km/h (25 mph) and 105 km/h (65 mph) were depicted. The taillights on the lead vehicle were randomly shown as bright, or 80% dimmer which is typical of older taillights or aged retroreflective materials. Results showed that observers’ ability to recognize closing from separating worsened with increased distance, dimmer taillights and lower closing speeds. Observers perceived brighter taillights to be closer. Also, at greater distances, observers did not recognize closing speeds as well.


2017 ◽  
Vol 831 ◽  
pp. 779-825 ◽  
Author(s):  
Mohammad Mohaghar ◽  
John Carter ◽  
Benjamin Musci ◽  
David Reilly ◽  
Jacob McFarland ◽  
...  

The effect of initial conditions on transition to turbulence is studied in a variable-density shock-driven flow. Richtmyer–Meshkov instability (RMI) evolution of fluid interfaces with two different imposed initial perturbations is observed before and after interaction with a second shock reflected from the end wall of a shock tube (reshock). The first perturbation is a predominantly single-mode long-wavelength interface which is formed by inclining the entire tube to 80$^{\circ }$ relative to the horizontal, yielding an amplitude-to-wavelength ratio, $\unicode[STIX]{x1D702}/\unicode[STIX]{x1D706}=0.088$, and thus can be considered as half the wavelength of a triangular wave. The second interface is multi-mode, and contains additional shorter-wavelength perturbations due to the imposition of shear and buoyancy on the inclined perturbation of the first case. In both cases, the interface consists of a nitrogen-acetone mixture as the light gas over carbon dioxide as the heavy gas (Atwood number, $A\sim 0.22$) and the shock Mach number is $M\approx 1.55$. The initial condition was characterized through Proper Orthogonal Decomposition and density energy spectra from a large set of initial condition images. The evolving density and velocity fields are measured simultaneously using planar laser-induced fluorescence (PLIF) and particle image velocimetry (PIV) techniques. Density, velocity, and density–velocity cross-statistics are calculated using ensemble averaging to investigate the effects of additional modes on the mixing and turbulence quantities. The density and velocity data show that a distinct memory of the initial conditions is maintained in the flow before interaction with reshock. After reshock, the influence of the long-wavelength inclined perturbation present in both initial conditions is still apparent, but the distinction between the two cases becomes less evident as smaller scales are present even in the single-mode case. Several methods are used to calculate the Reynolds number and turbulence length scales, which indicate a transition to a more turbulent state after reshock. Further evidence of transition to turbulence after reshock is observed in the velocity and density fluctuation spectra, where a scaling close to $k^{-5/3}$ is observed for almost one decade, and in the enstrophy fluctuation spectra, where a scaling close to $k^{1/3}$ is observed for a similar range. Also, based on normalized cross correlation spectra, local isotropy is reached at lower wave numbers in the multi-mode case compared with the single-mode case before reshock. By breakdown of large scales to small scales after reshock, rapid decay can be observed in cross-correlation spectra in both cases.


1997 ◽  
Vol 23 ◽  
pp. S145
Author(s):  
F. Gens ◽  
J.P. Remenieras ◽  
S. Diridollou ◽  
Y. Gall ◽  
F. Patat ◽  
...  

Sign in / Sign up

Export Citation Format

Share Document