Sign & Graphic Processing: An International Journal (SIPIJ) Vol. two, No . two, June 2011
VEHICLE TRAFFIC MONITORING USING KALMAN FILTER AND
Amir Salarpour1and Arezoo Salarpour2and Mahmoud Fathi2and MirHossein Dezfoulian1
Department of Computer Architectural, BuAliSina University or college, Hamedan, Iran a.salarpour, dezfoulian @basu. ac. ventosear
Division of Computer system Engineering, Iran University of Science and Technology, Tehran, Iran
Vehicle monitoring has a wide variety of applications. The image resolution of the video obtainable from many traffic camera system is low. In many cases intended for tracking multiple object, distinguishing them from another just isn't easy because of the similarity. In this paper we all describe a method, for monitoring multiple things, where the items are automobiles. The number of vehicles is not known and varies. We find all shifting objects, and then for tracking of vehicle we use the kalman filter and color feature and distance of it from frame to another. So the method can separate and monitoring all vehicles individually. The proposed protocol can be placed on multiple going objects.
Kalman filtering, occlusion, energetic contour
1 ) INTRODUCTION
There are many of traffic monitoring systems being used. Traffic cameras give a more flexible means of monitoring traffic. These cameras not only can be used in basic tasks just like counting automobiles, they also have the to be utilized in more complex applications like tracking. Multiple subject tracking is an important research theme in laptop vision. It has the ability of deal with the only object issues such as obturation with backdrop changing appearance, illumination, non rigid motion and the variable object troubles such as obturation between objects and subject confusion. In  tracking fix number of objects. In  an effective algorithm in order to multiple persons is provided.  proposed a bayesian tracker for tracking multiple blob. Various tracking algorithms have been proposed in the literary works, including strategies templates and native features  Kalman filter systems  and shape . The mean-shift protocol was first adopted as an effective tracking approach in .
In tracking devices two complications must be regarded: prediction and correction. Forecast problem: predict the location of the object getting tracked over the following frame, that is identify a region in which the likelihood of finding object is large. Correction problem: identify the item in the next framework within chosen region. A well-known solution pertaining to prediction is definitely Kalman filter, a recursive estimator of state of any dynamic system. Kalman filtration system have been used in online video tracking . To predict the search region more effectively, mean-shift was coupled with Kalman filter in . To enhance the ability of meanDOI: twelve. 5121/sipij. 2011. 2201
Signal & Image Control: An International Record (SIPIJ) Volume. 2, Number 2, 06 2011
move tracking with respect to scale improvements, a scale selection device was launched in  based on Lindeberg's theory. Mean-shift has been found in the past to moving targets in sequences of FLIR imagery  and human being bodies (i. e., nonrigid objects) in . The static correction problem requires a similarity metric to compare candidate pairs of thing in earlier and current frame. This is actually the correspondence of object in two structures. Matching metrics in static correction problem is essential. A traffic monitoring problem is data association, seeking the true placement of moving target, when there are more than one valid applicant. This occurs in muddle scenes.
The success or failure of any tracking algorithm is dependent a lot for the degree the fact that tracked thing can be recognized from its surroundings . In particular, the set of features used by the tracking formula to represent the object(s) becoming tracked takes on a major function in checking performance.
Recommendations: for stage correspondenceвЂќ, IEEE conference of Computer Perspective and Pattern Recognition, 704709.
Zhimin Fan, Jie Zhou, Dashan Gao and Zhiheng Li, (2002) " Contour Extraction And Checking Of
J Lou, H Yang, Wei Ming Hu, Tieniu Tan, (2002) " Image vehicle checking using an increased