Vehicle Speed Determination using HMM
This paper presents an application of computer vision methods for traffic vehicle tracking and speed detection. The application is utilizing background subtraction, object detection, feature selection and then object tracking. These methods combined together gives functional capabilities to the system to initiate automated vehicle tracking and to determine their speeds. To find the speed of vehicle in a high road we introduce a statistical probability model namely Hidden Markov Model (HMM) .In this method we will train the HMM using cars videos with speeds within the specified range which we want to consider. After training we test the car speeds on the trained HMM, if car in the video have a speed within the specified range the HMM will give a high probability value and if it is less or more than the specified range it will give a low probability value.
Keywords - Hidden Markov Model (HMM), Feature Extraction, Blob Identification, Object Detection, Object Tracking, Kalman Filter.