Multiple moving vehicle speed estimation using Blob analysis

Abstract

Nowadays, traffic problems including vehicles monitoring for their speed is the need of this modern world. The video-based surveillance system is used to monitor and identifying the moving vehicles on road for safety measurement of public. The proposed work is aimed to find out the over-speeding of multiple vehicles through an automatic vehicle video tracking system. The method build for the automatic vehicles tracking systems is based on a couple of steps; initially the frame differencing is applied for the object detection. In the frame differencing, the background model is computed and the current frame is subtract from previous frame which is computed during background modeling for the feature extraction. The blob analysis is then used to track the detected object or the region of interest. Ultimately the speed is estimated to compute and find out the over speeding of vehicles. The Euclidian distance determines the centroid of every region in both horizontal and vertical direction. The overall accuracy achieved by our proposed system is 90% for the detection and measurement of speeds of multiple vehicles.