Development of prototype for vehicle speed measurement
Rohit M. Waregaonkar, S. M. Deokar, Sanika S Patankar, Jayant V Kulkarni · 2017
Motion estimation and tracking are significant activities in many computer vision applications. Real-time surveillance is a vigorous element in any circumstances or environment where there is a great necessity of security for both personal and commercial property and assets. This paper aims to develop a prototype for measurement of vehicle speed. The vehicle speed is estimated using algorithm based on background modeling and blob detection. Absolute differencing method isolates foreground and background efficiently. The blob detection technique provides coordinates in the form of centroid resulting in the display of the vehicle movement. The performance of the system has been tested on videos captured on typical Indian highways using AXIS IP camera. Also, the algorithm is tested in the lab setup. A Raspberry Pi 3 processor and IP camera were used as hardware for this project. To overcome the shortcoming of formerly recommended methodologies, here vehicle speed estimation is done in Python using OpenCV. The speed is calculated by mathematical formulation based on the distance traversed by vehicle. The error in the speed estimation is observed to be 9 %.