Suggest a Mathematical Model to Measure the Speed of Vehicles Via Video


  • University of Thi-Qar, College of Education for pure Sciences, Department of Computer Science
  • University of Thi-Qar, College of Education for pure Sciences, Department of Computer Science



ACF, Euclidean distance, fps.


Identification of vehicles violating a certain speed by measuring the speed of vehicles in a nonintrusive
manner with low cost and good results is very important, as it directly contributes to helping the
traffic police instead of using radar, lidar or other expensive devices, the use of video analysis to estimate
the speed of vehicles involved in traffic accidents is becoming increasingly common. In most cases, the
estimate is based on in situ reference measurements or data derived using photogrammetry techniques.
In this research, the suggested method estimates the average vehicle speed via computing the
distance between consecutive tires using the "Pixel Image Scale Coefficient" and converts the results to
accurate measurements. Using an installed camera or a mobile phone device at video frames per second
(fps of 30 frames per second, a mathematical function to identify the speed of a passing vehicle
predicated on its motion pattern vector and transform that speed from measurements of pixels in
succeeding frames to realistic measurements is presented. Three steps were used in this work to
implement the technique: In the first stage, a vehicle is distinguished from other objects utilizing artificial
intelligence, particularly the Aggregated Channel Features(ACF) vehicle detectors algorithm. In the
second step, automobiles are monitored by tracking successive video frames and then the third step,
which is the working core of this The research, is to calculate the speed of the detected vehicles in
successive video frames through the proposed mathematical equation that mainly depends on the product
of the obtained Euclidean distance, and then convert this result to what is proportional to the realistic
measurements of speed, which really is km/h.
The project work was produced on a 64-bit system with an Intel Core i7 processor using Matlab

R2021a. The software was used to analyze a series of locally captured movies that were used as the data
source for a data set used to calculate the average vehicle speed. The proposed program's results were
compared with the established driving velocity in order to verify the findings. A simulation of a vehicle
moving at a known speed was employed. The mistake rate as a result of the vehicle speeds varied from 1
to 5 km/h.