Visual Vehicle Counting and Classification System

  • Manu Singh, Rajnesh Singh, Vibhash Yadav
Keywords: Video analysis, Adaptive background estimation, Traffic Planning, traffic counting, turning movement counts

Abstract

Motion capturing involves the Identification of objects in motion and plotting their activity by analyzing them. When considered concerning a video sequence, motion capture can be defined as a procedure of unmasking things in action in lines of frames using specific and proficient digital image processing techniques. Methodology: To ingress need contrasting time for trial and execution shows the difference in pace and memory requisites.   Results: The result obtained shows the performance of the algorithms and models in the given conditions. It also reflects the best suitable environment for the techniques. The work was accomplished by testing the algorithms for numerous sequences of the input video. The identification rate analysis gives the identification rate of foreground pixels for various colors divergent from the background model. Conclusions: The objective of computer vision is to imitate the human eye utilizing advanced digital images through three principle handling parts that are executed consistently, i.e., acquisition of pictures, picture processing, and picture investigation and comprehension. It has gained a lot of attention from researchers in an enormous field. The basic purpose of this study was to prospecting tracking and computer-aided object detection techniques. For this purpose, several methods were observed, studied critically, and used. Originality: In the study, an improved frame differencing method for tracking an object is briefly reviewed, and an enhanced version of the algorithm was also implemented using MATLAB. The proposed algorithm was tested over distinct video sequences. It was observed that the objects in motion were identified with minimal error rate compared to a traditional frame differencing method. Limitations: One of the major challenges in tracking objects in action is to design algorithms or techniques for monitoring the things present in disrupted or random videos, like videos attained from broadcast news networks or home videos. These videos contain noise, and some of them may be unstructured, compressed, and denominationally having edited pieces obtained in different ways by moving cameras.

Published
2021-10-28
How to Cite
Vibhash Yadav, M. S. R. S. (2021). Visual Vehicle Counting and Classification System. Design Engineering, 7775- 7789. Retrieved from http://thedesignengineering.com/index.php/DE/article/view/5820
Section
Articles