Robust Vehicle Detection and Counting Algorithm Employing a Convolution Neural Network and Optical Flow.

Ahmed Gomaa,Moataz M Abdelwahab,Mohammed Abo-Zahhad,Tsubasa Minematsu,Rin-Ichiro Taniguchi

doi:10.3390/s19204588

Abstract

Automatic vehicle detection and counting are considered vital in improving traffic control and management. This work presents an effective algorithm for vehicle detection and counting in complex traffic scenes by combining both convolution neural network (CNN) and the optical flow feature tracking-based methods. In this algorithm, both the detection and tracking procedures have been linked together to get robust feature points that are updated regularly every fixed number of frames. The proposed algorithm detects moving vehicles based on a background subtraction method using CNN. Then, the vehicle’s robust features are refined and clustered by motion feature points analysis using a combined technique between KLT tracker and K-means clustering. Finally, an efficient strategy is presented using the detected and tracked points information to assign each vehicle label with its corresponding one in the vehicle’s trajectories and truly counted it. The proposed method is evaluated on videos representing challenging environments, and the experimental results showed an average detection and counting precision of 96.3% and 96.8%, respectively, which outperforms other existing approaches.

Highlights

Detecting and counting vehicles on the road are very important tasks for the traffic information analysis that can be used in traffic control and management to ensure a safe transportation system.Recently [1], vision-based vehicle detection and counting using image processing techniques provide more advantages than traditional intelligent transportation techniques [2], like microwave or magnetic detectors
Vehicle detection can be categorized into two groups [1]: detection methods based on vehicle appearance, and detection methods based on vehicle motion
A new and robust vehicle detection and counting approach was proposed by developing a three-step approach

Summary

Introduction

Detecting and counting vehicles on the road are very important tasks for the traffic information analysis that can be used in traffic control and management to ensure a safe transportation system. [1], vision-based vehicle detection and counting using image processing techniques provide more advantages than traditional intelligent transportation techniques [2], like microwave or magnetic detectors. An important stage in vehicle detection and counting is the elimination of the static background from the moving objects in a challenging environment. The most recent studies in intelligent transportation systems focus on vehicle detection [3,4,5,6,7,8,9,10,11,12].

Methods

Results

Conclusion