Author(s)

Likai Wang¹, Weiwei Liu¹, Dan Li^2,*

Affiliation(s)

¹ Traffic police detachment of Xuzhou Public Security Bureau Xuzhou, China;

² Xuzhou University of Technology, Xuzhou, China.

Corresponding Author

Dan Li, email: lidanonline@163.com

ABSTRACT

At present, with the development of society, road vehicles are increasing, and the highway carrying capacity is relatively insufficient, resulting in serious traffic congestion and frequent accidents. How to select effective tracking methods and extract invariant features needs to be improved. Aiming at some problems existing in urban road vehicle tracking, this paper proposes a vehicle tracking algorithm in complex environment based on continuous adaptive Meanshift algorithm and integrating local invariant features. The improved algorithm can adaptively update different feature weights when dealing with vehicle deformation, frost and fog weather, background noise interference, light variation, occlusion and other problems. The complementarity between features is insufficient, and has good robustness to complex environment.

KEYWORDS

Traffic monitoring; target tracking; Camshift; feature fusion.

CITE THIS PAPER

Wang Likai, Liu Weiwei, Li Dan. Research on traffic target tracking method in complex environment based on machine vision. Eurasia Journal of Science and Technology. 2022, 4(1): 29-34.

REFERENCES

[1] Gao, J. . "Analysis of the Application of TrafficMonitoring System in Traffic Unraveling during SubwayConstruction." Electrical Technology of IntelligentBuildings (2020). 

[2] Wang, F. , M. Liu , and S. Wang . "Kalman filtertracking of sequence spot centroid ablated by femtosecondlaser." Microwave and Optical Technology Letters63.2(2021). 

[3] Kumar, D. ."Hybrid Unscented Kalman Filter withRare features for Underwater Target tracking using PassiveSonar Measurements." Optik - International Journal forLight and Electron Optics 226.no. 3(2021):165813. 

[4] Du, S. , and Q. Deng . "Unscented Particle FilterAlgorithm Based on Divide-and-Conquer Sampling forTarget Tracking." Sensors 21.6(2021):2236. 

[5] Yang, J. , et al. "Particle filter algorithm optimized bygenetic algorithm combined with particle swarmoptimization." Procedia Computer Science187.4(2021):206-211. 

[6] Fang, C. , et al. "Comparative study on poultry targettracking algorithms based on a deep regression network." Biosystems Engineering 190(2020):176-183. 

[7] Hu, B. , G. Chen , and Q. Liu . "UAV attitude anglemeasurement system based on machine vision." 2020 IEEE5th Information Technology and Mechatronics EngineeringConference (ITOEC) IEEE, 2020. 

[8] Kumar, M. , and S. Gupta . "2D-human facerecognition using SIFT and SURF descriptors of face'sfeature regions." The Visual Computer 37.11(2021). 

[9] Parashivamurthy, R. , C. Naveena , and Y. Kumar . "SIFT and HOG features for the retrieval of ancientKannada epigraphs." IET Image Processing (2021). 

[10] Lv, W. , et al. "Research and Application ofIntersection Clustering Algorithm Based on PCA FeatureExtraction and K-Means." Journal of Physics: ConferenceSeries 1861.1(2021):012001 (7pp). 

[11] Zhang, Y. , D. Xiao , and Y. Liu . "AutomaticIdentification Algorithm of the Rice Tiller Period Based onPCA and SVM." IEEE Access PP.99(2021):1-1.