Virtual Reality Assisted Visual Tracking Modeling of Machine Vehicles
This paper proposes a method for visual tracking modeling of machine vehicles assisted by virtual reality tech-nology,aiming to address the high cost of training tracking models in the complex real-world driving environment.Firstly,a virtual driving environment for machine vehicles is constructed using virtual reality technology,and image data from the virtu-al environment is collected to build a simulation dataset.Secondly,the network structure within the tracking model is opti-mized based on the performance of the virtual vehicle in the simulated driving environment and the predictive accuracy of the tracking model,among other metrics,to enhance the model's performance.Finally,the real-time visual tracking effect of the tracking model is achieved in the virtual environment.Experimental results show that compared to the traditional model and training methods,the improved model increases the tracking accuracy by 5.2%in the simulated environment dataset,and the tracking accuracy of the improved model in the real environment dataset is improved by 5.9%to 81.2%,and virtual reality ef-fectively assist in visual tracking modeling for machine vehicles.
万方数据
维普
