Research And Application Of Key Technologies Of Real-Time Realistic Enhancement System For Virtual Render Scene

Computer graphics is a huge research field of computer science,which can help in the production phase of many fields such as design field,mechanical manufacturing,architecture and so on.The virtual reality technology produced by the intersection of graphics and other disciplines can also assist professional training such as power grid maintenance and military training.In recent years,with the development of technology,users in some fields have higher requirements for the quality of the virtual scene,and require the system to quickly give feedback to their input.Based on the principle of physics-based rendering,this paper analyzes the problems and limitations of the existing methods.Then analyzes the scene characteristics,so as to do specific processing in the lighting calculation and achieve the purpose of improving the realism of the rendering result.Secondly,since the requirement of real-time rendering applications is to be able to receive user input,process information and give feedback quickly.This paper also improves the algorithm execution efficiency and the rendering system’s whole running speed to increase the overall rendering rate of the application system.The first work of this paper proposes a global illumination enhancement method based on screen space depth data analysis to improve the limitation of existing global illumination algorithms.It can analyze the characteristics of the current scene by analyzing the intermediate data products of the rendering pipeline.This method considers the position of the object in the scene when calculating the shading result,combines the auxiliary information such as the depth buffer generated in the preprocessing stage of the scene data,compares the shading position and the depth of the surrounding sampling points,determines the shading strategy,so as to achieves the goal of improving the realism of the rendering result.The second work proposes a dynamic trade-off algorithm for real-time detection of system state.This method records the average rendering frame rate during the running process of the rendering scene as the basis for measuring the current system performance,and updates the detection values of performance influencing factors in real time.Therefore,dynamically adjusting the parameter variables of the soft shadow generation algorithm to achieve a balance between image quality and running speed.The last work proposes a state division method based on data multiplexing according to the hardware performance bottleneck of the rendering system and the depth information that is the basis of various rendering algorithms.This method divides the state of objects that need to be rendered in the scene,analyzes the data that remains unchanged under certain conditions,and reuses some information that has been calculated to reduce the cost of the subsequent rendering process.That is,by reducing redundant calculations,the overall performance of the system can be further improved while maintaining the rendering quality.