基于蒙特卡洛模拟的虚拟测试场景生成方法
A Method for Generating Virtual Test Scenario Based on Monte Carlo Simulation
姚明菊 1李志远 1羊雪玲 1马军生2
作者信息
- 1. 吉利学院智能科技学院,四川 成都 610000
- 2. 国防科技大学信息通信学院,湖北 武汉 430000
- 折叠
摘要
为了提高用户在虚拟测试场景中的体验感,提高虚拟测试场景的真实度,提出基于蒙特卡洛模拟的虚拟测试场景生成方法.采用Meanshift分割方法分割输入图像,并对分割后的图像块实施融合,标记图像区域,获取对应的深度图,融合上述深度图获得虚拟测试场景的深度图.采用透射率估计方法融合虚拟测试场景的深度图,通过梯度融合方法融合归一化处理低照度图像,将支持向量机学习方法引入场景生成过程中,控制场景生成过程.采用蒙特卡洛模拟方法消除场景中存在的噪声斑点,优化生成的虚拟测试场景.实验结果表明,研究方法的场景生成误差较小,生成的虚拟测试场景具有较高的逼真度和真实感.
Abstract
In order to improve the user experience in the virtual test scene and the fidelity of virtual test scene,this article puts forward a method of generating virtual test scene based on Monte Carlo simulation.At first,Meanshift segmentation method was adopted to segment the input image,and then the image blocks were fused.Meanwhile,the image regions were marked to obtain depth maps correspondingly.After that,these depth maps were integrated to form a depth map of virtual test scene.Moreover,transmittance estimation method was adopted to fuse the depth maps of virtual test scene,and then the gradient fusion method was adopted to fuse and normalize the low-light images.More-over,the support vector machine learning method was introduced into the scene generation to control the process.Fi-nally,the Monte Carlo simulation method was used to eliminate noise speckles in the scene and thus to optimize the generated virtual test scene.The experimental results show that the scene error generated by the proposed method is small,and the generated virtual scene has high fidelity and realism.
关键词
蒙特卡洛模拟/透射率估计方法/梯度融合方法/虚拟测试场景生成Key words
Monte Carlo simulation/Transmittance estimation/Gradient fusion method/Virtual test scene gener-ation引用本文复制引用
出版年
2024
NETL
NSTL
万方数据