虚拟现实场景图像点云大数据展示仿真

Virtual Reality Scene Image Point Cloud Big Data Display

虚拟现实技术可以利用计算机图像建模技术构建虚拟空间,通过视觉效果增强现实效果。但是,由于虚拟场景特征点较多,且密度较大,复杂场景的三维重建和展示的难度较高。提出基于图像点云大数据的虚拟现实场景展示方法。利用Kinevt v2设备获取虚拟现实场景两帧图像及点云数据,通过对点云数据的滤波处理,优化虚拟现实场景重构精度。基于滤波处理,提取虚拟现实场景两帧图像的特征点,并匹配特征点,构建三维场景重建模型。将三维场景重构模型与配准后的点云数据、场景颜色信息相融合,完成模型的可视化,实现虚拟现实场景的展示。实验结果证明了研究方法展示的虚拟现实场景特征信息更完整,在图像滤波处理前后,研究方法的虚拟现实场景特征点提取效率均更高,总耗时在7ms以内,且研究方法应用下图像处理器占用的空间内存低于20MB。

At present,virtual reality technology can use modeling technology to build a virtual space and enhance the reality effect.However,due to the large number of feature points and high density of virtual scenes,it is difficult to reconstruct and display the 3D complex scene.In this article,a method of displaying virtual reality scenes was proposed based on point cloud big data.At first,we used Kinevt v2 to obtain two images of virtual reality scenes as well as point cloud data,and then filtered the point cloud data,thus optimizing the reconstruction accuracy of virtual reality scenes.Based on the filtering,we extracted the feature points in the two images and matched them.After that,we built a 3D scene reconstruction model.Moreover,we integrated the model with the registered point cloud data and color information to complete the visualization of the model.Finally,we realized the display of virtual reality scenes.Following conclusions can be drawn from experimental results.The feature information of virtual reality scenes displayed by the proposed method is more complete.Before and after image filtering,the extraction efficiency of virtual reality scene feature points is higher,and the total time is less than 7ms.In addition,the space memory occupied by the image processor is less than 20MB.