Kinect Fusion和OpenGL之间的集成

问题描述

亲爱的，

我知道Kinect Fusion界面为可视化及其网格提供了重建体积的阴影图像。

我试图在OpenGL中检索和显示网格，但速度很慢。

主要问题是
INuiFusionReconstruction :: CalculateMesh 方法，每次网格检索花费大约1.2秒。

我想知道是否有更有效的方法来集成（可视化）由OpenGL中的KinectFusion重建的卷。

我想从点云到点音量然后回到网格我浪费了很多资源。

谢谢！

解决方案

CalculateMesh很慢，但作为替代方案，您可以使用高于1的voxelStep调用它。   

另一种方法是使用INuiFusionReconstruction :: CalculatePointCloud，然后锁定并复制点云图像框中的位，将采用这种格式，每像素六个浮点数：（posX，posY，posZ，normX，normY，normZ）。

获取该数据并将其发送到OpenGL以进行渲染。这种方法应该更快。

Dear all,

I know the Kinect Fusion interface provides the shaded image of reconstructed volume for visualization and its mesh as well.

I tried to retrieve and display the mesh in OpenGL but the speed was quite slow.

The main problem was INuiFusionReconstruction::CalculateMesh method which took about 1.2 seconds for each mesh retrieval.

I was wondering if there is any more efficient way to integrate (visualize) the volume reconstructed by KinectFusion in OpenGL.

As I guess from point cloud to volume and then back to mesh I am wasting a lot of resources.

Thank you!

解决方案

CalculateMesh is slow, but as an alternative you could call it with a voxelStep of higher than 1.  

An alternative would be to use INuiFusionReconstruction::CalculatePointCloud and then lock and copy the bits out of the point cloud image frame, which will be in this format, six float per pixel: (posX, posY, posZ, normX, normY, normZ).

Take that data and send it to OpenGL to render as you like. That approach should be faster.

这篇关于Kinect Fusion和OpenGL之间的集成的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！