转换深度Kinect以查找真实世界坐标 [英] Convert Depth Kinect to find the real world coordinates

查看:167
本文介绍了转换深度Kinect以查找真实世界坐标的处理方法,对大家解决问题具有一定的参考价值,需要的朋友们下面随着小编来一起学习吧!

问题描述

大家好,

我通过互联网搜索使用Kinect传感器查找对象的真实世界坐标,但是我仍然无法理解如何做到这一点。

I searched over the Internet for finding real-world coordinates of an object using the Kinect sensor, however I still can not understand how to do it.

 

我正在开发c#程序,用于跟踪在空间中移动的对象。

I'm developing c# program that track an object that moves in the space.

使用Kinect传感器深度,我找到了从相机到物体的实际距离。

Using the Kinect sensor depth, I found the actual distance from the camera to the object.

现在,我想要找到对象的真实世界坐标(x,y)(深度为z)。

Now, I want to find the real-world coordinates (x,y) of the object (while the depth is z).

我找到了已计算的距离(mm)的以下公式。

I found the following formula for the distance(mm) that is already calculated.




  (对象的实际高度(mm))未知,而其他可以从场景计算。

The  (real height of objec(mm)) is unknown, while the others can be calculated from the scene.

让我们假设我的图像为640 * 480结果,所以图像高度是480px? 

Lets assume I have an image of 640*480 resulotion, so the image height is 480px ? 

如何计算Kinect焦点l ength,它是一个已知的参数吗?

How to calculate Kinect focal length, Does it a known pararmeter ?

有人可以给出一个详细的例子(而不是转发到只用文字解释的论文) 用于计算距离时参数的方法该对象已经确定。

Can someone give a detailed example (and not forwarding to a paper that explain only in words)  for the way to calculate the parameters when the distance to the object is already determined.

谢谢 

Yoni  

推荐答案

你可以获得ColorCamera的焦距,但如果你当帧被映射到深度空间时,已经考虑了深度点。

You can get the focal length of the ColorCamera, but if you are looking at depth points. This is already taken into account as the frame is mapped to depth space.

http://msdn.microsoft.com/en-us/library/microsoft.kinect.colorcamerasettings_members.aspx

另一个选项是使用坐标映射将深度点映射到Skeleton(世界)空间,以获得相对于传感器的世界空间坐标系中的点

The other option is to map the depth point to Skeleton(world) space using coordinate mapping to get the points in a world space coordinate system relative to the sensor

http://msdn.microsoft.com/en-us/library/microsoft.kinect.coordinatemapper.mapdepthpointtoskeletonpoint.aspx

我能想到的唯一另一个选择是将相机校准到已知的世界点。有几个线程讨论做这样的事情。这将是SDK外部的东西。

The only other option I can think of is doing your own calibration of the camera to known world points. There are several threads that discuss doing something like this. That would be something external to the SDK.


这篇关于转换深度Kinect以查找真实世界坐标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IT屋!

查看全文
登录 关闭
扫码关注1秒登录
发送“验证码”获取 | 15天全站免登陆