抓住单个深度框架的问题 [英] Problem to grab a single Depth frame

问题描述

我们开发了一个采用单个DepthColor帧的应用程序。因为我们使用了许多Kinect传感器，所以我们必须在未使用的kinect上保持红外线。



我们正在用C＃开发我们的解决方案。要拍摄一帧，我们首先打开红外线，然后我们用事件AllFramesReady等待第一帧。问题是我们收到的第一帧并不总是好的。有时
没关系，但它可以有一些无效的像素。无效像素为黑色。通常，帧可以具有这些无效像素的10-80％之间。我们认为红外线不是完全活跃的。



我们试图等待一些帧。第三帧总是好的。问题是我们有时间问题。等待3帧可能需要90毫秒（30毫秒帧）。我想知道我们现在有什么问题，我们可以在没有无效像素的情况下抓取一个
的帧。



$
这里有一些我们想要抓住的框架例子：

好​​框架：

http://social.msdn.microsoft.com/Forums/getfile/299348

错误的框架：

http://social.msdn.microsoft.com/Forums/getfile/299349



$
感谢您的帮助，



Maxime 

解决方案

您的图像没有显示。

深度框架中会有噪音，您需要分析您可能从自然阳光和其他来源获得的红外干扰量的区域。通常情况下，您将分析几个帧的深度，或者在特定区域上关注
，并在该区域中假设未知深度值。

< a href ="http://social.msdn.microsoft.com/Forums/en-US/5cda686f-1af6-4f2b-befb-d00c57f8a98e/how-to-compare-two-depth-frame-data"> http：/ /social.msdn.microsoft.com/Forums/en-US/5cda686f-1af6-4f2b-befb-d00c57f8a98e/how-to-compare-two-depth-frame-data

最新的SDK确实提供了一种方法，可以从您不想从中提取数据的其他传感器关闭IR。这会有所帮助，但同样，你的图像中仍会有一定程度的噪音，这将是你在代码中如何处理
的问题。

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

Hi,

We have developped an application who take a single DepthColor Frames. Because we use many Kinect Sensors, we have to keep the infrared off on the unused kinect.

We are developping our solution with C#. To take a frame, we first turn on the infrared, then we wait the first frame with the event AllFramesReady. The problem is the first frame we receive is not always good. Sometime it's ok, but it can have some invalide pixel. The invalid pixels are black. Generally, the frame can have between 10-80% of these invalid Pixels. We think that the infrared is not completly active.

We tried to wait some frame. The third frame is always good. The problem is we have a timing problem. Wait 3 frames can take 90ms (30 ms by frame). I would like to know if we have an issue to now when we can grab a frame without the invalid pixels.


Here some exemples of frame we want to grab:
A good frame:
http://social.msdn.microsoft.com/Forums/getfile/299348

A bad frame:
http://social.msdn.microsoft.com/Forums/getfile/299349


Thanks for your help,

Maxime 

解决方案

Your images are not showing up.

There is going to be noise in the depth frame and you will need to analyze the area for the amount of IR interference you might be getting from natural sunlight and other sources. Typically you are going to analyze depth over a couple of frames or focus on a particular region and make an assumption amount unknown depth values in that area.

http://social.msdn.microsoft.com/Forums/en-US/5cda686f-1af6-4f2b-befb-d00c57f8a98e/how-to-compare-two-depth-frame-data

The latest SDK does provide a way to turn off the IR from the other sensors you are not trying to pull data from. This will help a bit, but again, you will still have some level of noise in the image and it will be a matter of how you are going to treat this in your code.

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

这篇关于抓住单个深度框架的问题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！