有没有办法缝合16位图像 [英] Is there a way to stitch 16 bit images

问题描述

大家好，



我正在进行一个图像拼接项目，需要将两个16位灰度图像拼接成单个图像。但这就是问题所在？



问题

您知道没有合适的图书馆（at-）至少在开源中，我认为）可以缝合16位图像，没有任何损失。我已经尝试过EmguCv，其他github项目等等...... Emgu CV缝针但不适用于16位图像。当我保存图像时，它会转换为8位图像。



其他github项目是一样的:(



无处不在，答​​案是否定



我的尝试：



你可以说有一个名为 Hugin 的软件很有前途。是的我明白了，但我不能用它这就是问题。



并且针脚应该自动执行而无需任何用户干预。



所以，

我很多时候都遇到过这个链接，这是一篇关于图像拼接的精彩文章。



使用Accord.NET进行自动图像拼接 [ ^ ]



仔细观察甚至哈哈这里写的rris算法再次支持最多8位。我的问题是，有没有办法让它适用于16位？



最后，

在无助的情况下，我可以只有回头看看这里的开发人员可以帮助我。而且我不想要美味的勺子饲料。我只需要从你的角度来看一个可以解决我的问题的想法，但不仅仅是算法或其他东西。



任何代码样本都会更好。 ：）

解决方案

嗯，有一个应用程序可以做到这一点：



图像复合编辑器 - Microsoft Research [ ^ ]

我不是这方面的专家，但似乎你的情况很少见，如果是扫描原始位图，你可以直接去创建自己的拼接位图（这不是很难），可以参考;



位图初学者指南 [ ^ ]



但是如果你已经有一个压缩文件，我认为你不能避免在缝合时丢失更多的信息。

Hi guys,

I am working on an image stitching project which requires two 16 bit gray-scale images to be stitched to a single image. But that's what the problem is?

Problem
You know that there is no proper libraries (at-least in opensource I think) which can stitch 16 bit images and without any loss. I have tried EmguCv, other github projects etc and etc... The Emgu CV stitches but not for 16 bit images. When I save the image, it converts into an 8 bit image.

The other github projects are just the same :(

"Everywhere I go, the answer is No"

What I have tried:

You can say that there is a software called Hugin which is promising. Yes yes I understand, but I can't use it and that's the problem.

And also the stitch should perform automatically without any user intervention.

So,
I came across this link a lot of times and its a great article on image stitching.

Automatic Image Stitching with Accord.NET[^]

Looking this on closer look even the harris algorithm written here supports up to 8 bit again. My question is, Is there a way to make it work for the 16 bit?

And finally,
In an helpless situation I can only turn back and see the developers here who can help me with it. And I don't want a delicious spoon feed. I just need an idea from your perspective which can solve my problem but not just an algorithm or something.

Any code samples would be even better. :)

解决方案

Well, there's an app that does it:

Image Composite Editor - Microsoft Research[^]

I am not expert in this matter, but seems your case is rare, if it is scanned raw bitmap, you can directly go for create your own stitched bitmap(which is not very hard), may refer to;

A Beginners Guide to Bitmaps[^]

But if you already have a compressed file, i think you cant avoid loosing some more information when stitch.

这篇关于有没有办法缝合16位图像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！