计算机视觉中的多视图几何
Multiple view geometry in computer vision
Richard Hartley · 初版 2003 年
3人推荐
6位读者提及
6次提及
计算机视觉中的一个基本问题是根据现实场景的多幅图像来理解其结构。作者涵盖了相关的几何原理，以及采用 camera projection matrices、fundamental matrix 和 trifocal tensor 对其进行的代数表示。
Hacker News 上怎么说
正在加载完整评论……
如果你对更深入的数学和 3D Geometry 感兴趣，该领域的圣经是 Hartley 和 Zisserman 写的《计算机视觉中的多视图几何》（Multiple View Geometry in Computer Vision）。Szeliski 的书非常偏实战，但比起严谨的推导，它更像是一本“公式堆砌+文献综述”的书。很多 3D Geometry 算法都是建立在纯粹漂亮的数学推导之上的……
If you are interested in more maths and 3D Geometry, the bible of the field is "Multiple view geometry in computer vision" by Hartley and Zisserman. Szielski's book is very hands on, but is it more of a formulas + litterature review book than a clean derivation book. A lot of 3D Geometry algorithms are based on clean derivations…
我买纸质技术书是因为我喜欢在上面随手记笔记，方便在阅读时理清概念。比如我就有一本 Hartley 和 Zisserman 写的《计算机视觉中的多视图几何》，写得极好，空白处画满了我在尝试彻底搞懂（grok）它时留下的草图和笔记。能随手翻翻书、复习那些完整的概念，感觉还挺棒的……
I buy physical tech books because I enjoy writing notes in them to keep track of concepts while reading. For example I have a copy of the excellent Multiple View Geometry in Computer Vision by Hartley and Zisserman filled with little sketches and notes I took in the margins while trying to grok the book, and it’s kinda nice to be able to thumb through the book and review concepts complete…
……打包封装得很好。这就意味着，不像某些网课那样，我能得到足够具体的反馈，从而知道自己的心智模型（mental model）对不对。这门课还用到了 Hartley 和 Zisserman 的书《计算机视觉中的多视图几何》。虽然这是一本非常详尽、绝对值得买的书，但对自学者来说确实有点劝退。把这门课和书结合起来看，刚好能互相平衡一下。
…well-packaged. That means that unlike some online courses, I've been able to get enough tangible feedback to know whether my mental model is correct. The course also uses Hartley and Zisserman's book, Multiple View Geometry in Computer Vision. While it's a very comprehensive book that's definitely worth owning, it's also kind of intimidating for self-learners. The course and the book together help to balance each other.
完全没错，我一看到那篇博文也立刻想到了 2d3。我是因为那本（如果允许我评价的话，非常宝贵的）书《计算机视觉中的多视图几何》（http://www.amazon.com/Multiple-View-Geometry-Computer-Vision... ）才了解到这家公司的。我没记错的话（IIRC），这家公司是由 Andrew Fitzgibbon（http://research.microsoft.com/en-us/um/people/awf/）掌舵的，不过现在看来觉得有点奇怪，因为他……
Exactly, I also though of 2d3 as soon as I have seen that blog post. I came to know the company since it is linked by the (very valuable, if I may) book "Multiple View Geometry in Computer Vision", http://www.amazon.com/Multiple-View-Geometry-Computer-Vision... . IIRC the company is run by Andrew Fitzgibbon http://research.microsoft.com/en-us/um/people/awf/ , which now seems odd to me since he's…
> 《计算机视觉中的多视图几何》，Richard Hartley 和 Andrew Zisserman 著，[117]（这里有一些样本章节，CVPR 教程可以看这里）呵，大约 10 年前我读了这本书，搞懂了几件事：* 三角测量（triangulate）……
> Multiple View Geometry in Computer Vision, Richard Hartley and Andrew Zisserman, [117] (some sample chapters are available here, CVPR Tutorials are available here) Heh, about 10 years ago I read that book and figured out a few things: * triangulate…
每当我看到类似的东西，我就会想起《计算机视觉中的多视图几何》的前言，很多这种“巧妙的数学”都在那里讲过。从本质上说，十年前被认为无解的问题，得益于几何学的发展，现在不仅能解，而且还相当简单。只要稍微……
Whenever I see something like this I'm reminded of the introduction to "Multiple View Geometry in Computer Vision", which is where a lot of this "clever math" is described. Essentially, problems considered unsolvable ten years ago are not only solvable, but simple, due to advances in geometry. With just a little…