图像识别从屏幕中提取各种形状 [英] Image recognition to extract various shapes from screen

问题描述

大家好，



我有一个想法，我想实践，不知道从哪里开始文档。



照相机可以识别脸部的方式相同，我想创建一个识别来自谷歌地图，雅虎地图等的火车轨道的工具。来源没关系，肯定会涉及到一些数学问题。



有些satelitte图像不够清晰，我的软件需要把它们放在一起即使是轨道的模糊/阴影部分。

有些事实是有用的，比如火车轨道大部分时间都是相互比较的，跟踪点通常具有相同的形状（特别是在分拣场中，它很可能是所有点具有相同的半径）。

就像火车建模一样，预定义的轨道部分已经知道长度和半径，因此软件应该通过可用的形状并返回最佳匹配。



欢迎任何想法。再一次，我发布在这里是为了缩小搜索正确的图像识别技术，形状匹配或应该调用适用于我的问题的技术。



万分感谢，

Dani

Hi all,

I have an idea that I want to put in practice and don''t know where to start with documentation.

The same way photo cameras can recognize face, I want to create a tool that recognize train tracks from Google maps, Yahoo maps, .. so on. The source doesn''t matter, it should be some mathematics involved for sure.

Some satelitte images are not clear enough and my software will need to put the pieces together even for blured/shaded portions of the tracks.
Some facts are to be useful such as train tracks are most of the time paralel to each other, track points usualy have the same shape (especially in a sorting yard it''s highly possible that all points have the same radius).
Just as for train modelling, there are predefined track portions with already known length and radius so the software should go through available shapes and return the best match.

Any ideas are welcomed. Once again, I am posting here in order to narrow the search for the right techniques for image recognition, shape matching or what ever should be called that applies to my problem.

Thanks a million,
Dani

推荐答案

我曾经需要做的任何图像处理，我搜索了SIGRAPH [ ^ ]以及我的搜索词，并找到了解释数学的论文，然后我从那里编写了代码。这是一个非常模糊的请求，我不知道有多少人可以帮助你，除非有人知道现有的图书馆。正如您所见，面部识别是非常可能的，但这不是一项微不足道的任务。

Any image processing I''ve ever needed to do, I searched for SIGRAPH[^] as well as my search terms, and found a paper that explained the math, then I wrote the code from there. This is a pretty vague request, I''m not sure how much more anyone can help you, unless someone knows of an existing library. Face recognition is very possible, as you''ve seen, but it''s not a trivial task.

这篇关于图像识别从屏幕中提取各种形状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！