Vxl is an open-source computer vision library. The project homepage isHttp://vxl.sourceforge.net/Is "a collection of C ++ Libraries
Designed for Computer Vision Research and Implementation ", it indicates that this library is a varietyCode. A little bit, there is also a bundler for SFM, it seems that it is indeed a lot of effort to collect, if you do not want to find it, directly use this library is also a good choice.
However, the biggest trouble for this library is that it only providesSource codeSo I had to compile it myself. Fortunately, it was compiled with cmake. Otherwise, I still don't know how much time it will take. It is reasonable to use cmake for compilation, but there are still some problems.
This article is the original blogger, reprint please note, this article address: http://blog.csdn.net/houston11235/article/details/8146344
The first problem is cmake. Do not place the source code in a path with Chinese characters. Otherwise, the configuration file cannot be opened.
The second problem is that this library does not support compiling into a dynamic link library and can only be compiled into a static library. Although the cmake configuration contains build_shared_library, however, if you select this option, you will encounter various errors that cannot open the Lib. This is described in the FAQ. For details, seeHttp://vxl.sourceforge.net/vxl-users-faq.html.
After compilation, you can start to use it. However, it seems that the code and library organization are not as organized as opencv. the header file and code are mixed together, And the include path is also cool when being added, let's talk about it later, and there may be unexpected gains.