Load Captures in OpenCV

后端 未结 2 1698
清歌不尽
清歌不尽 2021-01-21 03:59

Which video formats can we use in OpenCV? Can anything in addition to AVI and load from camera be used? If these are the only supported formats, is a video converter required to

相关标签:
2条回答
  • 2021-01-21 04:24

    OpenCV "farms out" video encoding and decoding to other libraries (e.g., ffmpeg and VFW). Also, have a look at the highgui source directory to see all of the VideoCapture wrappers available (specifically pay attention to the cap_* implementations). AVI is merely a container, and really isn't that critical to what video codecs that OpenCV can read. AVI can contain several different combinations of video, audio, and even subtitle streams. See my other answer about this. Here is also a quick article explaining the differences between containers and codecs.

    So, if you're on Linux make sure ffmpeg supports decoding the video codec you are interested in processing. You can check what codecs your version of ffmpeg supports with the following command:

    ffmpeg -formats
    

    On Windows, you'll want to make sure you have plenty of codecs available to decode various types of video like the K-Lite Codec Pack.

    0 讨论(0)
  • 2021-01-21 04:34

    I'm not sure how up-to-date it is, but this OpenCV wiki page gives a good overview of what codecs are supported. If looks like AVI is the only format with decent cross-platform support. Your options are either to do the conversion using an external converter (like you suggest) or write code that uses a video library to load the image and create the appropriate cv::Mat or IplImage * header for the data.

    Unless you're processing huge quantities of video I suggest taking the path of least resistance and just converting the videos to AVI (see the above link for the details of what OpenCV supports). Just be careful to avoid lossy compression: it will wreck havoc with a lot of image processing algoritms.

    0 讨论(0)
提交回复
热议问题