Construction of fire identification alarm platform based on deep learning framework----OpenCV3.1.0 reading DAV video data error

When building the deep learning-caffe framework, read the DAV video data using OpenCV3.1.0, and the parsing h264 data errors occur:

[H264] top block unavailable for requested intra4x Mode-1 at 44 0

[H264] Error while Decodin MB 0, ByteStream 6336

......

At this time, the display H264 parsing failed, image flower screen, resulting in fire recognition failure.

C:\Program Files (x86) \opencv\opencv\build\x64\vc14\bin\opencv_ffmpeg310_64.dll parsing H264 failure, it will OpenCV3.0.0 under the Opencv_ The Ffmpeg300_64.dll file is modified to Opencv_ffmpeg310_64.dll and placed in the appropriate location. H264 data in DAV format can be successfully parsed