Facing 'libavfilter' not found error while doing cmake for NVIDIA-DALI library on ubuntu version

– Installed OpenCV 3.4.4 which NVIDIA DALI recommended.
– below link for NVIDIA DALI reference.
https://docs.nvidia.com/deeplearning/sdk/dali-developer guide/docs/quickstart.html

–below link for OpenCV installation:
Ubuntu 18.04: How to install OpenCV - PyImageSearch

—NVIDIA DALI using below versions:
-DALI version: 0.10.0dev -Found PythonInterp: /usr/bin/python (found version “2.7.15”) -Found CUDA: /usr (found suitable version “9.1”, minimum required is “8.0”) -Found OpenCV: /usr/local/include;/usr/local/include/opencv (found suitable version “3.4.4”, minimum required is “2.0”)

–This is the error(No package ‘libavfilter’ found) i am getting.

Found CUDA: /usr (found suitable version “9.1”, minimum required is “8.0”)
-Building WITHOUT nvJPEG
-Building WITHOUT JpegTurbo
-Found OpenCV: /usr/local/include;/usr/local/include/opencv (found suitable version “3.4.4”, minimum required is “2.0”)
OpenCV libraries:
opencv_core;opencv_imgproc;opencv_imgcodecs

  • Found PythonLibs: /usr/lib/x86_64-linux-gnu/libpython2.7.so
    -pybind11 v2.2.4
    -Found Protobuf: /usr/lib/x86_64-linux-gnu/libprotobuf.a;-lpthread (found suitable version “3.0.0”, minimum required is “2.0”)

  • Enabling TensorFlow TFRecord file format support

  • Found PkgConfig: /usr/bin/pkg-config (found version “0.29.1”)

  • Checking for module ‘libavformat’

  • Found libavformat, version 58.20.100
    -Checking for module ‘libavcodec’

  • Found libavcodec, version 58.35.100
    -Checking for module ‘libavfilter’
    No package ‘libavfilter’ found

CMake Error at /usr/share/cmake-3.10/Modules/FindPkgConfig.cmake:415 (message):
A required package was not found
Call Stack (most recent call first):
/usr/share/cmake-3.10/Modules/FindPkgConfig.cmake:593 (_pkg_check_modules_internal)
cmake/Dependencies.cmake:176 (pkg_check_modules)
CMakeLists.txt:45 (include)
-Configuring incomplete, errors occurred!

Hello, did you solved this issue?

Hi,
Sorry for the late response, we expect to get the questions on GitHub directly.
I recommend using dockerized build according to this guide.
You may lack the FFmpeg installed. By the way, we are using a custom OpenCV build with CUDA disabled as we don’t want anything running on the GPU when the user asks to use CPU based implementation backed up by OpenCV.