How to check for successful install (including in virtualenv).

[aditjha]

Hello, thank you for your build script for the nano. I ran the script overnight and I think it finished building, however is there a way to check if it built successfully? And is there any way you know how to get this built opencv to work with a virtual environment on the nano, i ran the script from inside the virtual environment but it still doesn't find module cv2. Help would be super appreciated thank you!

currently in my /tmp/ folder i see a folder build-opencv, however I don't know if it is supposed to be like that.

[mdegans]

Hi, @aditjha

To check for a sucessful install:

$ opencv_version --verbose

You should see something like:

 $ opencv_version --verbose

General configuration for OpenCV 4.3.0 =====================================
  Version control:               4.3.0

  Extra modules:
    Location (extra):            /tmp/build_opencv/opencv_contrib/modules
    Version control (extra):     4.3.0

  Platform:
    Timestamp:                   2020-05-14T20:55:02Z
    Host:                        Linux 4.9.140-tegra aarch64
    CMake:                       3.10.2
    CMake generator:             Unix Makefiles
    CMake build tool:            /usr/bin/make
    Configuration:               RELEASE
...

To check if the python module is installed:

 $ python3
Python 3.6.9 (default, Apr 18 2020, 01:56:04) 
[GCC 8.4.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import cv2
v>>> cv2.__version__
'4.3.0'

To check if GPU support works in python:

>>> cv2.cuda.printCudaDeviceInfo(0)
*** CUDA Device Query (Runtime API) version (CUDART static linking) *** 

Device count: 1

Device 0: "Xavier"
  CUDA Driver Version / Runtime Version          10.20 / 10.20
  CUDA Capability Major/Minor version number:    7.2
  Total amount of global memory:                 15815 MBytes (16582963200 bytes)
  GPU Clock Speed:                               1.38 GHz
  Max Texture Dimension Size (x,y,z)             1D=(131072), 2D=(131072,65536), 3D=(16384,16384,16384)
  Max Layered Texture Size (dim) x layers        1D=(32768) x 2048, 2D=(32768,32768) x 2048
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total number of registers available per block: 65536
  Warp size:                                     32
  Maximum number of threads per block:           1024
  Maximum sizes of each dimension of a block:    1024 x 1024 x 64
  Maximum sizes of each dimension of a grid:     2147483647 x 65535 x 65535
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             512 bytes
  Concurrent copy and execution:                 Yes with 1 copy engine(s)
  Run time limit on kernels:                     No
  Integrated GPU sharing Host Memory:            Yes
  Support host page-locked memory mapping:       Yes
  Concurrent kernel execution:                   Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support enabled:                No
  Device is using TCC driver mode:               No
  Device supports Unified Addressing (UVA):      Yes
  Device PCI Bus ID / PCI location ID:           0 / 0
  Compute Mode:
      Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) 

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version  = 10.20, CUDA Runtime Version = 10.20, NumDevs = 1

Please note you must use cv2.cuda.foo instead of cv2.foo if you want your code to use the GPU. This means most code for opencv you find on the internet must be rewritten to use equivalents in the cuda module (if they exist) if you want GPU acceleration. Otherwise your code will use the CPU exclusively (OpenCV is an Intel supported product) and this will be very very slow on an Arm CPU like Nano.

A virtual environment will not include system python packages unless you specify it on creation with an option like --system-site-packages in the case of virtualenv. If for some reason this doesn't work you should be able to copy or link opencv from it's installation path to your virtual env's python path. Like:

ln -s /usr/local/lib/python3.6/dist-packages/cv2/ /path/to/venv/lib/python3.6/site-packages/

(this is untested personally but should work)

Easier if you want repeatability might be the docker images here. To get gpu support you must run with the runtime --nvidia flag when you docker run. There are tags for JetPack 4.3(jp-r32.3.1...) as well as 4.4 (jp-r32.4.2...).

Hope that answers your questions!

[Sreerag-ibtl]

I tried the symlink but it didn't worked. So, I copied the directory cv2 from usr/local/lib/python3.6/site-packages to [path to venv]/lib/python3.6/site-packages and that worked.

[beniroquai]

Hey, thanks for the information! Did anybody try that with a miniconda/miniforge installation?

After installing it using your script and linking it I get the following error when importing cv2:

(imswitch) uc2@uc2-desktop:~/miniforge3/envs/imswitch/lib/python3.7$ python
Python 3.7.12 | packaged by conda-forge | (default, Oct 26 2021, 05:43:52) 
[GCC 9.4.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import cv2
['', '/home/uc2/miniforge3/envs/imswitch/lib/python3.7/cv2/python-3.7', '/home/uc2/miniforge3/envs/imswitch/lib/python37.zip', '/home/uc2/miniforge3/envs/imswitch/lib/python3.7', '/home/uc2/miniforge3/envs/imswitch/lib/python3.7/lib-dynload', '/home/uc2/miniforge3/envs/imswitch/lib/python3.7/site-packages', '/home/uc2/Downloads/imswitch']
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/uc2/miniforge3/envs/imswitch/lib/python3.7/cv2/__init__.py", line 96, in <module>
    bootstrap()
  File "/home/uc2/miniforge3/envs/imswitch/lib/python3.7/cv2/__init__.py", line 86, in bootstrap
    import cv2
  File "/home/uc2/miniforge3/envs/imswitch/lib/python3.7/cv2/__init__.py", line 96, in <module>
    bootstrap()
  File "/home/uc2/miniforge3/envs/imswitch/lib/python3.7/cv2/__init__.py", line 23, in bootstrap
    raise ImportError('ERROR: recursion is detected during loading of "cv2" binary extensions. Check OpenCV installation.')
ImportError: ERROR: recursion is detected during loading of "cv2" binary extensions. Check OpenCV installation.
>>> 

My cmake is slightly modified:

configure () {
    local CMAKEFLAGS="
        -D BUILD_EXAMPLES=OFF
        -D BUILD_opencv_python2=OFF
        -D BUILD_opencv_python3=ON
        -D CMAKE_BUILD_TYPE=RELEASE
        -D CMAKE_INSTALL_PREFIX=${PREFIX}
#        -D CUDA_ARCH_BIN=5.3,6.2,7.2
#        -D CUDA_ARCH_PTX=
#        -D CUDA_FAST_MATH=ON
#        -D CUDNN_VERSION='8.0'
        -D PYTHON3_EXECUTABLE=~/miniforge3/envs/imswitch/bin/python
        -D PYTHON3_LIBRARY=~/miniforge3/envs/imswitch/lib/python3.7m.dylib
        -D PYTHON3_INCLUDE_DIR=~/miniforge3/envs/imswitch/include/python3.7m
        -D PYTHON3_PACKAGES_PATH=~/miniforge3/envs/imswitch/lib/python3.7/site-packages
        -D EIGEN_INCLUDE_PATH=/usr/include/eigen3
        -D ENABLE_NEON=ON
#        -D OPENCV_DNN_CUDA=ON
        -D OPENCV_ENABLE_NONFREE=ON
        -D OPENCV_EXTRA_MODULES_PATH=/tmp/build_opencv/opencv_contrib/modules
        -D OPENCV_GENERATE_PKGCONFIG=ON
#        -D WITH_CUBLAS=ON
#        -D WITH_CUDA=ON
#        -D WITH_CUDNN=ONF
        -D WITH_GSTREAMER=ON
        -D WITH_LIBV4L=ON
        -D WITH_OPENGL=OFF"

Any ideas?