Jetson 学习笔记(十一)：jetson agx xavier 源码编译ffmpeg(3.4.1)和opencv(3.4.0)

ZZY_dl

已于 2022-04-12 19:15:41 修改

阅读量3.1k

点赞数

分类专栏：嵌入式 Jetson-AI 文章标签： opencv 人工智能计算机视觉

于 2021-11-05 16:58:28 首次发布

本文链接：https://blog.csdn.net/m0_51004308/article/details/121165563

版权

嵌入式 Jetson-AI 专栏收录该内容

26 篇文章 20 订阅

订阅专栏

文章目录

编译需求
编译ffmpeg
编译opencv
测试
编译问题
附录

编译需求

gcc -----------------------------7.5.0
g++ ---------------------------- 7.5.0
ubuntu ----------------------- 18.04
python ------------------------ 3.6.9
ffmpeg ------------------------ 3.4.1
opencv ------------------------ 3.4.0

编译ffmpeg

下载网址：http://www.ffmpeg.org/releases/

安装依赖

sudo apt-get install -y nasm yasm cmake mercurial
sudo apt-get install -y libxcb-xfixes0-dev pkg-config texinfo wget zlib1g-dev
sudo apt-get install -y libtool libva-dev libvdpau-dev libvorbis-dev libxcb1-dev libxcb-shm0-dev
sudo apt-get install -y libfreetype6-dev libsdl2-dev libtheora-dev
sudo apt-get install -y autoconf automake build-essential git libass-dev
sudo apt-get install libx264-dev libx265-dev libnuma-dev libgtk-3-dev

安装ffmpeg

cd ffmpeg-4.3

sudo -s
# 安装了yasm和nasm 就不需要加–disable-yasm
./configure --enable-shared   --enable-libx264  --enable-gpl  --prefix=/usr/local/ffmpeg

make && make install

如果要make有缓存通过make clean即可

配置环境ffmpeg变量

sudo gedit /etc/profile
export LD_LIBRARY_PATH=$/usr/local/ffmpeg/lib:$LD_LIBRARY_PATH
source /etc/profile

建立软链接

cd /usr/bin
sudo ln -s /home/lqs/Downloads/ffmpeg-3.4.1/ffprobe ffprobe 
sudo ln -s /home/lqs/Downloads/ffmpeg-3.4.1/ffmpeg ffmpeg 
sudo ln -s /home/lqs/Downloads/ffmpeg-3.4.1/ffplay ffplay

查看ffmpeg

ffmpeg

如果要通过自己编译后的ffmpeg文件来反编译opencv

1.未编译前的源代码复制到aarch64-linux-gnu文件夹下(共9个文件)
cp -r libavcodec libavfilter libavresample libpostproc libswscale libavdevice libavformat libavutil libswresample /usr/include/x86_64-linux-gnu

在这里插入图片描述

2.将编译后的头文件全部复制到aarch64-linux-gnu文件夹下(共8个文件)
sudo cp -r /usr/local/ffmpeg/include/*   /usr/include/x86_64-linux-gnu

卸载ffmpeg

sudo make uninstall
sudo apt-get --purge remove ffmpeg
ffmpeg -version

复制完了我们可以开始编译opencv了

编译opencv

下载好了opencv的源代码之后，我们进行解压，在配置依赖环境

sudo apt-get install libopencv-dev build-essential cmake git libprotobuf-dev 
sudo apt-get install libgtk2.0-dev pkg-config python-dev python-numpy
sudo apt-get install libsnappy-dev gtk+-3.0 yasm Doxygen 
sudo apt-get install libdc1394-22 libopenblas-dev liblmdb-dev 
sudo apt-get install libdc1394-22-dev libjpeg-dev libpng12-dev libtiff4-dev 
sudo apt-get install libjasper-dev libhdf5-serial-dev 
sudo apt-get install libavcodec-dev libavformat-dev libswscale-dev libxine-dev 
sudo apt-get install libgstreamer0.10-dev libgflags-dev
sudo apt-get install libgstreamer-plugins-base0.10-dev libv4l-dev libtbb-dev 
sudo apt-get install libqt4-dev libfaac-dev libgoogle-glog-dev
sudo apt-get install libmp3lame-dev libopencore-amrnb-dev libopencore-amrwb-dev 
sudo apt-get install libtheora-dev libvorbis-dev protobuf-compiler liblapack-dev
sudo apt-get install libxvidcore-dev x264 v4l-utils unzip liblapacke-dev 
sudo apt-get install libavresample-dev libgphoto2-dev libavutil-dev
sudo apt-get install --no-install-recommends libboost-all-dev ccache libatlas-base-dev

cmake

jetson

cmake -D WITH_IPP=OFF -D BUILD_opencv_python3=YES -D INSTALL_PYTHON_EXAMPLES=OFF -D BUILD_EXAMPLES=OFF -D BUILD_opencv_python3=ON -D BUILD_opencv_python2=OFF -D BUILD_opencv_java=OFF -D PYTHON3_EXECUTABLE=/usr/bin/python3.6 -D PYTHON3_INCLUDE_DIR=/usr/include/python3.6 -D PYTHON3_INCLUDE_PATH=/usr/include/python3.6 -D PYTHON3_LIBRARY=/usr/lib/aarch64-linux-gnu/libpython3.6m.so.1.0 -D HAVE_opencv_python3=ON -D PYTHON3_NUMPY_INCLUDE_DIRS=/usr/local/lib/python3.6/dist-packages/numpy/core/include -D PYTHON3_PACKAGES_PATH=/usr/local/lib/python3.6/dist-packages -D HAVE_opencv_python3=ON -D OPENCV_ENABLE_MEMORY_SANITIZER=ON -D WITH_QT=OFF -D ENABLE_CXX11=ON -D WITH_CUDA=OFF -D FFMPEG_INCLUDE_DIR=/usr/local/ffmpeg/include -D FFMPEG_LIB_DIR=/usr/local/ffmpeg/lib -D ENABLE_PRECOMPILED_HEADERS=OFF ..

ubuntu

cmake -D WITH_IPP=OFF -D BUILD_opencv_python3=YES -D INSTALL_PYTHON_EXAMPLES=OFF -D BUILD_EXAMPLES=OFF -D BUILD_opencv_python3=ON -D BUILD_opencv_python2=OFF -D BUILD_opencv_java=OFF -D PYTHON3_EXECUTABLE=/usr/bin/python3.6 -D PYTHON3_INCLUDE_DIR=/usr/include/python3.6 -D PYTHON3_INCLUDE_PATH=/usr/include/python3.6 -D PYTHON3_LIBRARY=/usr/lib/x86_64-linux-gnu/libpython3.6m.so.1.0 -D HAVE_opencv_python3=ON -D PYTHON3_NUMPY_INCLUDE_DIRS=/usr/lib/python3/dist-packages/numpy/core/include -D PYTHON3_PACKAGES_PATH=/usr/lib/python3/dist-packages -D HAVE_opencv_python3=ON -D OPENCV_ENABLE_MEMORY_SANITIZER=ON -D WITH_QT=OFF -D ENABLE_CXX11=ON -D WITH_CUDA=OFF -D FFMPEG_INCLUDE_DIR=/usr/local/ffmpeg/include -D FFMPEG_LIB_DIR=/usr/local/ffmpeg/lib -D ENABLE_PRECOMPILED_HEADERS=OFF ..

make

sudo make -j8 #这里的8表示线程的数量
cp /home/lqs/Downloads/opencv-3.4.0/build/lib/python3/* /home/lqs/anaconda3/lib/python3.7/site-packages/

install

make install

check

python
import cv2
cv2.__version__

添加环境变量

#在/etc/ld.so.conf.d/opencv.conf添加安装opencv的路径：
sudo gedit /etc/ld.so.conf.d/opencv.conf 
/usr/local/lib
sudo ldconfig       # 使配置生效
#在/etc/bash.bashrc中添加pkgconfig路径：
sudo gedit /etc/bash.bashrc
"""
PKG_CONFIG_PATH=$PKG_CONFIG_PATH:/usr/local/lib/pkgconfig  
export PKG_CONFIG_PATH 
"""
source /etc/bash.bashrc      # 使配置生效
sudo updatedb

至此，安装成功，可以使用命令pkg-config --modversion opencv3查看是否安装成功以及版本号
卸载opencv
sudo apt-get purge libopencv* 卸载板子上的opencv

卸载编译安装的opencv

cd /home/lqs/Downloads/opencv-3.4.0/build/
sudo make uninstall  && cd .. && rm -rf build
cd /usr && find . -name "*opencv*" | xargs sudo rm -rf
pkg-config --modversion opencv

测试

打开rtsp摄像头

import cv2
cap = cv2.VideoCapture("rtsp://admin:a12345678@10.16.55.150:555/Streaming/Channels/1")
ret, frame = cap.read()
while ret:
    ret, frame = cap.read()
    cv2.imshow("frame",frame)
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break
cv2.destroyAllWindows()
cap.release()

打开rtsp摄像头并获取ntp时间戳

import cv2

video = r'/home/lqs/Documents/face_detect_yolov4_yolov4tiny_ssd-master/action_detection/BalanceBeam.avi'
cap = cv2.VideoCapture("rtsp://admin:a12345678@10.16.55.150:555/Streaming/Channels/1")
while cap.isOpened():
    frame_exists, curr_frame = cap.read()
    # if frame_exists:
    k = cap.getRTPTimeStampSeconds()
    l = cap.getRTPTimeStampFraction()
    time_shift = 0x100000000
    #because in the getRTPTimeStampSeconds() 
    #function, seconds was multiplied by 0x10000000 
    seconds = time_shift * k
    m = (time_shift * k) + l
    print("Imagetimestamp: %i" % m)
    cv2.imshow('1',curr_frame)
    keyCode = cv2.waitKey(1) & 0xFF         
    if keyCode == 27:# ESC键退出
        break
cap.release()
cv2.destroyAllWindows()

编译问题

问题1 libavcodec/libx264.c:282:9: error: ‘x264_bit_depth’ undeclared (first use in this function); did you mean ‘x264_picture_t’?
解决办法：打开libx264.c，找到相应的代码行，将“x264_bit_depth”改成“X264_BIT_DEPTH”即可
问题2 ffbuild/common.mak:60: recipe for target ‘libavdevice/v4l2.o’ failed
解决办法：链接
问题3 /usr/include/aarch64-linux-gnu/sys/videoio.h:45:10: fatal error: opencv2/core/core_c.h: No such file or directory
解决办法：链接
问题4：target ‘libavutil/libavutil.so.55’ failed make: *** [libavutil/libavutil.so.55] Error 1
解决办法：编译后的源代码不干净，需要重新修改源代码来进行编译
问题5： Package ‘gtk±3.0’ not found
解决办法：sudo apt-get install build-essential libgtk-3-dev
问题6： Package ‘gtk±2.0’ not found
解决办法： apt-get install libgtk2.0-dev
问题7:fatal error: sys/videoio.h No such file or directory
解决办法：如果编译ffmpeg出现这个错，直接注释掉c4l2…h文件下video.h和types.h的头函数。
sudo cp /usr/include/opencv2/videoio/videoio_c.h /usr/include/x86_64-linux-gnu/sys
sudo ln -s /usr/include/x86_64-linux-gnu/sys/videoio_c.h /usr/include/x86_64-linux-gnu/sys/videoio.h

附录

Jetson AGX Xavier编译CUDA版本的OpenCV4.5.0

ZZY_dl

关注

0
点赞
踩
8

收藏

觉得还不错? 一键收藏
打赏
0
评论
Jetson 学习笔记(十一)：jetson agx xavier 源码编译ffmpeg(3.4.1)和opencv(3.4.0)

文章目录编译ffmpeg编译opencv编译ffmpeg下载网址：http://www.ffmpeg.org/releases/安装依赖sudo apt-get install -y nasm yasm cmake mercurialsudo apt-get install -y libxcb-xfixes0-dev pkg-config texinfo wget zlib1g-devsudo apt-get install -y libtool libva-dev libvdpau-dev
复制链接

扫一扫