opencv和ffmpeg适配指南
0.前言
感谢Yuzuki大佬提供的两个支持包，原贴链接：V853/V851s 增加 OpenCV, FFmpeg 编译包 | 全志在线开发者论坛
opencv支持包：[opencv.tar.gz](https://forums.100ask.net/uploads/short-url/5yNatU5BBvsoRGv8EGnIuOMEeaB.gz)
ffmpeg支持包：ffmpeg.tar.gz
1.安装opencv和ffmpegopencv和ffmpeg适配指南
将下载的两个压缩包opencv.tar.gz和ffmpeg.tar.gz传入Ubuntu中，可以放置在任意目录中，假设我放在/home/book/workspaces目录下
book@100ask:~/workspaces$ ls
tina-v853-open opencv.tar.gz ffmpeg.tar.gz
解压两个压缩包输入
tar -xzvf opencv.tar.gz
tar -xzvf ffmpeg.tar.gz
例如：
book@100ask:~/workspaces$ tar -xzvf opencv.tar.gz
opencv/
opencv/patches/
opencv/patches/010-uclibc-ng.patch
opencv/patches/020-l_tmpnam.patch
opencv/Makefile
opencv-sample/
opencv-sample/opencv-camera/
opencv-sample/opencv-camera/src/
opencv-sample/opencv-camera/src/main.cpp
opencv-sample/opencv-camera/src/Makefile
opencv-sample/opencv-camera/Makefile
opencv-sample/yolov5-post-opencv/
opencv-sample/yolov5-post-opencv/src/
opencv-sample/yolov5-post-opencv/src/main.cpp
opencv-sample/yolov5-post-opencv/src/Makefile
opencv-sample/yolov5-post-opencv/Makefile
book@100ask:~/workspaces$ tar -xzvf ffmpeg.tar.gz
ffmpeg/
ffmpeg/patches/
ffmpeg/patches/030-h264-mips.patch
ffmpeg/patches/050-glibc.patch
ffmpeg/patches/010-pkgconfig.patch
ffmpeg/Config.in
ffmpeg/Makefile
将解压出来的文件拷贝到tina-v853-open/openwrt/package/目录下
book@100ask:~/workspaces$ cp -rfd opencv tina-v853-open/openwrt/package/
book@100ask:~/workspaces$ cp -rfd opencv-sample tina-v853-open/openwrt/package/
book@100ask:~/workspaces$ cp -rfd ffmpeg tina-v853-open/openwrt/package/
查看该目录可看到复制过来的三个文件
book@100ask:~/workspaces$ ls tina-v853-open/openwrt/package/
allwinner feeds ffmpeg opencv opencv-sample thirdparty
2.编译opencv和ffmpeg
进入Tina根目录下，配置Tina环境，选择100ASK_V853-PRO开发板。
book@100ask:~/workspaces$ cd tina-v853-open/
book@100ask:~/workspaces/tina-v853-open$ source build/envsetup.sh
NOTE: The SDK(/home/book/workspaces/tina-v853-open) was successfully loaded
load openwrt... ok
Please run lunch next for openwrt.
load buildroot,bsp...ok
Please run ./build.sh config next for buildroot,bsp.
book@100ask:~/workspaces/tina-v853-open$ lunch
You're building on Linux
Lunch menu... pick a combo:
1 v853-100ask-tina
2 v853-vision-tina
Which would you like? [Default v853-100ask]: 1
Jump to longan autoconfig
/home/book/workspaces/tina-v853-open/build.sh autoconfig -o openwrt -i v853 -b 100ask -n default
========ACTION List: mk_autoconfig -o openwrt -i v853 -b 100ask -n default;========
options :
INFO: Prepare toolchain ...
INFO: kernel defconfig: generate /home/book/workspaces/tina-v853-open/kernel/linux-4.9/.config by /home/book/workspaces/tina-v853-open/device/config/chips/v853/configs/100ask/linux-4.9/config-4.9
INFO: Prepare toolchain ...
make: Entering directory '/home/book/workspaces/tina-v853-open/kernel/linux-4.9'
*** Default configuration is based on '../../../../../device/config/chips/v853/configs/100ask/linux-4.9/config-4.9'
#
# configuration written to .config
#
make: Leaving directory '/home/book/workspaces/tina-v853-open/kernel/linux-4.9'
INFO: clean buildserver
Usage:
kill [options] <pid> [...]
Options:
<pid> [...] send signal to every <pid> listed
-<signal>, -s, --signal <signal>
specify the <signal> to be sent
-l, --list=[<signal>] list all signal names, or convert one to a name
-L, --table list all signal names in a nice table
-h, --help display this help and exit
-V, --version output version information and exit
For more details see kill(1).
INFO: prepare_buildserver
进入Tina的配置界面
book@100ask:~/workspaces/tina-v853-open$ make menuconfig
进入如下目录中，选中opencv-camera和yolov5-post-opencv
> lizard
> opencv-sample
<*> opencv-camera............ opencv camera, capture video and display to fb0
< > yolov5-post-opencv....................... yolov5 post process with opencv
其中opencv-camera是使用opencv通过video获取图像并显示在fb0节点上；yolov5-post-opencv为使用opencv编写的yolov5后处理程序。
这里选中opencv-camera完成后保存并退出Tina配置界面。
编译opencv和例程，在Tina根目录下输入make
book@100ask:~/workspaces/tina-v853-open$ make
...
等待编译完成。
编译完成后，将镜像重新烧写到开发板上。