Transplant opencv to the development board, and collect and use the camera on the Development Board 6410 (1), opencv6410
Recently, we have been porting opencv qt to the development board tiny6410 for human eye detection and face recognition. We can run programs on virtual machines, but there are a bunch of errors on the Development Board.
Porting qt is easy to write.
Then install QT4.7
① Copy the compressed package to the host/opt directory
② # Mkdir-p/opt/mini6410
③ # Cd/opt/mini6410
④ # Tar zxvf/opt/qt-everywhere-opensource-src-4.7.0.tar.gz
⑤ Execute the following command to execute the configuration before Qt4.7.0 Compilation
# Cd/opt/mini6410/qt-everywhere-opensource-src-4.7.0
# Echo yes |. /configure-prefix/opt/Qt4.7-opensource-embedded arm-xplatform qws/linux-arm-g ++-no-webkit-qt-libtiff-qt-libmng-qt-mouse -tslib-qt-mouse-pc-no-mouse-linuxtp-no-neon (Be sure to note that, A space cannot be entered incorrectly. execution takes some time. If it passes the code instantly, it indicates that it is incorrect)
7 # make & make install (check whether there are more Qt4.7 folders under/OPT after execution)
Finally, deploy Qt4.7 on mini6410.
① # Cd/opt
② # Tar zcvf qt4.7.tgz Qt4.7 package the compiled Qt4.7 on PC
③ Place the compressed file in the Development Board/opt directory.
④ # Rm/usr/local/Trolltech/QtEmbedded-4.7.0-arm/-rf
⑤ # Cd/opt
# Tar zxvf/sdcard/qt4.7.tgz
Step ④ Delete the Qt4.7 provided by the friendly arm to ensure sufficient space for the Qt4.7 compiled by ourselves.
I also summarized a pair of problems when porting opencv. Generally, there are many errors according to the online tutorials. Here I will list all the modifications and you will be able to follow these steps. If you fail to contact me
Summarize the changes:
Remove WITH_TIFF WITH_CUDA
Change CMAKE_BUILD_TYPE to Release
For details about how to modify the CMAKE_INSTALL_PREFIX path, do not overlap with the pc library.
Go to opencv-arm/CMakeCache.txt and find CMAKE_EXE_LINKER_FLAGS: STRING =''
Replace with: CMAKE_EXE_LINKER_FLAGS: STRING =-lpthread-ldl-lrt
Basically, it is successfully transplanted here, but to use the development board, you need to copy your library to the Development Board/usr/lib. The touch screen can be used only after self-starting. self-starting and font library porting are in my other blog. You can check it out.
This is the next article about camera usage. Many problems
When we port opencv to the 6410 Development Board recently, a BUG always occurs: soft lockup detected on CPU #0.
It should be because the dependent library is not fully installed,
We need to use the 6410 Development Board and ensure the quality of the Development Board.
This depends on the product. Currently, the online promotional materials for the 6410 Development Board are everywhere. From the very beginning, youjian technology has dominated the world, to the present, various types of cloud6410 Development Boards have emerged, and the prices are also uneven.
However, I believe that, as a product, the first is stability, and the second is after-sales. The classic 6410 Development Board is tested by the customer, and the stability is not enough. Therefore, you must be cautious when selecting the 6410 Development Board, especially for enterprises or similar projects.
I personally recommend youjian technology, one of the first companies in China to make the 6410 development board. They have always adhered to the design philosophy of improving performance, reliable quality, and the latest configurations in the industry. They are very mature in ARM. With the serial number, you can upgrade the 6410 Development Board for free and obtain first-hand information about Samsung! Recommended!