About transplanting face recognition function to linux development board

Study notes about transplanting face recognition function to linux development board

Links to my other projectsSmart storage project based on linux
Face recognition based on QtSmart
home
transplantation based on STM32 Face recognition to Linux development boardSmart
vehicle system (Qt) based on the IMX6ULL development board of punctual atoms

---

foreword

In the previous study, I successfully ran through the face recognition module in ubuntu. So how to transplant face recognition to the linux development board?

---

提示：以下是本篇文章正文内容，下面案例可供参考

1. Preparation

In fact, the most difficult part of this function is the environment construction work, and the logic code part is very simple. The first is to install the openssl library, libcurl library, jsoncpp library. But these three libraries do not mean that they can be completed simply by using the apt-get install command, because the default library obtained by compiling in this way is the x86 architecture library, which cannot be run on the board, so cross-compilation is required.

2. Compilation steps of dependent libraries

The first is to install the openssl, jsoncpp, and libcurl libraries, then enter the makefile, and replace the environment variables of the default compiler with the environment variables of the cross compiler (it’s simple to say, but it’s actually quite complicated. For details, you can refer toXunwei Electronics has a license plate recognition projectVideo, here is the entrance of Xunwei Electronics' license plate recognition project

Then you have to go to Baidu Smart Cloud Platform to download the corresponding API package to the project directory for easy calling.

After the above library is installed, install cmake, use cmake to compile opencv into arm architecture, and finally copy the include/lib/ folder (openssl, libcurl, jsoncpp) of the library just compiled to the root file system The /lib folder, and then modify the environment variables of the root file system. You have to copy the opencv library to the /lib directory of the root file system.
In fact, when installing the library, there will be compatibility issues with various versions of the library, so it is not simple

Summarize

提示：这里对文章进行总结：

For example: Now I just understand the general process, and I will definitely encounter other difficulties in the specific implementation process. I hope I can keep learning, and if the transplant is successful, I will continue to improve this post