Intel® Distribution of OpenVINO™ Toolkit
Community assistance about the Intel® Distribution of OpenVINO™ toolkit, OpenCV, and all aspects of computer vision-related on Intel® platforms.
6163 Discussions

OpenVino inference engine fails on Raspberry Pi 4+NCS 2+OpenCV compiled from source

hamze60
New Contributor I
792 Views

There is an issue in using NCS2 on RaspberryPi 4 B, when OpenCV is compiled from source. I explained the error here, because I guess it is an issue about OpenCV compile.

0 Kudos
5 Replies
JesusE_Intel
Moderator
792 Views

Hi hamzeh,

Could you provide a detailed description of the error you are seeing and steps to reproduce? Are you using one of our sample applications or your own application? Are you using a pre-trained model or a custom trained model? 

Did you follow the Install OpenVINO™ toolkit for Raspbian* OS document?

Regards,

Jesus

hamze60
New Contributor I
792 Views

Jesus E. (Intel) wrote:

Hi hamzeh,

Could you provide a detailed description of the error you are seeing and steps to reproduce? Are you using one of our sample applications or your own application? Are you using a pre-trained model or a custom trained model? 

Did you follow the Install OpenVINO™ toolkit for Raspbian* OS document?

Regards,

Jesus

Hello, I opened a topic about this on opencv forum, here and explained the problem there.
fortunately I found a solution. When opencv is compiled with cross-compile, problem does not happen. Only with direct compiling it happens.

Shubha_R_Intel
Employee
792 Views

Dear ahangari, hamzeh,

Thanks for sharing your victory and findings with the OpenVino community ! It helps tremendously when we share information with each other.

Shubha

 

RTasa
New Contributor I
792 Views
Can you give a step by step version of what to do. Compiling building setting all the compiler switches is not obvious. Not even sure what cross compiling is. Why are you using Docker? Do you need to install Docker. Just setting this up to get to the point where you can use and modify models still seems difficult
hamze60
New Contributor I
792 Views

Because directly compiling on Arm boards is generally slow. You can compile once with Docker on your PC, and keep the compiled files for future usage. Please refer to the cross-compile page, I cited earlier. all steps are there.

Reply