Please provide complete information as applicable to your setup.
• Hardware Platform (GPU: GTX 1660 Super) • DeepStream Version: 6.1.1 • JetPack Version (valid for Jetson only) • TensorRT Version • NVIDIA GPU Driver Version (11.6) • Issue Type (bugs)
Assume i have working on Japanese car license plate, detect and OCR it.
The problem is when i display the license plate on the OSD screen, it show (yes the forums can not encoding it also). Like image below.
The expected text that should be show as “”
I tried to change the font mutiple times, using obj_meta.text_params.font_params.font_name = "Noto Sans" but it doesn’t work.
I tried install Japanese language for my Ubuntu 20.04.
1. Install Chinese Language package .
sudo apt-get install language-pack-zh-hans
2. Set the Chinese language enviroment
export LANG=ja_JP.UTF-8
export LANGUAGE="ja_JP:ja:en_US:en"
Then restart the PC and rerun pipeline, even change the font again using obj_meta.text_params.font_params.font_name = "Noto Sans" but it doesn’t work.
How can i solve this issue, please let me know. Thanks.
Also, i tried to type in “asdwasdw” but it doesnt print or notice anything that there is no font named “asdwasdw”. And i thing it may good to know what was the choosen font it is.
You install a Chinese Language package, it cannot support Japanese language. You can try to install Japanese Language package and see if it works well.
Yes. I test that with deepstream-test-1. It works well, too. Perhaps there is a problem with the Japanese language output of your model itself. You need to check your code.
Okey after recheck it, i found the reason of my issue.
It turn out my app is run inside a docker container, so there is no font that support japanese character on it.
So i copy a japanese font, then install the font inside docker, and then select the font name using: py_nvosd_text_params.font_params.font_name