How to Integrate YoloV8 in Sample Object Detector (sample_object_detector_tracker) Pipeline

Autonomous Vehicles DRIVE AGX Orin DRIVE AGX Orin General
1

Please provide the following info (tick the boxes after creating this topic):
Software Version
[*] DRIVE OS 6.0.10.0
DRIVE OS 6.0.8.1
DRIVE OS 6.0.6
DRIVE OS 6.0.5
DRIVE OS 6.0.4 (rev. 1)
DRIVE OS 6.0.4 SDK
other

Target Operating System
[*] Linux
QNX
other

Hardware Platform
DRIVE AGX Orin Developer Kit (940-63710-0010-300)
DRIVE AGX Orin Developer Kit (940-63710-0010-200)
[*] DRIVE AGX Orin Developer Kit (940-63710-0010-100)
DRIVE AGX Orin Developer Kit (940-63710-0010-D00)
DRIVE AGX Orin Developer Kit (940-63710-0010-C00)
DRIVE AGX Orin Developer Kit (not sure its number)
other

SDK Manager Version
2.1.0
other

Host Machine Version
native Ubuntu Linux 20.04 Host installed with SDK Manager
[*] native Ubuntu Linux 20.04 Host installed with DRIVE OS Docker Containers
native Ubuntu Linux 18.04 Host installed with DRIVE OS Docker Containers
other

Issue Description
As the sample_object_detector_tracker application is detecting car class but we want to detect more object classes like pedestrian, person, traffic light etc. As the sample_object_detector_tracker uses tensorRT_model.bin & tensorRT_model.bin.json for car detection how can we accommodate other classes in the current object detection pipeline of sample_object_detector_tracker.

2

Dear @PA_GN,
I am assuming yolov8 model is working with sample object detector. Could you check adding a if section for the needed classes in interpretOutput() to include other classes like car?

4

这题，我会。如果是中国的，可以加微信交流：18667105980

5

Hi @deng.shigang,

我不是来自中国，但我们可以在 Nvidia 论坛上的个人消息中联系吗？

6

这样会比较麻烦。如果有其他联系方式会更好。我们这边是直接在dw中使用 TensorRT来部署模型，然后会使用opencv的相关函数，所以你也需要交叉编译一下opencv。建议参考github的项目：https://github.com/triple-Mu/YOLOv8-TensorRT.git。

为啥不直接用dw的dnn模块呢，因为他这个模块，缺乏相关的模型前处理和后处理相关函数，文档也不清晰。

7

Hi @deng.shigang,

可以通过 Skype 进行交流吗？
你有什么文件可供我参考吗？
感谢您的支持。

8

Thanks @deng.shigang for your support to dev community.
@PA_GN , Once you have yolo v8 ONNX model, you need to use tensorRT_Optimization tool to generate TensorRT model that can be integrated into DW. To use, DW DNN supported preprocessing, you can update the tensorRT_model.bin.json file with preprocessing parameters(DriveWorks SDK Reference: DataConditioner Interface). Any additional pre/post processing needs to be handled separately in the sample.

10

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.