nvidia TX2 DeepStream示例

nvidia示例。

https://github.com/NVIDIA-AI-IOT/redaction_with_deepstream/

示例

示例地址:https://github.com/NVIDIA-AI-IOT/deepstream_reference_apps 。克隆其到 `/opt/nvidia/deepstream/deepstream_reference_apps 。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
模型文件:/opt/nvidia/deepstream/deepstream-5.0/samples/models

示例文档:https://docs.nvidia.com/metropolis/deepstream/dev-guide/text/DS_ref_app_github.html


安装依赖:
查看:/opt/nvidia/deepstream/deepstream/sources/apps/sample_apps/deepstream-app/README
命令:
sudo apt-get install libgstreamer-plugins-base1.0-dev libgstreamer1.0-dev \
libgstrtspserver-1.0-dev libx11-dev libjson-glib-dev
编译:make (验证SDK、依赖是否完备)。

将 anomaly 拷贝到 /opt/nvidia/deepstream/deepstream/sources/apps/sample_apps/ :
cp /opt/nvidia/deepstream/deepstream_reference_apps/anomaly/ /opt/nvidia/deepstream/deepstream/sources/apps/sample_apps/

进入目录:
cd /opt/nvidia/deepstream/deepstream/sources/apps/sample_apps/anomaly
编译库:
cd plugins/gst-dsdirection/
make && sudo make install

编辑 ./anomaly/apps/deepstream-anomaly-detection-test/dsanomaly_pgie_config.txt 文件。


2个示例:
cd /opt/nvidia/deepstream/deepstream_reference_apps/anomaly/apps/deepstream-anomaly-detection-test
./deepstream-anomaly-detection-app /opt/nvidia/deepstream/deepstream/samples/streams/sample_720p.h264

cd /opt/nvidia/deepstream/deepstream_reference_a
-to-back-detectors
./back-to-back-detectors /opt/nvidia/deepstream/deepstream/samples/streams/sample_720p.h264

出错及解决:

1
2
3
4
5
6
提示:
nvbuf_utils: Could not get EGL display connection:
临时设置环境环境: export DISPLAY=:0

One element could not be created. Exiting.

其他Github上的APP例子:

360度智能停车App

演示了一个或者多个360度视角的视频流的矫正功能。从一个CSV文件中读取摄像机矫正参数,然后将矫正后的过道和区域画面,呈现在屏幕上。

https://github.com/NVIDIA-AI-IOT/deepstream_360_d_smart_parking_application

DeepStream SDK for Redaction

重点展示如何在DeepStream里,对一个视频流,同时对人脸区域和车牌区域进行模糊, 也展示了当该DeepStream流水线风格的应用程序在运行的时候,动态的添加删除视频流通道。

https://github.com/NVIDIA-AI-IOT/redaction_with_deepstream

异常检测参考应用

使用并行管道处理2个流的示例应用程序。

https://github.com/NVIDIA-AI-IOT/deepstream_reference_apps/tree/master/anomaly

是否有关于动态流管理的例子吗?

https://github.com/NVIDIA-AI-IOT/deepstream_reference_apps/tree/master/anomaly

里面包括了在管道运行时动态添加或删除通道。

https://github.com/NVIDIA-AI-IOT/redaction_with_deepstream/