• Hardware Platform (Jetson NX )
• DeepStream Version 5.1
• JetPack Version (4.5.1)
• TensorRT Version 7.1.3
• Issue Type( questions)
I refer to GitHub( https://github.com/NVIDIA-AI-IOT/yolov4_ Now I need to send the image with bbox and other metadata to the custom server. What should I do? The data source is multi RTSP
1.In the example of deep stream image meta test, the MP4 file is normal, rtsp:1920 *1080, the screen flickers like it’s blocked, and it’s saved as a green image. I see many people raise this problem, but it’s still not solved in the new version.
2. In the example of deepstream-test5, we hope to add the inference result (full frame with bbox, as shown in the figure below) to the metadata. The current field has no image, and the analysis data will be sent together through Kafka
sudo ./deepstream-image-meta-test rtsp://…
In the source code, only the category is changed to person, and the others are unchanged. Can you change the RTSP test of a new camera,I do see others reflecting that this problem still exists in ds5.1
ok, can you share a diff file with us about exact what’s the modification from you compared with the original sample code, we will try to repro it following your steps.
i use same test code as you attached, with rtsp stream source, only people encoded as picutures, but can not observe the issue, any difference between us?
I found that a small number of IP cameras can, and the resolution is 1920 * 1080. I don’t know why, so please try to replace other brands of IP cameras to reproduce this problem.