no total frames parameter for save output .mp4 file

Accelerated Computing Intelligent Video Analytics DeepStream SDK
#1

Hi,
I set “[sink1] type 3=file” to save ouput file ,and using opencv to read the saved mp4 file while running “detecting(deepstream-app -c deepstream_app_config_yoloV3.txt”.

I get a error number of total frame by “cap.get(cv2.CAP_PROP_FRAME_COUNT)”
–count_frame= -9.223372036854776e+18

But it’s correct after quit the process of “detecting(deepstream-app -c deepstream_app_config_yoloV3.txt”.

Why it’s incorrect while running the project? How to solve it?

config file

[source0]
enable=1
#Type - 1=CameraV4L2 2=URI 3=MultiURI
type=4
uri=rtsp://admin:admin@192.168.1.64:554/cam/realmonitor?channel=1&subtype=2

num-sources=1
gpu-id=0
# (0): memtype_device   - Memory type Device
# (1): memtype_pinned   - Memory type Host Pinned
# (2): memtype_unified  - Memory type Unified
cudadec-memtype=0

[sink0]
enable=1
#Type - 1=FakeSink 2=EglSink 3=File
type=2
sync=0
source-id=0
gpu-id=0
nvbuf-memory-type=0

[sink1]
enable=1
#Type - 1=FakeSink 2=EglSink 3=File
type=3
#1=mp4 2=mkv  3=mpeg4
container=2
#1=h264 2=h265
codec=1
sync=0
bitrate=10000000
#1=cbr 2=vbr
# rc-mode=2
iframeinterval=30
#1=baseline 2=main 3=high
#profile=3
#output-file=file:///home/nvidia/TX2/tensorRT/deepstream_reference_apps-AI/output/out.mp4
output-file=../../output/out.mp4
source-id=0

code:

[code]import cv2

print(cv2.version)

cap = cv2.VideoCapture("/home/nvidia/TX2/tensorRT/deepstream_sdk_v4.0_v5/output/out.mp4")
count = cap.get(cv2.CAP_PROP_FRAME_COUNT)

print(“count_frame=”, count)

print(“CAP_PROP_POS_FRAMES=”, cap.set(cv2.CAP_PROP_POS_FRAMES, 22))
position=cap.get(cv2.CAP_PROP_POS_FRAMES)
print(position)

#2

Hi,
The API is for getting information from a saved file. While the application is running, the file is lively being updated.

If you intend to access video frames, please use dsexample plugin.
https://docs.nvidia.com/metropolis/deepstream/4.0/dev-guide/index.html#page/DeepStream_Development_Guide%2Fdeepstream_custom_plugin.html

#3

Hello,
I just want to get the recognized frame one by one which including bbox ,and send it out by TCP.
So I want to save it as image firstly ,I have tried to save the frame using cv::imwrite() in the function of dsexample plugin->get_converted_mat,but no bbox for the image .

Where I can get the frame included bbox?

#4

Hi,
We have Encode + RTSP streaming in sink gruop. You may give it a try.
https://docs.nvidia.com/metropolis/deepstream/4.0/dev-guide/index.html#page/DeepStream_Development_Guide%2Fdeepstream_app_config.3.2.html%23wwpID0ENHA