You can implement your own specific app, to generate image user meta in pgie_src_pad_buffer_probe() with nvinfer plugin and use your customized nvmsgconv plugin to handle the image user meta.
please suggest to me how to work with pgie_src_pad_buffer_probe() in nvinfer plugin
The better way is to use pgie_src_pad_buffer_probe() in application with nvinfer src pad.
deepstream-image-meta-test is already the sample.
so difficult with pgie_src_pad_buffer_probe() in deepstream-test5-app-main.c and gstnvmsgconv.c, So different way i can get image at gstdsexample.cpp, are you think sgtnvmsgconv.c can get image like gstdsexample
can you give me an example code for save image in deepstream-test5-app flow your way, thanks so much
You may refer to Encoding jpeg image after OSD causes segfault
deepstream-test5 is a dynamic pipeline app, it is very hard to insert image saving into deepstream-test5 because there is limitation with nvds_obj_enc_create_context() interface.
if get image at OSD i think so difficult to get image path for custom payload to json message broker
It is a sample. You need to find the correct place to put pgie_src_pad_buffer_probe() according to your actual pipeline.
how about Gstdsexample can apply to Gstnvmsgconv for save image. In Gstdsexample i had save image work well.
Can i use Gstdsexample in Gstnvmsgconv like a lib ?
The dsexample can be a sample for you to modify gstnvmsgconv
i am checking and modify, an i will ask for problem. thank you
To modify gstnvmsgconv like gstdsexample not easy.
Can i creat kafka producer at gstdsexample.cpp ? @Fiona.Chen
To modify gstdsexample as nvmsgconv is not easy too.
hello, gstdsexample where are you put pgie_src_pad_buffer_probe() function
hello, in this post How to save frame as jpg and send the filename to kafka in deepstream 5.0 deepstream_test5 has an user say
dsexample is not before the msgconv by default, so you can’t receive the frame_image. You can move dsexample bin to the right position.
So how to do that @Fiona.Chen, help me
i have see
deepstream_transfer_learning_app_main.cpp work on deepstream_app, and it can save image over
But this lid write on C++, and Deepstream-test5-app write on C, So when i try include header
#include "image_meta_consumer.h" #include "image_meta_producer.h"
Get error while i run deepstream_transfer_learning_app is successful
image_meta_consumer.h:25:10: fatal error: condition_variable: No such file or directory #include <condition_variable> ^~~~~~~~~~~~~~~~~~~~ compilation terminated.
please help me
There are a lot of resources for resolving building errors with google.
std::condition_variable - cppreference.com
condition_variable is supported since C++11, please write correct makefile for your application.
This is DeepStream forum, let’s focus on DeepStream topics.
i has fix it, thank you so much