Where is problem of test4 pipline?

wilicyy · November 10, 2021, 1:51am

hi,
i am working jetson nx wiht deepstream 5.1 version,
i build python test4 pipline with RTSP-in&out and imagemete multistream together,
pls help me check where is wrong?

print("Linking elements in the Pipeline \n")
streammux.link(pgie)
pgie.link(tracker)
tracker.link(sgie1)
sgie1.link(sgie2)
sgie2.link(sgie3)
sgie3.link(nvvidconv1)
nvvidconv1.link(filter1)
filter1.link(tiler)
tiler.link(nvvidconv)
nvvidconv.link(nvosd)
nvosd.link(tee)
queue1.link(msgconv)
msgconv.link(msgbroker)
if is_aarch64() and not no_display:
    queue2.link(transform)
    transform.link(sink)
else:
    queue2.link(sink)
nvosd.link(nvvidconv_postosd)
nvvidconv_postosd.link(caps)
caps.link(encoder)
encoder.link(rtppay)
rtppay.link(sink)

and logs:

it hung there and don’t move…

thank you very much!

AastaLLL · November 10, 2021, 2:19am

Hi,

It looks the pipeline starts and no error.
Do you get the expected RTSP output?

Thanks.

wilicyy · November 10, 2021, 3:44am

i didn’t get RTSP output because main procedure hung there.
i checked again and maybe my cloud kafka a little problem and i will try again.
another quesiton is where is this lib file? i didn’t find that…

dlopen error: /opt/nvidia/deepstream/deepstream/lib/libnvds_nvmultiobjecttracker.so: cannot open shared object file: No such file or directory

wilicyy · November 15, 2021, 6:26am

hi,
my problem is like this:

my command is :
$ python3 test4_wang.py rtsp://admin:sgs123456@192.168.10.146/stream1 -p /opt/nvidia/deepstream/deepstream-5.1/lib/libnvds_kafka_proto.so --conn-str=“112.74.161.156;9092;mytest” -s 1

pls help to check my link where is wrong, thank you very much!

print("Linking elements in the Pipeline \n")
streammux.link(pgie)
pgie.link(tracker)
tracker.link(sgie1)
sgie1.link(sgie2)
sgie2.link(nvvidconv1)
nvvidconv1.link(filter1)
filter1.link(tiler)
tiler.link(nvvidconv)
nvvidconv.link(nvosd)
nvosd.link(tee)
queue1.link(msgconv)
msgconv.link(msgbroker)
if is_aarch64() and not no_display:
    queue2.link(transform)
    transform.link(sink)
else:
    queue2.link(sink)


queue3.link(nvvidconv_postosd)
nvvidconv_postosd.link(caps)
caps.link(encoder)
encoder.link(rtppay)
rtppay.link(sink)

AastaLLL · November 15, 2021, 7:17am

Hi,

Have you tried the kafka test program to check if it works?

/opt/nvidia/deepstream/deepstream-6.0/sources/libs/kafka_protocol_adaptor/

# To run test program:
$ ./test_sample_proto_sync
$ ./test_sample_proto_async

Thanks.

wilicyy · November 15, 2021, 8:55am

kafka works because i may run test5 correctly by same IP and port.

wilicyy · November 15, 2021, 9:04am

after long long time, it shows like this:

it seems uripad problem, i used rtsp and removed input file in pipline.

thanks!

wilicyy · November 18, 2021, 11:19am

hi,
i forgot connect cable with camera…
and my problem is caused by lib of kafka, and i have to update to deepstream 6.0.
i think kafka lib is different between 5.1 and 6.0.
i will feed back to you after updating.

thank you

AastaLLL · November 22, 2021, 4:53am

Hi,

Suppose you have made some update to the test4.
Could you share your source with us so we can try it in our environment.

Thanks.

wilicyy · November 25, 2021, 1:21pm

hi,
my custom test4 runs already, but maybe some defines were problem and caused test4 is not stable to run, and didn’t detect objects which i want, i am checking and testing that.

thank you very much!

wilicyy · November 25, 2021, 1:29pm

and a more question:
many coder-persons do like test5 writed by C language because it has config file to get all works easier!
NV should release a similar config file for python test4 similar to test5 i think.
right now, i am working python test4 and test5 together, but i want to use custom test5 in my project of next year very soon.
pls advice me a easy and stable way to take images by test5?
dsexample or deepstream-transfer-learning-app?
i want to get full frame with bbox together.

thank you very much!

wilicyy · November 25, 2021, 1:31pm

sorry upload wrong photo for test4 problem, it should be this:

AastaLLL · November 29, 2021, 4:01am

Hi,

The log indicates that there is a missing attribute called helmet in your custom pyds.NvDSPersonObject.
This error is triggered when an object is detected, so the test4 will run correctly if no object is detected.

Since this is related to the custom implementation, we need the source code to figure out what is going wrong.
Would you mind sharing the source with us?

Thanks.

wilicyy · November 29, 2021, 5:58am

hi,
thank you for you time to help me.
i didn’t finish modify all codes still because i am trying to work enj_image on test5 recently,
but pls help me take a look about my python test4 codes, i have to finish both way for my project.
pls check attached file.

thank you very much!
test4_wang.py (38.0 KB)

AastaLLL · December 6, 2021, 4:02am

Hi,

The error occurs since you try to access a HELMET variable from a PERSON object.

However, the error is not consistent to the source you shared.
Do you get the error with this script? Or you fixed it already?

Thanks.

wilicyy · December 6, 2021, 6:41am

hi,
thanks for your kind and your time!
i wrote report for my project at last week, and i will delete test4 sgie and try again tonight.
i would like to ask a question about dsexample, when i enable it and run test5, two bbox shows but the resize didn’t work, what’s this problem? should i add dsexample in create common pipline?
Processing: dsexample3.png…

wilicyy · December 6, 2021, 6:42am

wilicyy · December 6, 2021, 8:38pm

hi,
i delete sgie2 and test4 runs no error, but only runs few seconds then exit automatically, i will check codes again.
pls help me for dsexample probelm and it is urgently for my project.
another problem is enc_jpeg from nvifer pad on test5:
i got codes which released patch by you from the forums, but when i make that codes, it shows cuda prop error, how should i do?

the part:
int current_device = -1;
cudaGetDevice(&current_device);
struct cudaDeviceProp prop;
cudaGetDeviceProperties(&prop, current_device);

thank you so much!!!

AastaLLL · December 10, 2021, 6:08am

Hi,

Do you mind sharing more about the problem you are facing currently?

The patch you shared above is trying to get the GPU property.
Do you want the information?

For the resize issue, do you use our default bounding box parser?
Or a customized implementation?

Thanks.

wilicyy · December 10, 2021, 6:19am

hi,
for GPU property compile problem i have past and forget that pls.
now’s problem is dsexample and enc_image by your patch.
1\the first dsexample, when i enable=1 it shows that 2 fake boxes on menu only, and no pictures to save i used default bounding box parser, but my model is custom yolo4, how should i do? i also checked fourms for more informations, should i have to modify out CV MAT in convert-mat funtions to get dsexample to work ?
2\i used attached patch for deepstream-app’s enc objects to save images by user_meta, but get nothing images, too

deepstream_app.patch (5.8 KB)