Deepstream-transfer-learning-app saving wrong image file name

andrea_vighi · October 19, 2020, 7:42am

Same problem…

How can i solve it?
Thanks

andrea_vighi · October 19, 2020, 9:39am

I can’t solve the problem, any suggestions?
Thank you

Amycao · October 20, 2020, 8:56am

I can not repro the green issue after upgrade, can you specify how to repro your issue?

andrea_vighi · October 20, 2020, 9:17am

Deepstream 5.0
4 RTSP sources (hikvision cameras)
I’m using this config file:

[application]
enable-perf-measurement=1
perf-measurement-interval-sec=5
#gie-kitti-output-dir=streamscl

[tiled-display]
enable=1
rows=2
columns=2
width=1280
height=720
gpu-id=0
#(0): nvbuf-mem-default - Default memory allocated, specific to particular platform
#(1): nvbuf-mem-cuda-pinned - Allocate Pinned/Host cuda memory applicable for Tesla
#(2): nvbuf-mem-cuda-device - Allocate Device cuda memory applicable for Tesla
#(3): nvbuf-mem-cuda-unified - Allocate Unified cuda memory applicable for Tesla
#(4): nvbuf-mem-surface-array - Allocate Surface Array memory, applicable for Jetson
nvbuf-memory-type=0

[source0]
enable=1
#Type - 1=CameraV4L2 2=URI 3=MultiURI 4=RTSP
type=4
uri=rtsp://IP
num-sources=1
#drop-frame-interval=2
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

[source1]
enable=1
#Type - 1=CameraV4L2 2=URI 3=MultiURI 4=RTSP
type=4
uri=rtsp://IP
num-sources=1
#drop-frame-interval=2
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

[source2]
enable=1
#Type - 1=CameraV4L2 2=URI 3=MultiURI 4=RTSP
type=4
uri=rtsp://IP
num-sources=1
#drop-frame-interval=2
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

[source3]
enable=1
#Type - 1=CameraV4L2 2=URI 3=MultiURI 4=RTSP
type=4
uri=rtsp://IP
num-sources=1
#drop-frame-interval=2
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

[osd]
enable=1
gpu-id=0
border-width=1
text-size=15
text-color=1;1;1;1;
text-bg-color=0.3;0.3;0.3;1
font=Serif
show-clock=0
clock-x-offset=800
clock-y-offset=820
clock-text-size=12
clock-color=1;0;0;0
nvbuf-memory-type=0

[streammux]
gpu-id=0
##Boolean property to inform muxer that sources are live
live-source=1
batch-size=4
##time out in usec, to wait after the first buffer is available
##to push the batch even if the complete batch is not formed
batched-push-timeout=40000

Set muxer output width and height

width=1920
height=1080
#enable to maintain aspect ratio wrt source, and allow black borders, works
##along with width, height properties
enable-padding=0
nvbuf-memory-type=0

If set to TRUE, system timestamp will be attached as ntp timestamp

If set to FALSE, ntp timestamp from rtspsrc, if available, will be attached

attach-sys-ts-as-ntp=1

config-file property is mandatory for any gie section.

Other properties are optional and if set will override the properties set in

the infer config file.

[primary-gie]
enable=1
gpu-id=0
model-engine-file=…/…/…/…/…/samples/models/Primary_Detector/resnet10.caffemodel_b4_gpu0_int8.engine
batch-size=4
#Required by the app for OSD, not a plugin property
bbox-border-color0=1;0;0;1
bbox-border-color1=0;1;1;1
bbox-border-color2=0;0;1;1
bbox-border-color3=0;1;0;1
interval=0
gie-unique-id=1
nvbuf-memory-type=0
#config-file=config_infer_primary_ds_transfer_learning.txt
config-file=/opt/nvidia/deepstream/deepstream/samples/configs/tlt_pretrained_models/config_infer_primary_peoplenet.txt

[tracker]
enable=1

For the case of NvDCF tracker, tracker-width and tracker-height must be a multiple of 32, respectively

tracker-width=640
tracker-height=384
#ll-lib-file=/opt/nvidia/deepstream/deepstream-5.0/lib/libnvds_mot_iou.so
#ll-lib-file=/opt/nvidia/deepstream/deepstream-5.0/lib/libnvds_nvdcf.so
ll-lib-file=/opt/nvidia/deepstream/deepstream-5.0/lib/libnvds_mot_klt.so
#ll-config-file required for DCF/IOU only
#ll-config-file=tracker_config.yml
#ll-config-file=iou_config.txt
gpu-id=0
#enable-batch-process and enable-past-frame applicable to DCF only
enable-batch-process=1
enable-past-frame=0
display-tracking-id=1

[tests]
file-loop=0

[img-save]
enable=1
output-folder-path=./output
save-img-cropped-obj=1
save-img-full-frame=0
frame-to-skip-rules-path=capture_time_rules.csv
second-to-skip-interval=5
min-confidence=0.5
max-confidence=1.0
min-box-width=5
min-box-height=5

If I set 1 RTSP source work fine, with more RTSP i’ve the green issue…

Amycao · October 26, 2020, 6:54am

We have internal bug to track this issue.

andrea_vighi · November 25, 2020, 9:07am

News about this problem?
I obtain the “green output” with:

HikVision ThermoCamera - RTSP source: 1920*1080P 8FPS H.265
Ganz - RTSP source: 1080*1920P 22FPS H.265

I’ve no problem with:

HikVision - RTSP source: 1920*1080P 8FPS H.265.

What could be the problem?
because with some cameras it works well and with others I have this problem?
Thanks a lot

Amycao · November 30, 2020, 3:28am

Can you get the output if using sink type 2 EglSink? try to run with your camera without cropping the objects into file, run other sample like deepstream-app, just with sink type EglSink.
Can you provide the camera codec parameter which have the issue?

andrea_vighi · December 14, 2020, 8:55am

I’m using EglSink for all cameras, but I’ve the issue with same parameter of the camera that works well.
1080P
bitrate variable
max bitrate 5325
20 fps
h265

That’s why I don’t understand why…

andrea_vighi · December 17, 2020, 8:18am

i noticed that green images are saved even with an mp4 video (h264 main profile).

Once converted to h264 high profile the problem disappears.

So I tried to change the camera configurations but, both with main and with high profile, the problem remains

PhongNT · December 17, 2020, 8:30am

can you set width=752 height=480 in group streammux, and run again

Amycao · December 18, 2020, 8:34am

You mean you run sample with EglSink with above parameters on all cameras, but will have same issue on the cameras you mentioned in comment 29 which have green output failure running Deepstream-transfer-learning-app？

andrea_vighi · December 18, 2020, 9:30am

exactly.
I have problems with all hikvision and ganz ip cameras

Amycao · December 28, 2020, 5:58am

Sorry for a late reply.
I am so confused about your coments, first you said you just have green output issue with HikVision - RTSP source: 1920*1080P 8FPS H.265.
but latest comment show all the cameras have issue.
Can you specify details your failure case one by one?

andrea_vighi · January 7, 2021, 8:16am

It is a problem that occurs randomly … sometimes it happens even with an mp4 video file. I can’t understand why. Maybe a codec incompatibility?

andrea_vighi · January 22, 2021, 9:08am

with this setting, for an rtsp stream, now it works.

Regarding this, however, I confirm that with rtsp multi sources sometimes I still get the green output images and often sources id for the saved images are wrong.

Any solution regarding this problems?
Thanks

PhongNT · January 22, 2021, 9:35am

jetpack 4.5 has release, can you reinstall os, and check

brian0b6iu · October 8, 2021, 2:18pm

I can confirm that I still get this issue with jetpack 4.5.1 / deepstream 5.1

PhongNT · October 8, 2021, 3:40pm

Refer to this post: Object_id and source id in file name of image save by deepstream-transfer-learning-app is wrong - #13 by XiangjiBU

brian0b6iu · October 8, 2021, 3:54pm

Sorry, my problem was related to the green images that are saved. It was solved by the following post: