search

Smorodov / Multitarget-tracker

Multiple Object Tracker, Based on Hungarian algorithm + Kalman filter.
Apache License 2.0
2.2k stars 652 forks source link

no video output #335

Closed calderonsin closed 3 years ago

calderonsin commented 3 years ago

Hello! I'm trying to use the multitracker, I'm trying to use the YOLO detector with Darknet backend and nothing happened:

./MultitargetTracker /home/juccalderondo/videos/Acevedo_6-18am_6-38am.MP4 -e=5 -a=1 -o=/home/juccalderondo/videos/Acevedo_6-18am_6-38am_detection_2.avi -r=/home/juccalderondo/videos/Acevedo_6-18am_6-38am.csv -g=0 -sl=1 -s=/home/juccalderondo/Multitarget-tracker/data/settings.ini

the output is: Press: 'm' key for change mode: play|pause. When video is paused you can press any key for get next frame.

Press Esc to exit from video

Usage: MultitargetTracker [params] 1

    -a, --async (value:1)
            Use 2 theads for processing pipeline
    --batch_size, --bs (value:1)
            Batch size - frames count for processing
    -c, --config
            For CarsCounting: Config file of neural network: yolov4.cfg
    -e, --example (value:5)
            number of example 0 - MouseTracking, 1 - MotionDetector, 2 - FaceDetector, 3 - PedestrianDetector, 4 - OpenCV dnn objects detector, 5 - YOLO Darknet detector, 6 - YOLO TensorRT Detector, 7 - Cars counting
    --ed, --end_delay (value:0)
            Delay in milliseconds after video ending
    --ef, --end_frame (value:0)
            Play a video to this position (if 0 then played to the end of file)
    -g, --gpu (value:0)
            Use OpenCL acceleration
    --heat_map, --hm (value:0)
            For CarsCounting: Draw heat map
    --inf, --inference (value:darknet)
            For CarsCounting: Type of inference framework: darknet, ocvdnn
    -n, --names
            For CarsCounting: File with classes names: coco.names
    -o, --out (value:/home/juccalderondo/videos/Acevedo_6-18am_6-38am_detection_2.avi)
            Name of result video file
    -r, --res (value:/home/juccalderondo/videos/Acevedo_6-18am_6-38am.csv)
            Path to the csv file with tracking result
    -s, --settings (value:/home/juccalderondo/Multitarget-tracker/data/settings.ini)
            Path to the init file with tracking settings
    --sf, --start_frame (value:0)
            Start a video from this position
    --show_logs, --sl (value:1)
            Show Trackers logs
    -w, --weights
            For CarsCounting: Weights of neural network: yolov4.weights
    --wf, --write_n_frame (value:1)
            Write logs on each N frame: 1 for writing each frame

    1 (value:/home/juccalderondo/videos/Acevedo_6-18am_6-38am.MP4)
            movie file

OpenCL not used ParseTrackerSettings: /home/juccalderondo/Multitarget-tracker/data/settings.ini ... ParseTrackerSettings - readed ParseTrackerSettings: 1

calderonsin commented 3 years ago

I'm on Ubuntu 18.04.3 LTS, it is a docker container of a darknet images (https://hub.docker.com/r/inqdocker/darknet) and got no GUI. Can u lend me a nad?

Nuzhny007 commented 3 years ago

For working without GUI it need to build with option: _cmake . -DSILENTWORK=ON

P.S. I never try this project in docker.

calderonsin commented 3 years ago

Hello, thank u for u reply it worked, it ran the tracking but the .csv is empty and the video itself doesn't have any information about tracking

I used this command: ./MultitargetTracker /home/juccalderondo/videos/Acevedo_6-18am_6-38am.MP4 -e=5 -a=0 -o=/home/juccalderondo/videos/Acevedo_6-18am_6-38am_detection_2.avi -r=/home/juccalderondo/videos/Acevedo_6-18am_6-38am.csv -g=0 -sl=1 -s=/home/juccalderondo/Multitarget-tracker/data/settings.ini

Nuzhny007 commented 3 years ago

What it write in console?

calderonsin commented 3 years ago

these are the first line: Examples of the Multitarget tracking algorithm Usage: ./MultitargetTracker [--example]=<number of example 0..7> [--start_frame]= [--end_frame]= [--end_delay]= [--out]= [--show_logs]= [--async]= [--res]= [--settings]= [--batch_size=]

Press: 'm' key for change mode: play|pause. When video is paused you can press any key for get next frame.

Press Esc to exit from video

Usage: MultitargetTracker [params] 1

    -a, --async (value:0)
            Use 2 theads for processing pipeline
    --batch_size, --bs (value:1)
            Batch size - frames count for processing
    -c, --config
            For CarsCounting: Config file of neural network: yolov4.cfg
    -e, --example (value:5)
            number of example 0 - MouseTracking, 1 - MotionDetector, 2 - FaceDetector, 3 - PedestrianDetector, 4 - OpenCV dnn objects detector, 5 - YOLO Darknet detector, 6 - YOLO TensorRT Detector, 7 - Cars counting
    --ed, --end_delay (value:0)
            Delay in milliseconds after video ending
    --ef, --end_frame (value:0)
            Play a video to this position (if 0 then played to the end of file)
    -g, --gpu (value:0)
            Use OpenCL acceleration
    --heat_map, --hm (value:0)
            For CarsCounting: Draw heat map
    --inf, --inference (value:darknet)
            For CarsCounting: Type of inference framework: darknet, ocvdnn
    -n, --names
            For CarsCounting: File with classes names: coco.names
    -o, --out (value:/home/juccalderondo/videos/Acevedo_6-18am_6-38am_detection_2.avi)
            Name of result video file
    -r, --res (value:/home/juccalderondo/videos/Acevedo_6-18am_6-38am.csv)
            Path to the csv file with tracking result
    -s, --settings (value:/home/juccalderondo/MT2/data/settings.ini)
            Path to the init file with tracking settings
    --sf, --start_frame (value:0)
            Start a video from this position
    --show_logs, --sl (value:1)
            Show Trackers logs
    -w, --weights
            For CarsCounting: Weights of neural network: yolov4.weights
    --wf, --write_n_frame (value:1)
            Write logs on each N frame: 1 for writing each frame

    1 (value:/home/juccalderondo/videos/Acevedo_6-18am_6-38am.MP4)
            movie file

OpenCL not used ParseTrackerSettings: /home/juccalderondo/MT2/data/settings.ini ... ParseTrackerSettings - readed ParseTrackerSettings: 1 Video /home/juccalderondo/videos/Acevedo_6-18am_6-38am.MP4 was started from 0 frame with 29.97 fps Used GPU 0 0 : compute_capability = 610, cudnn_half = 0, GPU: NVIDIA TITAN Xp net.optimized_memory = 0 mini_batch = 1, batch = 8, time_steps = 1, train = 0 layer filters size/strd(dil) input output 0 Create CUDA-stream - 0 Create cudnn-handle 0 conv 32 3 x 3/ 1 416 x 416 x 3 -> 416 x 416 x 32 0.299 BF 1 max 2x 2/ 2 416 x 416 x 32 -> 208 x 208 x 32 0.006 BF 2 conv 64 3 x 3/ 1 208 x 208 x 32 -> 208 x 208 x 64 1.595 BF 3 max 2x 2/ 2 208 x 208 x 64 -> 104 x 104 x 64 0.003 BF 4 conv 128 3 x 3/ 1 104 x 104 x 64 -> 104 x 104 x 128 1.595 BF 5 conv 64 1 x 1/ 1 104 x 104 x 128 -> 104 x 104 x 64 0.177 BF 6 conv 128 3 x 3/ 1 104 x 104 x 64 -> 104 x 104 x 128 1.595 BF 7 max 2x 2/ 2 104 x 104 x 128 -> 52 x 52 x 128 0.001 BF 8 conv 256 3 x 3/ 1 52 x 52 x 128 -> 52 x 52 x 256 1.595 BF 9 conv 128 1 x 1/ 1 52 x 52 x 256 -> 52 x 52 x 128 0.177 BF 10 conv 256 3 x 3/ 1 52 x 52 x 128 -> 52 x 52 x 256 1.595 BF 11 max 2x 2/ 2 52 x 52 x 256 -> 26 x 26 x 256 0.001 BF 12 conv 512 3 x 3/ 1 26 x 26 x 256 -> 26 x 26 x 512 1.595 BF 13 conv 256 1 x 1/ 1 26 x 26 x 512 -> 26 x 26 x 256 0.177 BF 14 conv 512 3 x 3/ 1 26 x 26 x 256 -> 26 x 26 x 512 1.595 BF 15 conv 256 1 x 1/ 1 26 x 26 x 512 -> 26 x 26 x 256 0.177 BF 16 conv 512 3 x 3/ 1 26 x 26 x 256 -> 26 x 26 x 512 1.595 BF 17 max 2x 2/ 2 26 x 26 x 512 -> 13 x 13 x 512 0.000 BF 18 conv 1024 3 x 3/ 1 13 x 13 x 512 -> 13 x 13 x1024 1.595 BF 19 conv 512 1 x 1/ 1 13 x 13 x1024 -> 13 x 13 x 512 0.177 BF 20 conv 1024 3 x 3/ 1 13 x 13 x 512 -> 13 x 13 x1024 1.595 BF 21 conv 512 1 x 1/ 1 13 x 13 x1024 -> 13 x 13 x 512 0.177 BF 22 conv 1024 3 x 3/ 1 13 x 13 x 512 -> 13 x 13 x1024 1.595 BF 23 conv 1024 3 x 3/ 1 13 x 13 x1024 -> 13 x 13 x1024 3.190 BF 24 conv 1024 3 x 3/ 1 13 x 13 x1024 -> 13 x 13 x1024 3.190 BF 25 route 16 -> 26 x 26 x 512 26 reorg_old reorg_old / 2 26 x 26 x 512 -> 13 x 13 x2048 27 route 26 24 -> 13 x 13 x3072 28 conv 1024 3 x 3/ 1 13 x 13 x3072 -> 13 x 13 x1024 9.569 BF 29 conv 40 1 x 1/ 1 13 x 13 x1024 -> 13 x 13 x 40 0.014 BF 30 detection mask_scale: Using default '1.000000' Total BFLOPS 34.879 avg_outputs = 645209 Allocate additional workspace_size = 314.58 MB Loading weights from /home/juccalderondo/backup/yolo-obj_40000.weights... seen 64, trained: 2560 K-images (40 Kilo-batches_64) Done! Loaded 31 layers from weights-file

try to allocate additional workspace_size = 314.58 MB CUDA allocate done!

calderonsin commented 3 years ago

these are the last: Frame 34174: tracks = 0, time = 23 Frame 34175: tracks = 0, time = 19 Frame 34176: tracks = 0, time = 15 Frame 34177: tracks = 0, time = 20 Frame 34178: tracks = 0, time = 14 Frame 34179: tracks = 0, time = 25 Frame 34180: tracks = 0, time = 14 Frame 34181: tracks = 0, time = 12 Frame 34182: tracks = 0, time = 12 Frame 34183: tracks = 0, time = 13 Frame 34184: tracks = 0, time = 12 Frame 34185: tracks = 0, time = 12 algorithms time = 657.614, work time = 1294.54

Nuzhny007 commented 3 years ago

Ok. Do you work in docker now? This project has a piece of code for visual debugging tracking part: https://github.com/Smorodov/Multitarget-tracker/blob/master/src/Tracker/Ctracker.cpp#L156 Set and rebuild: #define DRAW_DBG_ASSIGNMENT 1

But if don't has a GUI then uncomment all lines: https://github.com/Smorodov/Multitarget-tracker/blob/master/src/Detector/YoloDarknetDetector.cpp#L189 https://github.com/Smorodov/Multitarget-tracker/blob/master/src/Detector/YoloDarknetDetector.cpp#L218 etc

calderonsin commented 3 years ago

I'm working on docker without GUI. should I uncomment all lines starting by "//std::cout"? thanks four u reply!

Nuzhny007 commented 3 years ago

Yes

calderonsin commented 3 years ago

okey, I made the changes and built again, i'm going to try the tracking again and tell u what happen Thank u!

calderonsin commented 3 years ago

it seems that it's detecting objects but the csv is still empty . these are the first lines: 1 (value:/home/juccalderondo/videos/Acevedo_6-18am_6-38am.MP4) movie file OpenCL not used ParseTrackerSettings: /home/juccalderondo/MT2/data/settings.ini ... ParseTrackerSettings - readed ParseTrackerSettings: 1 Video /home/juccalderondo/videos/Acevedo_6-18am_6-38am.MP4 was started from 0 frame with 29.97 fps Used GPU 0 0 : compute_capability = 610, cudnn_half = 0, GPU: NVIDIA TITAN Xp net.optimized_memory = 0 mini_batch = 1, batch = 8, time_steps = 1, train = 0 layer filters size/strd(dil) input output 0 Create CUDA-stream - 0 Create cudnn-handle 0 conv 32 3 x 3/ 1 416 x 416 x 3 -> 416 x 416 x 32 0.299 BF 1 max 2x 2/ 2 416 x 416 x 32 -> 208 x 208 x 32 0.006 BF 2 conv 64 3 x 3/ 1 208 x 208 x 32 -> 208 x 208 x 64 1.595 BF 3 max 2x 2/ 2 208 x 208 x 64 -> 104 x 104 x 64 0.003 BF 4 conv 128 3 x 3/ 1 104 x 104 x 64 -> 104 x 104 x 128 1.595 BF 5 conv 64 1 x 1/ 1 104 x 104 x 128 -> 104 x 104 x 64 0.177 BF 6 conv 128 3 x 3/ 1 104 x 104 x 64 -> 104 x 104 x 128 1.595 BF 7 max 2x 2/ 2 104 x 104 x 128 -> 52 x 52 x 128 0.001 BF 8 conv 256 3 x 3/ 1 52 x 52 x 128 -> 52 x 52 x 256 1.595 BF 9 conv 128 1 x 1/ 1 52 x 52 x 256 -> 52 x 52 x 128 0.177 BF 10 conv 256 3 x 3/ 1 52 x 52 x 128 -> 52 x 52 x 256 1.595 BF 11 max 2x 2/ 2 52 x 52 x 256 -> 26 x 26 x 256 0.001 BF 12 conv 512 3 x 3/ 1 26 x 26 x 256 -> 26 x 26 x 512 1.595 BF 13 conv 256 1 x 1/ 1 26 x 26 x 512 -> 26 x 26 x 256 0.177 BF 14 conv 512 3 x 3/ 1 26 x 26 x 256 -> 26 x 26 x 512 1.595 BF 15 conv 256 1 x 1/ 1 26 x 26 x 512 -> 26 x 26 x 256 0.177 BF 16 conv 512 3 x 3/ 1 26 x 26 x 256 -> 26 x 26 x 512 1.595 BF 17 max 2x 2/ 2 26 x 26 x 512 -> 13 x 13 x 512 0.000 BF 18 conv 1024 3 x 3/ 1 13 x 13 x 512 -> 13 x 13 x1024 1.595 BF 19 conv 512 1 x 1/ 1 13 x 13 x1024 -> 13 x 13 x 512 0.177 BF 20 conv 1024 3 x 3/ 1 13 x 13 x 512 -> 13 x 13 x1024 1.595 BF 21 conv 512 1 x 1/ 1 13 x 13 x1024 -> 13 x 13 x 512 0.177 BF 22 conv 1024 3 x 3/ 1 13 x 13 x 512 -> 13 x 13 x1024 1.595 BF 23 conv 1024 3 x 3/ 1 13 x 13 x1024 -> 13 x 13 x1024 3.190 BF 24 conv 1024 3 x 3/ 1 13 x 13 x1024 -> 13 x 13 x1024 3.190 BF 25 route 16 -> 26 x 26 x 512 26 reorg_old reorg_old / 2 26 x 26 x 512 -> 13 x 13 x2048 27 route 26 24 -> 13 x 13 x3072 28 conv 1024 3 x 3/ 1 13 x 13 x3072 -> 13 x 13 x1024 9.569 BF 29 conv 40 1 x 1/ 1 13 x 13 x1024 -> 13 x 13 x 40 0.014 BF 30 detection mask_scale: Using default '1.000000' Total BFLOPS 34.879 avg_outputs = 645209 Allocate additional workspace_size = 314.58 MB Loading weights from /home/juccalderondo/backup/yolo-obj_40000.weights... seen 64, trained: 2560 K-images (40 Kilo-batches_64) Done! Loaded 31 layers from weights-file

try to allocate additional workspace_size = 314.58 MB CUDA allocate done!

these are the last lines: Frame 34179: tracks = 0, time = 12 Detected 6 objects Finally 0 objects, 1, 2 Frame 34180: tracks = 0, time = 12 Detected 5 objects Finally 0 objects, 1, 2 Frame 34181: tracks = 0, time = 12 Detected 6 objects Finally 0 objects, 1, 2 Frame 34182: tracks = 0, time = 12 Detected 9 objects Finally 0 objects, 1, 2 Frame 34183: tracks = 0, time = 12 Detected 6 objects Finally 0 objects, 1, 2 Frame 34184: tracks = 0, time = 12 Detected 7 objects Finally 0 objects, 1, 2 Frame 34185: tracks = 0, time = 12

Nuzhny007 commented 3 years ago

I think that problem in this line: https://github.com/Smorodov/Multitarget-tracker/blob/master/src/Detector/YoloDarknetDetector.cpp#L184

  1. Try to comment all lines with white_list here: https://github.com/Smorodov/Multitarget-tracker/blob/master/example/examples.h#L657

  2. If the problem still persists then comment this line: https://github.com/Smorodov/Multitarget-tracker/blob/master/src/Detector/YoloDarknetDetector.cpp#L184

calderonsin commented 3 years ago

Hello, i tried both solutions and the csv is still empty and the video doesn't show tracking at all :( I suppose is because the non GUI envyroment

thanks!

Nuzhny007 commented 3 years ago

I think that here another problem. I see in your log:

Detected 6 objects
Finally 0 objects, 1, 2

Detector works but objects was absent. Comment also line 215: https://github.com/Smorodov/Multitarget-tracker/blob/master/src/Detector/YoloDarknetDetector.cpp#L215

calderonsin commented 3 years ago

Hello! The tracking is working now. I have a question, where can I read about the .csv? I wish to know what is the format of the output.

Thank u for your help!

Nuzhny007 commented 3 years ago

Here: https://github.com/Smorodov/Multitarget-tracker/blob/master/example/FileLogger.h#L119 First value - Frame id and other - detection parameters on this frame