Closed dcgrove closed 3 years ago
So I am not sure why it works, but I changed the config of my reolink camera from the sub stream to the higher resolution main stream and restarted the container. Then changed the config back to the sub stream and restarted the container again and now person detection is working as it should.
Not sure you are aware but you might want to remove your logs as your rtsp usernames and passwords are listed in it .
Not sure you are aware but you might want to remove your logs as your rtsp usernames and passwords are listed in it .
Good call. Thanks for the heads up.
Well it stopped working shortly after I posted this. 😤
What does it look like when you view the mjpeg feed from the camera in frigate after it stops working? Is that still showing a live view?
What does it look like when you view the mjpeg feed from the camera in frigate after it stops working? Is that still showing a live view?
The mjpeg camera works, but does not show any person detection. The car detection works as it should.
I can't think of a reason why it would stop working for a single object type. Tensorflow looks for all objects in a single pass, so there is nothing person specific about the detection pipeline. The only thing I can think of is that the person scores are just too low to meet the threshold value you have set. You should still see bounding boxes for person objects even if they are lower than the threshold. You could try lowering the threshold to see if that's the issue.
Here is a screen recording of the mjpeg camera when my daughter took the trash out. As you can see the bounding boxes for the car works, but there is no person detection.
Can you try removing the car from the tracked objects list to see if it works with person only?
Hello, person is the only object I have set up to detect, the link below is a screen recording of the mjpeg camera. No detection occurred.
https://share.icloud.com/photos/0L6chola6lg3nhWxhZRAt0ohQ
Here is my object config.
objects:
track:
- person
#- car
#- truck
#- bicycle
#- motorcycle
#- cat
#- dog
#- bird
#filters:
#person:
#threshold: 0.90
I recreated the container in Portainer and tested it again and person detection worked albeit rather slowly. Video and log files from portainer below.
https://share.icloud.com/photos/07Rg24hz9aRwDpuvEKHEZLfDA
2020-12-11T14:13:07.769669740Z Fontconfig error: Cannot load default config file,
2020-12-11T14:13:08.068115068Z ffprobe -v panic -show_error -show_streams -of json "rtsp://FamLr:XXXXXX@192.168.1.189/live",
2020-12-11T14:13:08.068901271Z Starting detection process: 18,
2020-12-11T14:13:08.069065098Z Attempting to load TPU as usb,
2020-12-11T14:13:08.205983451Z On connect called,
2020-12-11T14:13:10.201036515Z {'streams': [{'index': 0, 'codec_name': 'h264', 'codec_long_name': 'unknown', 'profile': '77', 'codec_type': 'video', 'codec_time_base': '1/30', 'codec_tag_string': '[0][0][0][0]', 'codec_tag': '0x0000', 'width': 1920, 'height': 1080, 'coded_width': 1920, 'coded_height': 1088, 'closed_captions': 0, 'has_b_frames': 0, 'pix_fmt': 'yuv420p', 'level': 40, 'chroma_location': 'left', 'field_order': 'progressive', 'refs': 1, 'is_avc': 'false', 'nal_length_size': '0', 'r_frame_rate': '15/1', 'avg_frame_rate': '15/1', 'time_base': '1/90000', 'start_pts': 36000, 'start_time': '0.400000', 'bits_per_raw_sample': '8', 'disposition': {'default': 0, 'dub': 0, 'original': 0, 'comment': 0, 'lyrics': 0, 'karaoke': 0, 'forced': 0, 'hearing_impaired': 0, 'visual_impaired': 0, 'clean_effects': 0, 'attached_pic': 0, 'timed_thumbnails': 0}}, {'index': 1, 'codec_name': 'pcm_alaw', 'codec_long_name': 'unknown', 'codec_type': 'audio', 'codec_time_base': '1/8000', 'codec_tag_string': '[0][0][0][0]', 'codec_tag': '0x0000', 'sample_fmt': 's16', 'sample_rate': '8000', 'channels': 1, 'bits_per_sample': 8, 'r_frame_rate': '0/0', 'avg_frame_rate': '0/0', 'time_base': '1/8000', 'start_pts': 2, 'start_time': '0.000250', 'bit_rate': '64000', 'disposition': {'default': 0, 'dub': 0, 'original': 0, 'comment': 0, 'lyrics': 0, 'karaoke': 0, 'forced': 0, 'hearing_impaired': 0, 'visual_impaired': 0, 'clean_effects': 0, 'attached_pic': 0, 'timed_thumbnails': 0}}]},
2020-12-11T14:13:10.205150293Z ffprobe -v panic -show_error -show_streams -of json "rtsp://Clayton:XXXXXX@192.168.1.142/live",
2020-12-11T14:13:10.742845887Z TPU found,
2020-12-11T14:13:12.432607371Z {'streams': [{'index': 0, 'codec_name': 'h264', 'codec_long_name': 'unknown', 'profile': '77', 'codec_type': 'video', 'codec_time_base': '1/20', 'codec_tag_string': '[0][0][0][0]', 'codec_tag': '0x0000', 'width': 1920, 'height': 1080, 'coded_width': 1920, 'coded_height': 1088, 'closed_captions': 0, 'has_b_frames': 0, 'pix_fmt': 'yuv420p', 'level': 40, 'chroma_location': 'left', 'field_order': 'progressive', 'refs': 1, 'is_avc': 'false', 'nal_length_size': '0', 'r_frame_rate': '15/1', 'avg_frame_rate': '10/1', 'time_base': '1/90000', 'start_pts': 32999, 'start_time': '0.366656', 'bits_per_raw_sample': '8', 'disposition': {'default': 0, 'dub': 0, 'original': 0, 'comment': 0, 'lyrics': 0, 'karaoke': 0, 'forced': 0, 'hearing_impaired': 0, 'visual_impaired': 0, 'clean_effects': 0, 'attached_pic': 0, 'timed_thumbnails': 0}}, {'index': 1, 'codec_name': 'pcm_alaw', 'codec_long_name': 'unknown', 'codec_type': 'audio', 'codec_time_base': '1/8000', 'codec_tag_string': '[0][0][0][0]', 'codec_tag': '0x0000', 'sample_fmt': 's16', 'sample_rate': '8000', 'channels': 1, 'bits_per_sample': 8, 'r_frame_rate': '0/0', 'avg_frame_rate': '0/0', 'time_base': '1/8000', 'start_pts': 4, 'start_time': '0.000500', 'bit_rate': '64000', 'disposition': {'default': 0, 'dub': 0, 'original': 0, 'comment': 0, 'lyrics': 0, 'karaoke': 0, 'forced': 0, 'hearing_impaired': 0, 'visual_impaired': 0, 'clean_effects': 0, 'attached_pic': 0, 'timed_thumbnails': 0}}]},
2020-12-11T14:13:12.443941801Z Camera capture process started for family_livingroom: 31,
2020-12-11T14:13:12.447874890Z Creating ffmpeg process...Camera process started for family_livingroom: 32,
2020-12-11T14:13:12.448131161Z ,
2020-12-11T14:13:12.448169086Z ffmpeg -hide_banner -loglevel panic -hwaccel vaapi -hwaccel_device /dev/dri/renderD128 -hwaccel_output_format yuv420p -avoid_negative_ts make_zero -fflags nobuffer -flags low_delay -strict experimental -fflags +genpts+discardcorrupt -rtsp_transport tcp -stimeout 5000000 -use_wallclock_as_timestamps 1 -i rtsp://FamLr:XXXXXX@192.168.1.189/live -f rawvideo -pix_fmt yuv420p pipe:,
2020-12-11T14:13:12.452117335Z Camera capture process started for formal_livingroom: 34,
2020-12-11T14:13:12.455200215Z Camera process started for formal_livingroom: 36,
2020-12-11T14:13:12.458038880Z Camera capture process started for driveway: 37,
2020-12-11T14:13:12.463527445Z Camera process started for driveway: 38,
2020-12-11T14:13:12.466724667Z Creating ffmpeg process...,
2020-12-11T14:13:12.466903677Z ffmpeg -hide_banner -loglevel panic -hwaccel vaapi -hwaccel_device /dev/dri/renderD128 -hwaccel_output_format yuv420p -avoid_negative_ts make_zero -fflags nobuffer -flags low_delay -strict experimental -fflags +genpts+discardcorrupt -rtsp_transport tcp -stimeout 5000000 -use_wallclock_as_timestamps 1 -i rtsp://admin:XXXXXX@192.168.1.19:554/h264Preview_01_sub -f segment -segment_time 10 -segment_format mp4 -reset_timestamps 1 -strftime 1 -c copy -an -map 0 /cache/driveway-%Y%m%d%H%M%S.mp4 -r 7 -f rawvideo -pix_fmt yuv420p pipe:,
2020-12-11T14:13:12.467489965Z Creating ffmpeg process...,
2020-12-11T14:13:12.467619425Z ffmpeg -hide_banner -loglevel panic -hwaccel vaapi -hwaccel_device /dev/dri/renderD128 -hwaccel_output_format yuv420p -avoid_negative_ts make_zero -fflags nobuffer -flags low_delay -strict experimental -fflags +genpts+discardcorrupt -rtsp_transport tcp -stimeout 5000000 -use_wallclock_as_timestamps 1 -i rtsp://Clayton:XXXXXX@192.168.1.142/live -f segment -segment_time 10 -segment_format mp4 -reset_timestamps 1 -strftime 1 -c copy -an -map 0 /cache/formal_livingroom-%Y%m%d%H%M%S.mp4 -f rawvideo -pix_fmt yuv420p pipe:,
2020-12-11T14:13:12.489134052Z * Serving Flask app "detect_objects" (lazy loading),
2020-12-11T14:13:12.489152433Z * Environment: development,
2020-12-11T14:13:12.489155351Z * Debug mode: off,
2020-12-11T14:13:17.492707882Z driveway: ffmpeg sent a broken frame. something is wrong.,
2020-12-11T14:13:17.492743575Z driveway: ffmpeg process is not running. exiting capture thread...,
2020-12-11T14:13:22.477346265Z Creating ffmpeg process...,
2020-12-11T14:13:22.477396681Z ffmpeg -hide_banner -loglevel panic -hwaccel vaapi -hwaccel_device /dev/dri/renderD128 -hwaccel_output_format yuv420p -avoid_negative_ts make_zero -fflags nobuffer -flags low_delay -strict experimental -fflags +genpts+discardcorrupt -rtsp_transport tcp -stimeout 5000000 -use_wallclock_as_timestamps 1 -i rtsp://admin:XXXXXX@192.168.1.19:554/h264Preview_01_sub -f segment -segment_time 10 -segment_format mp4 -reset_timestamps 1 -strftime 1 -c copy -an -map 0 /cache/driveway-%Y%m%d%H%M%S.mp4 -r 7 -f rawvideo -pix_fmt yuv420p pipe:,
2020-12-11T14:13:25.592244320Z driveway: ffmpeg sent a broken frame. something is wrong.,
2020-12-11T14:13:25.592319892Z driveway: ffmpeg process is not running. exiting capture thread...,
2020-12-11T14:13:32.494965708Z Creating ffmpeg process...,
2020-12-11T14:13:32.495005114Z ffmpeg -hide_banner -loglevel panic -hwaccel vaapi -hwaccel_device /dev/dri/renderD128 -hwaccel_output_format yuv420p -avoid_negative_ts make_zero -fflags nobuffer -flags low_delay -strict experimental -fflags +genpts+discardcorrupt -rtsp_transport tcp -stimeout 5000000 -use_wallclock_as_timestamps 1 -i rtsp://admin:XXXXXX@192.168.1.19:554/h264Preview_01_sub -f segment -segment_time 10 -segment_format mp4 -reset_timestamps 1 -strftime 1 -c copy -an -map 0 /cache/driveway-%Y%m%d%H%M%S.mp4 -r 7 -f rawvideo -pix_fmt yuv420p pipe:,
Try a simpler set of input args for the driveway
camera:
driveway:
ffmpeg:
input: 'rtsp://XXXXXX:XXXXX@192.168.1.19:554//h264Preview_01_sub'
input_args:
- '-avoid_negative_ts'
- make_zero
- '-fflags'
- +genpts+discardcorrupt
- '-use_wallclock_as_timestamps'
- '1'
- '-an'
- '-dn'
- '-flags'
- '-global_header'
With updated input args. It detected me after some time once, but failed to detect me again. This was all taken on one continuous recording.
Can you try removing your mask?
That did it! I don't know why the mask is the issue though. I followed the steps outlined in the docs to generate the image points. Do you know of a tool that can reverse the process to identify the polygon in an image based on points you feed it?
The best way to test in this version is to add the mask coordinates as a zone for the camera to see the polygon drawn on the preview feed.
Here is the zone drawn on the preview feed.
The mask is going to ignore any motion inside that polygon. Your mask needs to be the opposite.
Closing as this was user error. I appreciate the assistance getting this sorted out!
Describe the bug Hello, I have three cameras (2x wyzeV2 with RTSP firmware and 1x Reolink RLC410-5mp) set up. All three are being scanned for the objects set up in the config.yaml below. The reolink (driveway) camera has not sent an MQTT message for person detection in two days despite numerous people in the detection zone.. Last person screenshot for the reolink is below. car/truck detection for this camera is working as expected, and person/cat detection for the wyze camera's are working as expected.
Version of frigate 0.7.3
Config file
Logs
Frigate debug stats
FFprobe from your camera
Run the following command and paste output below
Screenshots
Computer Hardware
Camera Info: