Automated poultry or livestock welfare tracking
Various proxies can be used to qualitatively or quantitatively assess welfare of a poultry flock. For example, as shown and described herein, one or more of a video-based monitoring technique or audio-based monitoring technique can be used. Such monitoring can include a video-based (e.g., image based) technique where indicia of movement of individual members of the flock are automatically tracked. In addition, or instead, an audio-based technique can be used, such as for classification of vocalization. Various metrics can be stored or reported using such techniques and one or more alerts can be generated, such as indicative of flock welfare or a change in flock welfare. Techniques described herein can include identifying paths traversed by respective members of a flock (e.g., individual animals), and then extracting indications such as corresponding to a distribution of such paths. Such paths can be defined by corresponding movement trajectories associated with individual members.
1 . A method for assessing poultry or livestock welfare, comprising:
receiving video data corresponding to at least a portion of a flock within a field of view;
using the received video data, automatically establishing a plurality of movement trajectories corresponding to respective portions of the field of view, the movement trajectories corresponding to tracked individual members of the flock; and
automatically determining a first metric corresponding to a distribution of the movement trajectories including a degree of uniformity of movement of the flock;
wherein the determining the plurality of movement trajectories corresponding to respective portions of the field of view includes suppressing a contribution from non-flock objects within the field of view.
2 . The method of claim 1 , wherein the plurality of movement trajectories includes groups of trajectories corresponding to paths taken by tracked individual members of the flock.
3 . The method of claim 1 , wherein the first metric corresponds to a degree of movement of the flock; and
wherein the degree of movement comprises a relative indication of a proportion of the flock that have moved over a specified duration or versus a prior movement determination.
4 . The method of claim 1 , comprising determining a welfare indicator using the first metric or a value indicative of a population of first metrics.
5 . The method of claim 4 , wherein the welfare indicator comprises a determination that the flock is exhibiting evidence of foot lesions.
6 . The method of claim 4 , wherein the welfare indicator comprises a determination that the flock is experiencing incidence of sickness or infection.
7 . The method of claim 1 , comprising comparing the first metric or a value indicative of a population of first metrics to a threshold; and
in response, triggering an alert when the first metric traverses the threshold.
8 . The method of claim 7 , wherein the threshold is established in response to one or more of data indicative of (1) an absolute spatial distribution of flock members or (2) a relative distribution of flock members with respect to each other.
9 . The method of claim 7 , wherein the first metric is used to monitor or adjust an environmental parameter that includes one or more of light level, temperature, or ventilation state.
10 . The method of claim 1 , wherein substantially all individual members of the flock have associated movement trajectories within the field of view.
11 . The method of claim 1 , comprising receiving audio data corresponding to another portion of the flock; and
identifying flock-related vocalizations in the received audio data.
12 . The method of claim 11 , comprising automatically classifying respective vocalizations from amongst the identified vocalizations as to whether respective vocalizations indicate distress.
13 . The method of claim 12 , wherein the automatically classifying the respective vocalizations includes classifying whether the respective vocalizations indicate comfort or distress.
14 . The method of claim 12 , wherein the automatically classifying the respective vocalizations includes using a second metric indicative of respective vocalizations.
15 . The method of claim 14 , wherein the automatically classifying the respective vocalizations includes identifying a profile corresponding to at least one of a comfort vocalization or a distress vocalization and correlating a detected vocalization with the profile to provide the second metric.
16 . The method of claim 15 , wherein the profile comprises a temporal profile.
17 . The method of claim 15 , wherein the profile comprises a frequency-domain profile.
18 . The method of claim 12 , wherein the automatically classifying the respective vocalizations includes applying a machine-learning model trained to label an audio stream in manner indicative of a classified vocalization.
19 . The method of claim 11 , wherein receiving the audio data includes capturing audio using at least one directional microphone.
20 . The method of claim 19 , wherein the directional microphone has a shotgun receive pattern in a specified plane with a main lobe extending axially from the microphone in a boresight direction.
21 . The method of claim 1 , wherein the automatically establishing the plurality of movement trajectories and the automatically determining a first metric are performed using an embedded data processing system located at a livestock or poultry site being monitored.
22 . The method of claim 21 , wherein the embedded data processing system is located on or within a structure housing poultry or livestock.
23 . A processor-readable medium comprising instructions that, when performed by at least one processor circuit, cause the processor circuit to perform the method of claim 1 .
24 . A system for assessing livestock welfare, comprising:
an audio processor circuit coupled to a directional microphone, the directional microphone oriented to receive sound generated by a flock;
a video processor circuit coupled to a downward-facing imaging device arranged to image the flock;
a processor circuit coupled to the video processor circuit and the audio processor circuit or included as a portion of at least one of the video processor circuit or the audio processor circuit; and
a memory circuit communicatively coupled to the processor circuit;
wherein the processor circuit is configured to receive instructions stored using the memory circuit, the instructions, when executed, causing the processor circuit to:
receive video data from the video processor circuit, the video data corresponding to at least a portion of the flock within a field of view of the downward-facing imaging device;
using the received video data, automatically establish a plurality of movement trajectories corresponding to respective portions of the field of view, the movement trajectories corresponding to tracked individual members of the flock;
automatically determine a first metric corresponding to a distribution of the movement trajectories including a degree of uniformity of movement of the flock;
wherein the determining the plurality of movement trajectories corresponding to respective portions of the field of view includes suppressing a contribution from non-flock objects within the field of view.
25 . The system of claim 24 , further comprising the imaging device and the directional microphone.
26 . The system of claim 24 , wherein the plurality of movement trajectories includes groups of trajectories corresponding to paths taken by tracked individual members of the flock.
27 . The system of claim 24 , wherein the first metric corresponds to a degree of movement of the flock; and
wherein the degree of movement comprises a relative indication of a proportion of the flock that have moved over a specified duration or versus a prior movement determination.
28 . The system of claim 24 , wherein the instructions, when executed, cause the processor circuit to:
compare the first metric or a value indicative of a population of first metrics to a threshold; and in response, trigger an alert when the first metric traverses the threshold.
29 . The system of claim 24 , wherein the instructions, when executed, cause the processor circuit to determine a welfare indicator using the first metric or a value indicative of a population of first metrics.
30 . The system of claim 29 , wherein the welfare indicator comprises a determination that the flock is exhibiting evidence of foot lesions.
31 . The system of claim 29 , wherein the welfare indicator comprises a determination that the flock is experiencing incidence of sickness or infection.
32 . The system of claim 24 , wherein the instructions, when executed, cause the processor circuit to compare the first metric or a value indicative of a population of first metrics to a threshold; and
in response, to trigger an alert when the first metric traverses the threshold.
33 . The system of claim 32 , wherein the instructions, when executed, cause the processor circuit to establish the threshold in response to one or more of data indicative of (1) an absolute spatial distribution of flock members or (2) a relative distribution of flock members with respect to each other.
34 . The system of claim 32 , wherein the first metric is used to monitor or adjust an environmental parameter that includes one or more of light level, temperature, or ventilation state.
35 . The system of claim 24 , wherein substantially all individual members of the flock have associated movement trajectories within the field of view.
36 . The system of claim 24 , wherein the instructions, when executed, cause the processor circuit to receive audio data corresponding to another portion of the flock; and
identifying flock-related vocalizations in the received audio data.
37 . The system of claim 36 , wherein the instructions, when executed, cause the processor circuit to classify respective vocalizations from amongst the identified vocalizations as to whether respective vocalizations indicate distress.
38 . The system of claim 36 , wherein the instructions, when executed, cause the processor circuit to classify the respective vocalizations includes classifying whether the respective vocalizations indicate comfort or distress.
39 . The system of claim 38 , wherein the classifying the respective vocalizations includes using a second metric indicative of respective vocalizations.
40 . The system of claim 39 , wherein the classifying the respective vocalizations includes identifying a profile corresponding to at least one of a comfort vocalization or a distress vocalization and correlating a detected vocalization with the profile to provide the second metric.
41 . The system of claim 40 , wherein the profile comprises a temporal profile.
42 . The system of claim 39 , wherein the classifying the respective vocalizations includes applying a machine-learning model trained to label an audio stream in manner indicative of a classified vocalization.
43 . The system of claim 24 , wherein the directional microphone has a shotgun receive pattern in a specified plane with a main lobe extending axially from the microphone in a boresight direction.