Search

EP-4740730-A1 - ANIMAL HUSBNDRY METHOD AND SYSTEM

EP4740730A1EP 4740730 A1EP4740730 A1EP 4740730A1EP-4740730-A1

Abstract

Animal husbandry method and system, the method comprising the following steps: - keeping a group of animals roaming about freely in an enclosed roaming area; - automatically monitoring at least one animal; - repeatedly detecting a number of reference points of said animal; - analysing the relative positions of the detected reference points; - determining the pose of the animal from the analysed relative positions of the detected reference points; - determining from said pose the position of the animal's head relative to its body; - establishing a most likely animal behaviour or condition associated with said pose and head position. The method further develops and improves the monitoring of the animals. The determined animal pose is used for animal behaviour and/or condition analysis, allowing for specific, relevant follow-up actions, thus enabling improved animal welfare.

Inventors

  • VAN DER SLUIS, Joram Robin
  • Pena Trapero, Juan Manuel
  • VAN ADRICHEM, PAULUS JACOBUS MARIA

Assignees

  • Lely Patent N.V.

Dates

Publication Date
20260513
Application Date
20251111

Claims (20)

  1. Animal husbandry method comprising the following steps: - keeping a group of animals (1, 6, 7, 8, 9) roaming about freely in an enclosed roaming area; - automatically monitoring at least one animal (1, 6, 7, 8, 9); - repeatedly detecting a number of reference points (2, 3, 4, 5) of said animal (1, 6, 7, 8, 9); - analysing the relative positions of the detected reference points (2, 3, 4, 5); - determining the pose of the animal (1, 6, 7, 8, 9) from the analysed relative positions of the detected reference points (2, 3, 4, 5); - determining from said pose the position of the animal's head relative to its body; - establishing a most likely animal behaviour or condition associated with said pose and head position.
  2. Method according to claim 1, wherein the determining of the pose of the animal (1, 6, 7, 8, 9) is performed using a neural network.
  3. Method according to any one of claims 1 - 2, wherein the reference points (2, 3, 4, 5) are situated substantially on the spine of the animal (1, 6, 7, 8, 9).
  4. Method according to any one of claims 1 - 3, wherein the step of automatically monitoring at least one animal (1, 6, 7, 8, 9) includes tracking the location of said animal (1, 6, 7, 8, 9).
  5. Method according to claim 4, wherein, in the determining and establishing steps, the location and/or the surroundings of the animal (1, 6, 7, 8, 9) are taken into account.
  6. Method according to claim 5, wherein one of the known animal behaviours is drinking, wherein said behaviour is selected in case the determined pose and head position of the animal (1, 6, 7, 8, 9) indicates that its head is down in the vicinity of a drinking place.
  7. Method according to any one of claims 5 - 6, wherein one of the known animal behaviours is eating, wherein said behaviour is selected in case the determined pose and head position of the animal (1, 6, 7, 8, 9) indicates that its head is down in the vicinity of a feeding place (10).
  8. Method according to any one of claims 5 - 7, wherein one of the known animal behaviours is offspring drinking, wherein said behaviour is selected in case the determined pose and head position of the animal (1, 6, 7, 8, 9) indicates that its head is in the vicinity of the back side of a further, larger animal (1, 6, 7, 8, 9).
  9. Method according to claim 8, wherein the duration of the offspring drinking is determined.
  10. Method according to claim 9, wherein the determined duration of the offspring drinking is used to estimate the amount drunk.
  11. Method according to any one of claims 8 - 10, wherein it is registered which mother animal and/or which teats of said mother animal are involved in said offspring drinking.
  12. Method according to any one of claims 1 - 11, wherein one of the known animal behaviours is rolling over, wherein said behaviour is selected in case the determined pose of the animal (1, 6, 7, 8, 9) indicates that the pose has been alternating between two positions.
  13. Method according to claim 12, wherein one of the known animal conditions is mastitis, wherein said condition is selected in case the alternating frequency lies within a first interval.
  14. Method according to claim 12, wherein one of the known animal conditions is birthing, wherein said condition is selected in case the alternating frequency lies within a second interval.
  15. Method according to claim 14 when dependent on any one of claims 9 - 11, wherein an alarm is generated in case a minimum duration of offspring drinking has not been achieved within a set time interval after the birthing.
  16. Method according to any one of claims 5 - 15, wherein one of the known animal behaviours is heat, wherein said behaviour is selected in case the determined pose and head position of the animal (1, 6, 7, 8, 9) indicates that its head is near or on top of another animal (1, 6, 7, 8, 9).
  17. Method according to any one of claims 1 - 16, wherein one of the known animal behaviours is heat, wherein said behaviour is selected in case the determined pose of the animal (1, 6, 7, 8, 9) indicates that the animal (1, 6, 7, 8, 9) is restless.
  18. Method according to any one of claims 1 - 17, wherein the method further comprises the steps of: - determining whether the selected animal behaviour or condition necessitates a specific action associated with said behaviour; - executing said specific action in case of necessity.
  19. Method according to any one of claims 1 - 18, wherein the method further comprises the steps of: - automatically monitoring also at least one person and/or machine in said enclosed roaming area; - repeatedly detecting a number of reference points of said person and/or machine; - analysing the relative positions of the detected reference points; - determining the pose of the person and/or machine from the analysed relative positions of the detected reference points; - registering the determined pose.
  20. Animal husbandry system comprising a group of animals (1, 6, 7, 8, 9) roaming about freely in an enclosed roaming area, monitoring means and at least one processor with image analysis means, said processor being configured to perform the method of any one of claims 1 - 19.

Description

The present invention relates to an animal husbandry method and system, the method comprising the following steps: keeping a group of animals roaming about freely in an enclosed roaming area;automatically monitoring at least one animal;repeatedly detecting a number of reference points of said animal;analysing the relative positions of the detected reference points;determining the pose of the animal from the analysed relative positions of the detected reference points. Such methods are known in the art. Monitoring animals, such as cows in a stable, plays an important role in modern animal husbandry systems. Gathering reliable animal data within a group or herd is of increasing importance. EP3933766 deals with a system and method for monitoring livestock using a camera arranged above the livestock for gathering data relating to a motion of a tracked individual animal, using a set of identified reference points, each associated with a body part location. The motion data thus obtained are used to establish a mobility level and a mobility score for the animal. The mobility score can be used to detect mobility issues such as lameness. WO2022/077113 shows an animal management system with a number of imaging devices in which images are processed using artificial intelligence for detecting and locating animals. An identification or classification of the detected animal is generated with a confidence score. US2022/0398861 relates to an image-based animal object detection apparatus using bounding boxes and key points. With the aid of a neural network animal detection information is being extracted, such as information about the class and a pose of the animal object detected. The pose information of the animal object may distinguish various poses such as sitting, standing, mounting behaviour, rollover and dog sitting. There is a need for an improved, economical and animal friendly animal husbandry method and system. It is an object of the present invention to provide such an improved method and system. The invention achieves the object at least in part by means of a method according to claim 1, in particular an animal husbandry method comprising the following steps: keeping a group of animals roaming about freely in an enclosed roaming area;automatically monitoring at least one animal;repeatedly detecting a number of reference points of said animal;analysing the relative positions of the detected reference points;determining the pose of the animal from the analysed relative positions of the detected reference points;determining from said pose the position of the animal's head relative to its body;establishing a most likely animal behaviour or condition associated with said pose and head position. In this way, the method is versatile and further develops and improves the monitoring of the animals. A user-friendly and cheap method enabling improved animal welfare is thus realized. The invention is based on the insight that a determined animal pose in combination with a determined head position can be highly advantageously used for accurate animal behaviour and/or condition analysis, allowing for specific, relevant follow-up actions. This results in a secure method. The invention further achieves the object at least in part by means of a system according to claim 20, in particular an animal husbandry system comprising a group of animals roaming about freely in an enclosed roaming area, monitoring means and at least one processor with image analysis means, said processor being configured to perform the method of any one of claims 1 - 19. Suitable and advantageous embodiments are described in the dependent claims, as well as in the description below. According to a first embodiment of the invention, the determining of the pose of the animal is performed using a neural network. Such a neural network is highly suitable for such a task because of its learning capacity. In a further embodiment, the reference points are situated substantially on the spine of the animal. Such reference points are very suitable to determine the pose of an animal. Preferably, the number of reference points used is four or more. In practice, this has proven to give good results. Advantageously, the step of automatically monitoring at least one animal includes tracking the location of said animal. Knowing the location of an animal in the course of time is very important. According to a further embodiment, in the determining and establishing steps, the location and/or the surroundings of the animal are taken into account. In this way, these steps are more accurate. In accordance with yet another embodiment, one of the known animal behaviours is drinking, wherein said behaviour is selected in case the determined pose and relative head position of the animal indicates that its head is down in the vicinity of a drinking place. Thus, a reliable detection of the animal drinking is realised. In yet a further embodiment, one of the known animal behaviours is