Vision system with automatic teat detection

Patent metadata

Field	Value
Publication number	US-9807971-B1
Application number	US-201615239300-A
Country	US
Kind code	B1
Filing date	Aug 17, 2016
Priority date	Aug 17, 2016
Publication date	Nov 7, 2017
Grant date	Nov 7, 2017

Abstract

A system including a 3D camera, memory, and a processor. The processor is configured to obtain the 3D image, identify one or more regions within the 3D image comprising depth values greater than a depth value threshold, and apply the thigh gap detection rule set to identify a thigh gap region. The processor is further configured to demarcate an access region within the thigh gap region, demarcate a teat detection region, partition the 3D image within the teat detection region to generate a plurality of image depth planes, and examine each of the plurality of image depth planes. The processor is further configured to identify one or more teat candidates within the image depth plane, apply the teat detection rule set to the one or more teat candidates to identify one or more teats, and determine position information for the one or more teats.

First claim

The invention claimed is:

1. A teat detection method comprising:
obtaining, by a processor, a three-dimensional (3D) image of a rearview of a dairy livestock in a stall, wherein:
the dairy livestock is oriented in the 3D image with respect to:
an x-axis corresponding with a horizontal dimension of the 3D image,
a y-axis corresponding with a vertical dimension of the 3D image, and
a z-axis corresponding with a depth dimension into the 3D image; and
each pixel of the 3D image is associated with a depth value along the z-axis;

identifying, by the processor, one or more regions within the 3D image comprising depth values greater than a depth value threshold;
applying, by the processor, a thigh gap detection rule set to the one or more regions to identify a thigh gap region among the one or more regions, wherein the thigh gap region comprises an area between hind legs of the dairy livestock;
demarcating, by the processor, an access region within the thigh gap region, wherein the access region is defined by:
a first vertical edge,
a second vertical edge,
a first upper edge spanning between the first vertical edge and the second vertical edge, and
a first lower edge spanning between the first vertical edge and the second vertical edge;

demarcating, by the processor, a teat detection region, wherein the teat detection region is defined by:
a third vertical edge extending vertically from the first vertical edge of the access region,
a fourth vertical edge extending vertically from the second vertical edge of the access region,
a second upper edge spanning between the third vertical edge and the fourth vertical edge, and
a second lower edge spanning between the third vertical edge and the fourth vertical edge;

partitioning, by the processor, the 3D image within the teat detection region along the z-axis to generate a plurality of image depth planes;
examining, by the processor, each of the plurality of image depth planes, wherein examining each of the image depth planes comprises:
identifying one or more teat candidates within the image depth plane; and
applying a teat detection rule set to the one or more teat candidates to identify one or more teats; and

determining, by the processor, position information for the one or more teats.

2. The method of claim 1 , wherein:
the thigh gap detection rule set identifies a marker positioned between the hind legs of the dairy livestock at a lower edge of the 3D image; and
applying the thigh gap detection rule set to the one or more regions to identify the thigh gap region comprises:
discarding regions from the one or more regions that do not comprise the marker; and
identifying a region from among the one or more regions that comprises the marker as the thigh gap region.

3. The method of claim 1 , wherein:
the teat detection rule set indicates a minimum area value to be considered a teat; and
applying the teat detection rule set to identify the one or more teats comprises:
comparing the area of each of the one or more teat candidates to the minimum area value to be considered a teat;
discarding teat candidates from the one or more teat candidates with an area less than the minimum area value to be considered a teat; and
identifying a teat candidate from among the one or more teat candidates as a teat when the teat candidate has an area greater than or equal to the minimum area value to be considered a teat.

4. The method of claim 1 , wherein:
the teat detection rule set indicates a minimum height position with respect to the y-axis to be considered a teat; and
applying the teat detection rule set to identify the one or more teats comprises:
comparing the height position of each of the one or more teat candidates to the minimum height position to be considered a teat;
discarding teat candidates from the one or more teat candidates with an height position less than the minimum height position to be considered a teat; and
identifying a teat candidate from among the one or more teat candidates as a teat when the teat candidate has a height position greater than or equal to the minimum height position to be considered a teat.

5. The method of claim 1 , wherein:
the teat detection rule set indicates a minimum width value to be considered a teat; and
applying the teat detection rule set to the one or more teat candidates to identify the one or more teats comprises:
comparing the width of each of the one or more teat candidates to the minimum width value to be considered a teat;
discarding teat candidates from the one or more teat candidates with a width less than the minimum width value to be considered a teat; and
identifying a teat candidate from among the one or more teat candidates as a teat when the teat candidate has a width greater than or equal to the minimum width value to be considered a teat.

6. The method of claim 1 , wherein:
the teat detection rule set indicates a maximum width value to be considered a teat; and
applying the teat detection rule set to the one or more teat candidates to identify the one or more teats comprises:
comparing the width of each of the one or more teat candidates to the maximum width value to be considered a teat;
discarding teat candidates from the one or more teat candidates with a width greater than the maximum width value to be considered a teat; and
identifying a teat candidate from among the one or more teat candidates as a teat when the teat candidate has a width less than or equal to the maximum width value to be considered a teat.

7. The method of claim 1 , wherein:
the thigh gap detection rule set indicates a minimum area value to be considered the thigh gap region; and
applying the thigh gap detection rule set to the one or more regions to identify the thigh gap region comprises:
comparing the area of each of the one or more regions to the minimum area value to be considered the thigh gap region;
discarding regions from the one or more regions with an area less than the minimum area value to be considered the thigh gap region; and
identifying a region from among the one or more regions as the thigh gap region when the region has an area greater than or equal to the minimum area value to be considered the thigh gap region.

8. A vision system comprising:
a three-dimensional (3D) camera configured to obtain a 3D image of a rearview of a dairy livestock in a stall, wherein:
the dairy livestock is oriented in the 3D image with respect to:
an x-axis corresponding with a horizontal dimension of the 3D image,
a y-axis corresponding with a vertical dimension of the 3D image, and
a z-axis corresponding with a depth dimension into the 3D image; and

each pixel of the 3D image is associated with a depth value along the z-axis;

a memory configured to store:
a thigh gap detection rule set; and
a teat detection rule set; and

a processor communicatively coupled to the 3D camera and the memory, and configured to:
obtain the 3D image;
identify one or more regions within the 3D image comprising depth values greater than a depth value threshold;
apply the thigh gap detection rule set to the one or more regions to identify a thigh gap region among the one or more regions, wherein the thigh gap region comprises an area between hind legs of the dairy livestock;
demarcate an access region within the thigh gap region, wherein the access region is defined by:
a first vertical edge,
a second vertical edge,
a first upper edge spanning between the first vertical edge and the second vertical edge, and
a first lower edge spanning between the first vertical edge and the second vertical edge;

demarcate a teat detection region, wherein the teat detection region is defined by:
a third vertical

Assignees

• Technologies Holdings Corp

Inventors

• Foresman Mark A
• Prevost Bradley J

Classifications

• G06V10/751

  Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching · CPC title

• G06V10/44

  Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components · CPC title

• G06V10/25

  Determination of region of interest [ROI] or a volume of interest [VOI] · CPC title

• G06V10/50

  by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis · CPC title

• G06T7/001

  using an image reference approach · CPC title