Image analysis for making animal measurements

1. A computer-implemented image analysis process for making animal measurements, including:

accessing image data representing an image including an animal object representing an animal;

identifying coordinate values for a surface boundary of said animal object;

selecting one or more start points on the surface boundary, wherein selecting the start points includes determining whether two or more potential start points correspond to a predefined expected animal shape;

generating surface boundary data, representing at least a portion of the surface boundary, using a gradient of surface boundary coordinate values starting at each of the selected one or more start points;

generating distance data representing distances between a plurality of boundary locations on the surface boundary, using the surface boundary data;

generating size data, representing the animal's size, using the distance data; and

generating weight data, representing a weight of the animal, based on the size data, including accessing weight model data representing a predetermined weight model which associates size values with weight values for the animal.

2. The process as claimed in claim 1 , wherein determining whether the two or more potential start points correspond to the predefined expected animal image shape includes comparing an expected animal width to a distance between the two potential start points.

3. The process as claimed in claim 1 , further including selecting one or more subsets of the image and selecting the start points as points within the subset of the image, wherein selecting the start points includes selecting one or more respective search lines across the image and selecting the one or more start points as points on the respective search lines.

4. The process as claimed in claim 1 , wherein selecting the one or more start points includes analysing the gradient of the image data stored within a subset of the image.

5. The process as claimed in claim 4 , wherein generating the boundary data includes iteratively using at least one neighbourhood filter.

6. The process as claimed in claim 5 , wherein an analysis to determine the gradient within the neighbourhood filter commences at an edge of the image data.

7. The process as claimed in claim 5 , wherein an analysis to determine the gradient within the neighbourhood filter commences at a point in the image data.

8. The process as claimed in claim 5 , wherein the analysis of the gradient of the image data within the subset of the image includes selecting one or more maxima or minima from the gradient of the data within the subset.

9. The process as claimed in claim 1 , wherein selecting the one or more start points includes: determining a longitudinally central portion of the animal object using the image's intensity profile; and selecting the one or more start points in the longitudinally central portion.

10. The process as claimed in claim 9 , wherein the longitudinally central portion excludes a head portion and a tail portion of the animal object.

11. The process as claimed in claim 1 , wherein the surface boundary data include side boundary data, and wherein the side boundary data represents two longitudinal side boundaries of the animal object, and generating the side boundary data includes trimming at least one longitudinal end from different boundary portions associated with respective different sides of the animal object, based on the different boundary portions being closer than a predefined minimum distance associated with an expected minimum animal width.

12. The process as claimed in claim 11 , further including generating data representing at least one modelled end boundary, corresponding to estimated longitudinal ends of the animal object, using data representing an end model.

13. The process as claimed in claim 1 , wherein generating the size data includes generating width measurement data, representing lateral width measurements of the animal object, and/or area or section of area measurement data, representing area measurements of the animal object.

14. The process as claimed in claim 1 , wherein generating the size data includes using installation data representing imaging parameters of an imaging device and physical parameters of an apparatus for generating the image data.

15. The process as claimed in claim 1 , wherein the weight model includes a predetermined allometric relationship between size and weight, including a multivariate model associating size values and weight values.

16. The process as claimed in claim 1 , further including generating orientation data representing angular orientations of the distances, wherein generating the size data includes using the orientation data.

17. A computer-implemented image analysis process for making animal measurements, including:

accessing image data representing an image including an animal objectg representing an animal;

identifying coordinate values for a surface boundary of said animal object;

selecting one or more start points on the surface boundary;

generating surface boundary data, representing at least a portion of the surface boundary, using a gradient of surface boundary coordinate values starting at each of the selected one or more start points;

generating distance data representing distances between a plurality of boundary locations on the surface boundary, using the surface boundary data;

generating size data, representing the animal's size, using the distance data;

generating weight data, representing a weight of the animal, based on the size data, including accessing weight model data representing a predetermined weight model which associates size values with weight values for the animal, including determining whether the surface boundary data are relevant based on expected animal size limits for the animal object.

18. The process as claimed in claim 17 , wherein the expected animal size limits include acceptable animal widths, animal lengths, animal areas or sections of area.

19. An image analysis system for making animal measurements, including:

at least one image capture device for capturing an image of an animal, and

generating image data representing the image; and

a processor for:

detecting a boundary contour of an animal object representing the animal in the image by: starting at at least one start point on the boundary contour wherein selecting the at least one start point includes determining whether two or more potential start points correspond to a predefined expected animal shape; identifying coordinate values of neighbouring points along the boundary contour using the image's intensity gradient; generating boundary data representing at least a portion of the boundary contour by selecting a maxima or minima of the gradient of intensity values within subset array of the image data representing surface boundary coordinates of the animal; generating distance data, representing distances between a plurality of boundary locations on the surface boundary, using said surface boundary data; generating size data, representing the animal's size, using the distance data; and generating weight data, representing a weight of the animal, based on the size data, including accessing weight model data representing a predetermined weight model which associates size values with weight values for the animal.