System for high performance, AI-based dairy herd management and disease detection
Systems and methods for detecting udder disease based on machine learning methods and complementary supporting techniques are presented. Included are methods for assembling time sequences of images of each animal of a herd or set for subsequent use in per-animal image analysis for disease detection. Methods presented also include image pre-processing methods used prior to image analysis, resulting in contrast and resolution optimization such as appropriate image intensity level adjust wilt and resolution down sampling for more rapid and more accurate disease detection. Combinatorial techniques for compositing whole-udder images or udder-quarter images from partial images captures are described. Methods are provided for power usage optimization in regard to computing resources used in the computing-intensive AI analysis methods. Location-based and animal history-based detection refinements are incorporated into described systems. Further presented are methods for multi-modal and multi-factor detection of udder disease, as well as methods for infection type classification.
1. A method for automatically assembling sequences of images of udders of a set of dairy animals, with each animal identified and with each image having a timestamp, for use in detecting udder disease, comprising:
identifying an identity of each animal in the set of dairy animals, using one or more sensors, wherein each animal is an animal having an udder;
imaging all or a portion of the udder of each animal at a first time, creating one or more first images with an image capture time recorded as the timestamp;
assigning the one or more images and timestamps to each identified animal;
performing adjustments to individual pixel values of potential regions of interest of the one or more images within specified maxima and minima and according to a specified scaling function applicable to original pixel values, to increase image contrast for better detection of a diseased area;
identifying an identity of another animal, using the one or more sensors; and
sorting the images by animal and by capture time, for use in machine learning or change detection analysis against one or more reference images for the purpose of detecting udder disease.
2. The method as recited in claim 1 , wherein the identifying further includes scanning one or more tags on the animal.
3. The method as recited in claim 2 , wherein the one or more tags are RFID tags.
4. The method as recited in claim 1 , wherein the imaging of the udder at a first time and the imaging of the udder at a second time are performed using one or more thermal sensors.
5. The method as recited in claim 1 , wherein the one or more images are thermal images.
6. The method as recited in claim 5 , wherein the one or more sensors on or within the animal includes a GPS sensor.
7. The method as recited in claim 1 , wherein the analyzing further includes isolating the udder from within each image.
8. The method as recited in claim 7 , further comprising:
scanning the one or more sensors on or within the animal to determine any physical activity of the animal.
9. The method as recited in claim 8 , wherein the analyzing further includes analyzing any changes between the one or more first images and the one or more second images in view of the physical activity of the animal.
10. The method as recited in claim 1 , wherein the animal is a cow.
11. The method as recited in claim 1 , further comprising:
determining an activity level of the animal.
12. The method as recited in claim 11 wherein the activity level is determined by analyzing movement of the animal.
13. The method as recited in claim 12 , wherein the movement of the animal is determined using the one or more sensors.
14. The method as recited in claim 1 , further comprising:
identifying one or more images for further use in presenting or studying a time progression of udder disease within a given animal, wherein the one or more images are detected as containing possible diseased areas.
15. The method as recited in claim 1 , further comprising:
combining multiple partial image captures of a given animal udder through geometric transformations to form a complete or more nearly complete udder image for subsequent use in image analysis for disease detection.