EP4694675A1 - Device, method and system for pasture estimation and management - Google Patents

Device, method and system for pasture estimation and management

Info

Publication number
EP4694675A1
EP4694675A1 EP24789130.2A EP24789130A EP4694675A1 EP 4694675 A1 EP4694675 A1 EP 4694675A1 EP 24789130 A EP24789130 A EP 24789130A EP 4694675 A1 EP4694675 A1 EP 4694675A1
Authority
EP
European Patent Office
Prior art keywords
pasture
mass
data
grazing
geographical areas
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
EP24789130.2A
Other languages
German (de)
French (fr)
Inventor
Jeremy BRYANT
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Aimer Development Ltd
Original Assignee
Aimer Development Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Aimer Development Ltd filed Critical Aimer Development Ltd
Priority claimed from PCT/NZ2024/050040 external-priority patent/WO2024215212A1/en
Publication of EP4694675A1 publication Critical patent/EP4694675A1/en
Pending legal-status Critical Current

Links

Classifications

    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01K—ANIMAL HUSBANDRY; AVICULTURE; APICULTURE; PISCICULTURE; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
    • A01K29/00—Other apparatus for animal husbandry
    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01K—ANIMAL HUSBANDRY; AVICULTURE; APICULTURE; PISCICULTURE; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
    • A01K29/00—Other apparatus for animal husbandry
    • A01K29/005—Monitoring or measuring activity
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/24323—Tree-organised classifiers
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/01—Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/02—Agriculture; Fishing; Forestry; Mining
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/56—Extraction of image or video features relating to colour
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/13—Satellite images
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/188—Vegetation
    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01B—SOIL WORKING IN AGRICULTURE OR FORESTRY; PARTS, DETAILS, OR ACCESSORIES OF AGRICULTURAL MACHINES OR IMPLEMENTS, IN GENERAL
    • A01B79/00—Methods for working soil
    • A01B79/005—Precision agriculture
    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01C—PLANTING; SOWING; FERTILISING
    • A01C21/00—Methods of fertilising, sowing or planting
    • A01C21/007—Determining fertilization requirements
    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01K—ANIMAL HUSBANDRY; AVICULTURE; APICULTURE; PISCICULTURE; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
    • A01K2227/00—Animals characterised by species
    • A01K2227/10—Mammal
    • A01K2227/101—Bovine
    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01K—ANIMAL HUSBANDRY; AVICULTURE; APICULTURE; PISCICULTURE; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
    • A01K2227/00—Animals characterised by species
    • A01K2227/10—Mammal
    • A01K2227/103—Ovine
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10016—Video; Image sequence
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10024—Color image
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20076—Probabilistic image processing
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30181—Earth observation
    • G06T2207/30188—Vegetation; Agriculture

Definitions

  • the present technology relates to a device for the estimation of pasture mass and/or for pasture management.
  • the present technology also relates to a method for the estimation of pasture mass and/or for pasture management.
  • the present technology also relates to a system for the estimation of pasture mass and/or for pasture management.
  • a method of developing a pasture grazing allocation for a mob of animals includes obtaining pasture mass data for two or more target geographical areas of a farm, obtaining mob data for at least one group of grazing animals. Based on pasture mass data and mob data, determining a pasture grazing allocation of the group of animals in the two or more target geographical areas. The method may also include estimating pasture growth data for the two or more target geographical areas and based on pasture mass data, growth data and mob data, determining a pasture grazing allocation of the group of animals in the two or more target geographical areas. A pasture grazing allocation is communicated to a farmer or farm worker and includes at least one time and one or more target geographical area(s) for grazing the group of animals.
  • a method of developing a pasture grazing allocation for a group of animals including: obtaining pasture mass data for two or more target geographical areas, estimating pasture growth data for the two or more target geographical areas obtaining mob data for at least one group of grazing animals, and based on pasture growth data and mob data, determining a pasture grazing allocation of the group of animals in the two or more target geographical areas, wherein the pasture grazing allocation includes at least one time and one or more target geographical area(s) for grazing the group of animals.
  • the pasture mass data includes an estimate of pasture mass for the two or more target geographical areas. In some examples the pasture data includes an observation of the pasture mass for the two or more target geographical areas. In some examples the pasture mass data includes a partial observation of pasture mass and a pasture mass prediction based on the partial observation and historical pasture growth data.
  • the estimate of pasture mass includes obtaining a video of two or more target geographical areas and based on machine learning processing the video to making a prediction of pasture mass for the two or more target geographical areas.
  • the estimate of pasture mass includes obtaining pasture mass data using one or more of a plate-meter, C-Dax Pasture Meter, or human observation.
  • estimating pasture growth rates includes obtaining pasture mass data for the two or more target geographical areas at two or more distinct periods of time. In some examples estimating pasture growth rates includes a sum-of-squares regression.
  • the method includes establishing a target pasture mass for the two or more target geographical areas, and obtaining a deviationof the target pasture mass from the obtained pasture mass data for the two or more target geographical areas. In some examples the method includes displaying to a user, on a portable device screen, a graphical representation of the obtained pasture mass data for the two or more target geographical areas. In some examples the method includes displaying to a user, on a portable device screen, a graphical representation of the target pasture mass. In some examples the method includes displaying to a user, on a portable device screen, a graphical value representation of the obtained deviation. In some examples displaying to a user a graphical representation of the obtained pasture mass data for the two or more target geographical areas, includes organising the two or more target geographical areas according to obtained pasture mass.
  • the method includes establishing predicting a future pasture mass for the two or more target geographical areas, and obtaining a second deviation from the target pasture mass of the predicted future pasture mass data for the two or more target geographical areas.
  • displaying to a user a graphical representation of the pasture mass data for the two or more target geographical areas includes organising the two or more target geographical areas according to the predicted future pasture mass.
  • the method includes determining, or predicting, a surplus pasture mass.
  • a surplus pasture mass includes a positive deviation, or positive second deviation, from the target pasture mass of the pasture mass, or predicted future pasture mass, for one or more of the two or more target geographical areas.
  • the method includes determining, or predicting, a deficit pasture mass.
  • a deficit pasture mass includes a negative deviation, or negative second deviation, from the target pasture mass of the pasture mass, or predicted future pasture mass, for one or more of the two or more target geographical areas.
  • the method can estimate feed supplementation and suggest paddock fertilisation.
  • the method will estimate whether a farm will be moving into a surplus or deficit of pasture mass, and suggest what paddocks to drop out of rotation, or how much additional supplement may be required.
  • the method utilizes an Artificial Intelligence (Al) or a Machine Learning (ML) process to learn pasture growth patterns for one or more of the two or more target geographical areas and give recommendations on pasture mass and/or for pasture management.
  • the method utilises smartphone camera optics and an Al or ML process to scan paddocks and obtain an immediate assessment of pasture mass.
  • the Al or ML process is trained using historical pasture mass data and/or historical pasture growth data for a geographical area. In some examples the method will monitor accumulated growth rates for individual paddocks over the course of the farming season. In some examples the Al or ML process is trained using pasture growth rates for individual paddocks over the course of a farming season. In some examples an Al or ML process is trained using a dataset of over 10,000 historical images with matching pasture mass estimates derived from a rising plate meter. In some examples the images are processed and Red: Green: Blue (RGB) statistics derived from each image. In some examples two machine learning approaches (XGBoost and Random Forest) are applied to develop models that predict pasture mass accurately from new processed images.
  • XGBoost and Random Forest two machine learning approaches
  • the method includes a digital version of the farm and paddock maps. In some examples the method includes a digital version of the farm and paddock maps in GeoJson format.
  • the method includes allowing a farmer to complete a 'part farm measurement'. For example, measure 10 of the 40 paddocks and get an estimate/derive the pasture mass for the rest. In some examples the method includes estimating likely pasture mass on the measurement day using derived and farmer entered grazing plans, pasture growth rates and learned growth information about each of the paddocks.
  • paddock growth rates and farm average pasture mass are calculated on previous and current pasture measures and days between measures.
  • a sum of squares regression-based approach is used to calculate relative growth rate solutions for each paddock based on historical growth rates. In some examples this result is used for future pasture growth rate and pasture mass predictions.
  • calculated relative and average growth rate solutions are used to calculate a total pasture production per paddock which is then presented in a screen visualisation.
  • a user say, a farmer
  • inputs an expectation of pasture growth rate for the next period of days say 14 days.
  • the system will use the latest pasture growth rate estimated.
  • the system will use a proprietary pasture growth forecaster/emulator.
  • a proprietary paddock and supplement allocation method is used to identify for each half day and mob (group or herd of animals) a paddock to graze for how long and when to move a mob to a new paddock and which paddock.
  • paddocks are grazed or grown and tracked to determine the next day or grazing pasture mass.
  • the method works for multiple mobs on a farm and ensures multiple mobs are not grazing the same paddock on the same day.
  • an animal supplement amount is suggested/allocated for a time period to reduce variation in mob total feed intake.
  • a proprietary optimisation method is used to find the amount of supplement to feed based on the degree of one or more of a pasture deficit.
  • a proprietary optimisation method is used to find the optimum number of paddocks to apply nitrogen to and the amount of nitrogen needed that will correct a forecast feed wedge and be profitable based on one or more of the degree of the deficit, milk price, amount of nitrogen already applied in the season and the cost of nitrogen, and the suggestion then delivered to the farmer via an App, txt, WhatsApp.
  • the pasture management plan includes one or more of paddocks to conserve, amount of supplements to feed, nitrogen to apply, paddocks to graze when and by which mob and projected average feed mass.
  • a user typically although not exclusively farmer, uses the technology by setting up details of a farm, area of paddock(s) and paddocks that can be grazed together, uploading or providing a map of the farm, entering pasture mass for paddocks or distinct areas of the farm.
  • the technology then uses one or more of proprietary algorithms, proprietary information, artificial intelligence (Al) and/or machine learning (ML) to provide predictions that assist the user to optimise pasture management functions.
  • results and/or historical data are used my Al and/or ML processes to continuously learn about the farm to improve future predictions.
  • a device including a processor, a processor readable memory, an input interface, and a display, wherein the processor readable memory includes instructions that when operated by the processor cause the device to perform examples of the foregoing method of the technology.
  • a system including a device having a processor, a processor readable memory, an input interface, and a display, wherein the processor readable memory includes instructions that when operated by the processor cause the device to perform examples of the foregoing method of the technology.
  • FIG. 1 schematically illustrates one example method of the invention
  • Figure 2 illustrates a mapping showing a plurality of geographical areas (e.g., paddocks) making up a farm on which the invention may be used,
  • geographical areas e.g., paddocks
  • FIG. 3 graphically illustrates an example distribution (wedge) of pastures masss for the example geographical areas of Figure 2,
  • FIG. 4 schematically illustrates another example method of the invention
  • Figure 5 schematically illustrates one example device/method for estimating or predicting pastures mass of a geographical area (paddock)
  • Figure 6 illustrates a user of a mobile device operating a method of one example of the invention obtaining images of pasture mass from which an estimation or prediction can be made
  • a geographical area paddock
  • Figure 6 illustrates a user of a mobile device operating a method of one example of the invention obtaining images of pasture mass from which an estimation or prediction can be made
  • FIG. 7 schematically illustrates yet another example method of the invention
  • farm When used herein “farm”, includes an area of land used for growing crops and rearing animals. In some examples the area of land is divided or categorised into two or more sub-areas such as paddocks or fields.
  • a first step 100 of the method includes creating a digital map of a farm including one or more geographical areas (say paddocks) of the farm.
  • Figure 2 shows a typical map having a plurality of geographical areas a, b, c, d, e, f, g, h, I, j, k, I, m, n, o, and p (say, paddocks a to p).
  • the digital version of the farm and paddock maps is drawing in a GeoJson format.
  • a proprietary solution is employed whereby artificial intelligence (Al) creates a mapping file by analysing an image of the farm (for example a google satellite image).
  • the digital map is manually drawn.
  • the next step 110 in one example of the technology is to obtaining pasture mass data for paddocks a to p.
  • the pasture mass data includes an estimate of pasture mass for the paddocks. In some examples this is done by a human observation of the pasture mass for the paddocks, or by using one or more of a plate-meter or C-Dax Pasture Meter.
  • the estimate of pasture mass can be obtained from a video of the paddocks and based on Al or machine learning (ML) processing of the video to making a prediction of pasture mass for the two or more target geographical areas.
  • the method utilises smartphone camera optics and an Al to scan paddocks and get an immediate assessment of pasture mass.
  • a user say a farmer or farm worker, uses a mobile video recording device to obtain a circular video of the pasture.
  • it is a 240-to-300-degree circle videos of the pasture.
  • the video is processed 510/610 by splitting the video into individual images of portions of the pasture.
  • the images are cropped or trimmed to remove sky, horizon, or other non-pasture features, so that the images include only pasture.
  • a timestamp and GPS data is also extracted for each image: 520/620.
  • timestamp data is used to match pastures images against one or more areas of the farm, say paddocks a to p.
  • the video is obtained by a drone or camera mounted on an animal (cow as discussed later) or farm bike.
  • machine learning (ML) approaches for example XGBoost and Random Forest
  • ML machine learning
  • the Al models are trained using a dataset of spectral RGB data from over 10,000 individual images with matching pasture mass estimates derived from a rising plate meter or C-Dax Pasture Meter.
  • the images are processed and Red: Green: Blue (RGB) statistics derived from each image: 540/640.
  • obtaining pasture mass data includes obtaining an observation of the pasture mass, 650, or only some paddocks, and using a pasture mass prediction based on the partial observation and historical pasture growth data to fill in data for unobserved paddocks.
  • this partial observation method includes having a farmer complete a 'part farm measurement' of say a random selection of paddocks a to p and getting an estimate/deriving the rest using the Al, ML method.
  • this some examples include a method to estimate likely pasture mass on the measurement day using derived and farmer entered grazing plans, pasture growth rates and learned growth information about each of the paddocks. A farmer can then choose to use this estimate or measure mass of each individual paddock.
  • paddock growth rates and farm average pasture mass are calculated on previous and current pasture measures and days between measures.
  • pasture mass is the amount of pasture per hectare and is measured in kilograms of dry matter per hectare (kgDM/ha).
  • a next step 130 includes estimating pasture growth data for the two or more target geographical areas.
  • estimating pasture growth data includes estimating pasture growth rates and mass pasture.
  • estimating pasture growth rates includes obtaining pasture mass data for the two or more target geographical areas at two or more distinct periods of time.
  • estimating pasture growth rates includes a sum-of-squares regression.
  • the device, method and or system will monitor accumulated growth rates for individual paddocks over the course of the farming season, giving granular detail on each's paddock's respective growing ability.
  • One step 120 of some examples is to obtaining mob data for at least one group of grazing animals.
  • the animals are cows.
  • the animals are sheep.
  • the animals are other domesticated livestock.
  • the mob data may include a number and size of animals and/or a grazing requirement, or production data.
  • the animals are represented as a model such as described in inventor's other joint works 'Modeling Dairy Cow Variations in Genotype to Determine the Expected Ranges of Response to Supplements' (a copy of which is available at https://www.researchgate.net/publication/267222517 Modeling Dairy Cow Variations in Genoty pe to Determine the Expected Ranges of Response to Supplements), and 'Development and evaluation of a search simulation model that predicts dairy cattle performance based on animal genotype and environmental sensitivity information' (Agricultural Systems Volume 97, Issues 1-2, April 2008, Pages 13-25). The entire contents of both publications are incorporated herein by reference in their entirety.
  • Another step 140 includes determining a pasture grazing allocation of the group of animals in the two or more target geographical areas based on pasture growth data and mob data.
  • the pasture grazing allocation includes at least one time and one target geographical area for grazing the group of animals.
  • the method will calculate based on cow numbers, pasture mass and growth rates a pasture management and mob (animal) paddock rotation/feeding plan.
  • the method will utilise Al or ML to learn pasture growth patterns and give recommendations on pasture mass and/or for pasture management.
  • the method includes establishing a predicted future pasture mass 150 for the paddocks a-p, and based on predicted future pasture mass 150 modify the pasture management and mob (animal) paddock rotation/feeding plan.
  • the method establishes a target pasture mass 320 for the paddocks a-p, and calculates a deviation from the target pasture mass of the obtained pasture mass data for each paddock.
  • the pasture mass for each paddock is displayed to a user, on a portable device screen, in a graphical representation of the obtained pasture mass data for the paddocks.
  • displaying to a user a graphical representation of the obtained pasture mass data for the paddocks is ordered according to obtained pasture mass.
  • the target pasture mass is also displayed.
  • the obtained deviation is also displayed and represented as an optimisation score (range of 0 to 100%).
  • the method includes determining, or predicting, a surplus pasture mass.
  • a surplus pasture mass includes a positive deviation, or positive second deviation, from the target pasture mass of the pasture mass, or predicted future pasture mass, for one or more of the two or more target geographical areas.
  • the method includes determining, or predicting, a deficit pasture mass.
  • a deficit pasture mass includes a negative deviation, or negative second deviation, from the target pasture mass of the pasture mass, or predicted future pasture mass, for one or more of the two or more target geographical areas.
  • the system can estimate feed supplementation and suggest paddock fertilisation.
  • the method will estimate whether a farm will be moving into a surplus or deficit of pasture mass, and suggest what paddocks to drop out of rotation, or how much additional supplement may be required.
  • the method utilizes an Artificial Intelligence (Al) or a Machine Learning (ML) process to learn pasture growth patterns and give recommendations on pasture mass and/or for pasture management.
  • Al Artificial Intelligence
  • ML Machine Learning
  • the method utilises smartphone camera optics and an Al or ML process to scan paddocks and obtain an immediate assessment of pasture mass
  • FIG 4 schematically illustrates another example method of the invention.
  • a plurality of paddocks say paddocks a-p, including all or part of a farm are mapped as previously discussed with reference to step 100 of Figure 1 and Figure 2.
  • step 402 pasture mass data and mob data are obtained as previously discussed with reference to 110 and 120 of Figure 1.
  • the farmer also updates mob details, rotation lengths and pasture and supplement intake targets.
  • the method allows the farmer to complete a 'part farm measurement'. For example, measure, say, 10 of, say, 40 paddocks and use an Al or ML process to estimate/derive the rest.
  • estimates are based on derived and farmer entered grazing plans, pasture growth rates and learned growth information (see step 405) about each of the paddocks.
  • a "pasture wedge analysis" runs at step 403.
  • the method ranks paddocks from most mass to least as a bar graph (see Figure 3) and splits the ranked paddocks into three subsets: longest/high mass 330, middle/medium mass 340 and shortest/low mass 350 paddocks.
  • the longest subset includes, say, the 20% of paddocks with the highest pasture mass.
  • the shortest subset includes, say, the 20% of paddocks with the lowest pasture mass.
  • the middle subset includes, say, the remaining 60% of paddocks with a pasture mass between longest and shortest subsets.
  • the method calculates, for each of subset, the deviation of each paddock from a target line 320 and calculates the sum squares of error for each portion of the wedge.
  • a user can be presented with warnings based on the average deviation of the subset. For example, if there is a positive deviation about level 320 in a section of the pasture wedge it correlates as a 'current or emerging surplus' or cows are under-grazing/level being too much.
  • the method also calculate an optimisation level for the current wedge i.e., 100% optimization equals all paddocks lining up with the target line. This can be communicated to a user. This helps a farmer increase the optimisation by seeking scenarios that maximise the optimisation level.
  • a farmer may enter or generates pasture mass estimates on distinct days.
  • paddock growth rates and farm average pasture mass are calculated based on previous and current pasture measures and days between measures.
  • step 405 a squares regression-based approach is used to calculate relative growth rate solutions for each paddock a-p based on historical growth rates calculated in step 404. In some examples this result is stored and used for future pasture growth rate predictions. For example, a particular paddock might be growing at 120% of the average growth rate of other paddocks on the farm. This relative rate is then used as part of an optomisation to grow the paddock at a faster or slower rate in the future. In some examples the calculated relative and average growth rate solutions are also used to calculate a total pasture production per paddock which is then presented to the farmer in a visualisation. In some examples at step 406 a farmer inputs expected pasture growth rates for the next 14 days based on experience. In some examples the system will use the latest pasture growth rate estimated. In some examples an Al or ML method is used to forecaster/predict pasture growth rates.
  • a baseline proprietary paddock and supplement allocation method identifies for each within daytime period and each mob a paddock to graze and for how long and when to move the mob to a new paddock and to which paddock.
  • Each day the paddocks are tracked, grazed, and grown to determine the next day or grazing pasture mass based on data from steps 405 and 406.
  • the method is used for multiple mobs on the farm and ensures multiple mobs are not grazing the same paddock on the same day.
  • paddocks that don't quite have enough pasture for a full graze (or vice versa) supplement feed is indicated as required for that day to ensure even feed supply.
  • a paddock is identified as AM grazing only or PM grazing only.
  • mobs may be excluded/included from paddocks, differing ratios of pasture and supplement intake at AM and PM allocations, fixed and min-max supplement allocation to determine paddock and supplement allocations by mob.
  • a "forecast wedge analysis" is run based on the daily forecast pasture masss calculated in step 407.
  • an optimisation method at 409 identifies the optimum number of paddocks to conserve (i.e., leave out of the grazing rotation), and the suggestion is then delivered to the farmer via the App.
  • an optimisation method at 410 identifies an optimum number of paddocks to apply nitrogen to and the amount of nitrogen needed that will correct the forecast feed wedge and be profitable based on the degree of the deficit, milk price, amount of nitrogen already applied in the season and the cost of nitrogen, and the suggestion then delivered to the farmer via the App, txt, WhatsApp.
  • an optimisation method at 411 identifies the amount of supplement to feed based on the degree of the pasture deficit, and the suggestion then delivered to the farmer or farm worker via the App, text message, WhatsApp, email, or mobile device notification.
  • the relevant decisions e.g., paddocks to conserve, amount of supplements to feed, nitrogen to apply, paddocks to graze when and by which mob and projected average feed mass are then communicated to relevant farm staff, such as a farmer or farm worker, via App, SMS, text message, WhatsApp, email or mobile device notification.
  • a proprietary paddock and supplement allocation method is used to identify for each within day time period (e.g., AM and PM) and mob (group or herd of animals), a paddock(s) to graze for how long and when to move a mob to a new paddock(s) and which paddock(s).
  • day paddocks are tracked, grazed, and grown to determine the next day or grazing pasture mass.
  • the method works for multiple mobs and paddock(s) on a farm and ensures multiple mobs are not grazing the same paddock(s) on the same day.
  • each mob is given a colour or indication icon or symbol. Colours or indication icons or symbols can be overlaid on a visual map of the paddocks, such as the visualisation on Figure 2, as a visual indication to a farmer of where and when a mob should be grazing a particular paddock.
  • the pasture management plan includes one or more of paddocks to conserve, amount of supplements to feed, nitrogen to apply, paddocks to graze when and by which mob and projected average feed mass. In some examples this is communicated to relevant user and/or farm staff via App, or mobile device notification.
  • a user typically although not exclusively farmer or farm worker, uses the technology by setting up details of a farm, uploading or providing a map of the farm, entering pasture mass for paddocks or distinct areas of the farm.
  • the technology then uses one or more of proprietary algorithms, proprietary information, artificial intelligence (Al) and/or machine learning (ML) to provide predictions that assist the user to optimise pasture management functions.
  • Al artificial intelligence
  • ML machine learning
  • results and/or historical data are used my Al and/or ML processes to continuously learn about the farm to improve future predictions.
  • the technology includes, but are not limited to, one of more of the following:
  • Some examples of the technology operate over a single farm. Other examples of the technology operate over multiple farms.
  • Examples of the technology generate detailed and informed paddock and supplement allocation plans for each mob on the farm that optimise pasture usage and continues to learn the uniqueness of each farm and paddock. All plans have the flexibility of being fine-tuned and re-generated to suit each farmers individual requirements.
  • Examples of the technology are designed specifically as a digital assistant and coach it can be used directly by farm staff irrespective of experience rather than needing highly trained individuals to use and operate the software.
  • individual images with matching pasture mass estimates derived from a rising plate meter are used as training data form an Al or ML process to estimate pasture mass from images. Images are processed and Red:Green:Blue (RGB) statistics derived from each image. Using this data, two machine learning approaches (XGBoost and Random Forest) can be used to develop models that predict pasture mass accurately from new processed images.
  • XGBoost and Random Forest two machine learning approaches
  • images are obtained by using a mobile device to take video in each paddock.
  • the method processes the images and uses ML models to estimate pasture mass for each paddock.
  • a GPS stamp on the video is used to match the video to a particular paddock a to p rather than the farmer having to choose the paddock.
  • an imaging device may be provided on selected animals in a mob, say mounted on a collar or bridle. At certain times of the day i.e., when it is known that the cows (mobs) are about to enter paddocks or will be in paddocks, video footage of the pasture will be captured and the ML models will be used to estimate pre-grazing and post-grazing pasture mass, and grazing events for each paddock.
  • an imaging device may be provided on a farm/work bike. As a farmer goes in and out of paddocks or goes up and down a race, video will be captured, and the technology would process the images and use the ML models to estimate pasture mass for each paddock which is processed and calculated in the mobile device or in the cloud.
  • Figure 7 schematically illustrates yet another example method of the invention.
  • a 240-to-300-degree circular videos of pasture is captured, say via a smartphone.
  • the video is split into individual images and the images are trimmed to remove the sky and horizon, so they include only pasture: 610/620.
  • Machine vision models were training using XG Boost and Random Forest methods with matching video spectral RGB data and pasture mass measures from a platemeter, or C-Dax Pasture Meter. Feature engineering is used to identify the most informative spectral statistics for seasons and regions: 630/640. In some examples the trained machine vision models are used to estimate pasture mass for each new video and images. In other examples pasture masses are estimated manually using devices such as platemeters, C-Dax Pasture Meters or an experienced eye: 650.
  • paddock growth rates and farm average pasture mass are calculated.
  • trained Al and or ML method is used to estimate current paddock growth rates and farm average pasture mass based on historical training/trained data.
  • a grazing / pasture management plan is determined and communicated to a user: 680.
  • the plan is part based on a pasture mass forecast where all paddocks are grown up based on the user defined forecast pasture growth rate and/or previously calculated growth rate solutions. This relative rate is used to grow paddocks at a faster or slower rates depending on the prior learned knowledge.
  • the plan is part based on split AM and PM or AM/PM grazing days for paddock or combinations of paddocks that best matches the desired pasture intake and characteristics of the mob. Considerations include whether paddocks are only grazeable in the AM or PM, or corresponding buddy paddocks that are always grazed together because they share adjacent gates or could be grazed together.
  • the plan is part based on a pasture mass needed to meet total animal mob pasture demand. In some examples the plan is part based on supplement needs to fill pasture gaps for each mob.
  • a production output forecast can be estimated and communicated to a user: 690.
  • the technology may also be said broadly to consist in the parts, elements and features referred to or indicated in the specification of the application, individually or collectively, in any or all combinations of two or more of said parts, elements or features.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Business, Economics & Management (AREA)
  • Environmental Sciences (AREA)
  • Animal Husbandry (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Multimedia (AREA)
  • Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Human Resources & Organizations (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Marketing (AREA)
  • Evolutionary Computation (AREA)
  • Biodiversity & Conservation Biology (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Software Systems (AREA)
  • Agronomy & Crop Science (AREA)
  • Mathematical Physics (AREA)
  • Marine Sciences & Fisheries (AREA)
  • Computing Systems (AREA)
  • Mining & Mineral Resources (AREA)
  • Primary Health Care (AREA)
  • Development Economics (AREA)
  • Medical Informatics (AREA)
  • Game Theory and Decision Science (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Computational Linguistics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)

Abstract

A method of developing a pasture grazing allocation for a mob of animals includes obtaining pasture mass data for two or more target geographical areas of a farm, obtaining mob data for at least one group of grazing animals. Based on pasture mass data and mob data, determining a pasture grazing allocation of the group of animals in the two or more target geographical areas. The method may also include estimating pasture growth data for the two or more target geographical areas and based on pasture mass data, growth data and mob data, determining a pasture grazing allocation of the group of animals in the two or more target geographical areas. A pasture grazing allocation is communicated to a farmer or farm worker and includes at least one time and one or more target geographical area(s) for grazing the group of animals.

Description

Device, Method and System for Pasture Estimation and Management
Field of the Invention
The present technology relates to a device for the estimation of pasture mass and/or for pasture management. The present technology also relates to a method for the estimation of pasture mass and/or for pasture management. The present technology also relates to a system for the estimation of pasture mass and/or for pasture management.
Background
Arguably, for many farmers the farm herd, or mob of animals, along with the pasture on which they graze are the most important farm assets. Pasture is highly dependent on geography, weather conditions and good pasture management. Farmers generally cannot control geography and weather and good pasture management is a key factor in farm success.
Object of the Invention
It is an object of the technology to provide a device for the estimation of pasture mass and/or for pasture management. It is another object of the technology to provide a method for the estimation of pasture mass and/or for pasture management. It is yet another object of the technology to provide a system for the estimation of pasture mass and/or for pasture management.
Summary of the Technology
According to at least one example of the technology there is provide a method of developing a pasture grazing allocation for a mob of animals includes obtaining pasture mass data for two or more target geographical areas of a farm, obtaining mob data for at least one group of grazing animals. Based on pasture mass data and mob data, determining a pasture grazing allocation of the group of animals in the two or more target geographical areas. The method may also include estimating pasture growth data for the two or more target geographical areas and based on pasture mass data, growth data and mob data, determining a pasture grazing allocation of the group of animals in the two or more target geographical areas. A pasture grazing allocation is communicated to a farmer or farm worker and includes at least one time and one or more target geographical area(s) for grazing the group of animals.
According to another example of the technology there is provide a method of developing a pasture grazing allocation for a group of animals, the method including: obtaining pasture mass data for two or more target geographical areas, estimating pasture growth data for the two or more target geographical areas obtaining mob data for at least one group of grazing animals, and based on pasture growth data and mob data, determining a pasture grazing allocation of the group of animals in the two or more target geographical areas, wherein the pasture grazing allocation includes at least one time and one or more target geographical area(s) for grazing the group of animals.
In some examples the pasture mass data includes an estimate of pasture mass for the two or more target geographical areas. In some examples the pasture data includes an observation of the pasture mass for the two or more target geographical areas. In some examples the pasture mass data includes a partial observation of pasture mass and a pasture mass prediction based on the partial observation and historical pasture growth data.
In some examples the estimate of pasture mass includes obtaining a video of two or more target geographical areas and based on machine learning processing the video to making a prediction of pasture mass for the two or more target geographical areas.
In some examples the estimate of pasture mass includes obtaining pasture mass data using one or more of a plate-meter, C-Dax Pasture Meter, or human observation.
In some examples estimating pasture growth rates includes obtaining pasture mass data for the two or more target geographical areas at two or more distinct periods of time. In some examples estimating pasture growth rates includes a sum-of-squares regression.
In some examples the method includes establishing a target pasture mass for the two or more target geographical areas, and obtaining a deviationof the target pasture mass from the obtained pasture mass data for the two or more target geographical areas. In some examples the method includes displaying to a user, on a portable device screen, a graphical representation of the obtained pasture mass data for the two or more target geographical areas. In some examples the method includes displaying to a user, on a portable device screen, a graphical representation of the target pasture mass. In some examples the method includes displaying to a user, on a portable device screen, a graphical value representation of the obtained deviation. In some examples displaying to a user a graphical representation of the obtained pasture mass data for the two or more target geographical areas, includes organising the two or more target geographical areas according to obtained pasture mass.
In some examples the method includes establishing predicting a future pasture mass for the two or more target geographical areas, and obtaining a second deviation from the target pasture mass of the predicted future pasture mass data for the two or more target geographical areas. In some examples displaying to a user a graphical representation of the pasture mass data for the two or more target geographical areas, includes organising the two or more target geographical areas according to the predicted future pasture mass.
In some examples the method includes determining, or predicting, a surplus pasture mass. In some examples a surplus pasture mass includes a positive deviation, or positive second deviation, from the target pasture mass of the pasture mass, or predicted future pasture mass, for one or more of the two or more target geographical areas.
In some examples the method includes determining, or predicting, a deficit pasture mass. In some examples a deficit pasture mass includes a negative deviation, or negative second deviation, from the target pasture mass of the pasture mass, or predicted future pasture mass, for one or more of the two or more target geographical areas.
In yet other examples of the technology there is provided a method for calculating, based on a group of grazing animals, pasture mass and growth rates, a pasture management and mob (animal) paddock rotation/feeding plan.
In some examples the method can estimate feed supplementation and suggest paddock fertilisation.
In some examples the method will estimate whether a farm will be moving into a surplus or deficit of pasture mass, and suggest what paddocks to drop out of rotation, or how much additional supplement may be required. In some examples the method utilizes an Artificial Intelligence (Al) or a Machine Learning (ML) process to learn pasture growth patterns for one or more of the two or more target geographical areas and give recommendations on pasture mass and/or for pasture management. In some examples the method utilises smartphone camera optics and an Al or ML process to scan paddocks and obtain an immediate assessment of pasture mass.
In some examples the Al or ML process is trained using historical pasture mass data and/or historical pasture growth data for a geographical area. In some examples the method will monitor accumulated growth rates for individual paddocks over the course of the farming season. In some examples the Al or ML process is trained using pasture growth rates for individual paddocks over the course of a farming season. In some examples an Al or ML process is trained using a dataset of over 10,000 historical images with matching pasture mass estimates derived from a rising plate meter. In some examples the images are processed and Red: Green: Blue (RGB) statistics derived from each image. In some examples two machine learning approaches (XGBoost and Random Forest) are applied to develop models that predict pasture mass accurately from new processed images.
In some examples the method includes a digital version of the farm and paddock maps. In some examples the method includes a digital version of the farm and paddock maps in GeoJson format.
In some examples the method includes allowing a farmer to complete a 'part farm measurement'. For example, measure 10 of the 40 paddocks and get an estimate/derive the pasture mass for the rest. In some examples the method includes estimating likely pasture mass on the measurement day using derived and farmer entered grazing plans, pasture growth rates and learned growth information about each of the paddocks.
In some examples paddock growth rates and farm average pasture mass are calculated on previous and current pasture measures and days between measures.
In some examples a sum of squares regression-based approach is used to calculate relative growth rate solutions for each paddock based on historical growth rates. In some examples this result is used for future pasture growth rate and pasture mass predictions.
In some examples calculated relative and average growth rate solutions are used to calculate a total pasture production per paddock which is then presented in a screen visualisation. In some examples a user (say, a farmer) inputs an expectation of pasture growth rate for the next period of days (say 14 days). In some examples the system will use the latest pasture growth rate estimated. In some examples the system will use a proprietary pasture growth forecaster/emulator.
In some examples a proprietary paddock and supplement allocation method is used to identify for each half day and mob (group or herd of animals) a paddock to graze for how long and when to move a mob to a new paddock and which paddock. In some examples, paddocks are grazed or grown and tracked to determine the next day or grazing pasture mass. In some examples the method works for multiple mobs on a farm and ensures multiple mobs are not grazing the same paddock on the same day.
In some examples an animal supplement amount is suggested/allocated for a time period to reduce variation in mob total feed intake. In some examples a proprietary optimisation method is used to find the amount of supplement to feed based on the degree of one or more of a pasture deficit.
In some examples a proprietary optimisation method is used to find the optimum number of paddocks to apply nitrogen to and the amount of nitrogen needed that will correct a forecast feed wedge and be profitable based on one or more of the degree of the deficit, milk price, amount of nitrogen already applied in the season and the cost of nitrogen, and the suggestion then delivered to the farmer via an App, txt, WhatsApp.
In some examples the pasture management plan includes one or more of paddocks to conserve, amount of supplements to feed, nitrogen to apply, paddocks to graze when and by which mob and projected average feed mass.
In some examples of the technology a user, typically although not exclusively farmer, uses the technology by setting up details of a farm, area of paddock(s) and paddocks that can be grazed together, uploading or providing a map of the farm, entering pasture mass for paddocks or distinct areas of the farm. The technology then uses one or more of proprietary algorithms, proprietary information, artificial intelligence (Al) and/or machine learning (ML) to provide predictions that assist the user to optimise pasture management functions. In some examples, results and/or historical data are used my Al and/or ML processes to continuously learn about the farm to improve future predictions. According to some examples of the technology there is provide a device including a processor, a processor readable memory, an input interface, and a display, wherein the processor readable memory includes instructions that when operated by the processor cause the device to perform examples of the foregoing method of the technology.
According to some examples of the technology there is provide a system, the system including a device having a processor, a processor readable memory, an input interface, and a display, wherein the processor readable memory includes instructions that when operated by the processor cause the device to perform examples of the foregoing method of the technology.
Further examples of the technology, which should be considered in all its novel aspects, will become apparent to those skilled in the art upon reading of the following description which provides at least one example of a practical application of the technology.
Brief description of the drawings
One or more embodiments of the technology will be described below by way of example only, and without intending to be limiting, with reference to the following drawings, in which:
Figure 1 schematically illustrates one example method of the invention,
Figure 2 illustrates a mapping showing a plurality of geographical areas (e.g., paddocks) making up a farm on which the invention may be used,
Figure 3 graphically illustrates an example distribution (wedge) of pastures masss for the example geographical areas of Figure 2,
Figure 4 schematically illustrates another example method of the invention,
Figure 5 schematically illustrates one example device/method for estimating or predicting pastures mass of a geographical area (paddock), Figure 6 illustrates a user of a mobile device operating a method of one example of the invention obtaining images of pasture mass from which an estimation or prediction can be made, and
Figure 7 schematically illustrates yet another example method of the invention,
Description of examples of the technology
When used herein "farm", includes an area of land used for growing crops and rearing animals. In some examples the area of land is divided or categorised into two or more sub-areas such as paddocks or fields.
Referring to the Figures there is schematically illustrated several broad examples of a method of the technology for determining a pasture management and mob (animal) paddock rotation/feeding plan based on animal numbers and condition, pasture mass and historical growth rates.
A first step 100 of the method includes creating a digital map of a farm including one or more geographical areas (say paddocks) of the farm. Figure 2 shows a typical map having a plurality of geographical areas a, b, c, d, e, f, g, h, I, j, k, I, m, n, o, and p (say, paddocks a to p). In some examples of the invention the digital version of the farm and paddock maps is drawing in a GeoJson format. In other examples a proprietary solution is employed whereby artificial intelligence (Al) creates a mapping file by analysing an image of the farm (for example a google satellite image). In some examples the digital map is manually drawn.
The next step 110 in one example of the technology is to obtaining pasture mass data for paddocks a to p. In some examples the pasture mass data includes an estimate of pasture mass for the paddocks. In some examples this is done by a human observation of the pasture mass for the paddocks, or by using one or more of a plate-meter or C-Dax Pasture Meter. In other examples the estimate of pasture mass can be obtained from a video of the paddocks and based on Al or machine learning (ML) processing of the video to making a prediction of pasture mass for the two or more target geographical areas.
In some examples the method utilises smartphone camera optics and an Al to scan paddocks and get an immediate assessment of pasture mass. In some examples, say illustrated in Figures 5 and 6, a user, say a farmer or farm worker, uses a mobile video recording device to obtain a circular video of the pasture. In some examples it is a 240-to-300-degree circle videos of the pasture. In some examples the video is processed 510/610 by splitting the video into individual images of portions of the pasture. In some examples the images are cropped or trimmed to remove sky, horizon, or other non-pasture features, so that the images include only pasture. In some examples a timestamp and GPS data is also extracted for each image: 520/620. In some examples timestamp data is used to match pastures images against one or more areas of the farm, say paddocks a to p. In some examples the video is obtained by a drone or camera mounted on an animal (cow as discussed later) or farm bike.
In some examples machine learning (ML) approaches (for example XGBoost and Random Forest) are applied to train models that predict pasture mass accurately from the pasture images: 530/630. The Al models are trained using a dataset of spectral RGB data from over 10,000 individual images with matching pasture mass estimates derived from a rising plate meter or C-Dax Pasture Meter. In some examples the images are processed and Red: Green: Blue (RGB) statistics derived from each image: 540/640.
In some examples obtaining pasture mass data includes obtaining an observation of the pasture mass, 650, or only some paddocks, and using a pasture mass prediction based on the partial observation and historical pasture growth data to fill in data for unobserved paddocks. In some examples this partial observation method includes having a farmer complete a 'part farm measurement' of say a random selection of paddocks a to p and getting an estimate/deriving the rest using the Al, ML method. To do this some examples include a method to estimate likely pasture mass on the measurement day using derived and farmer entered grazing plans, pasture growth rates and learned growth information about each of the paddocks. A farmer can then choose to use this estimate or measure mass of each individual paddock. In some examples paddock growth rates and farm average pasture mass are calculated on previous and current pasture measures and days between measures.
In some examples pasture mass is the amount of pasture per hectare and is measured in kilograms of dry matter per hectare (kgDM/ha).
In some examples a next step 130 includes estimating pasture growth data for the two or more target geographical areas. In some examples estimating pasture growth data includes estimating pasture growth rates and mass pasture. In some examples estimating pasture growth rates includes obtaining pasture mass data for the two or more target geographical areas at two or more distinct periods of time. In some examples estimating pasture growth rates includes a sum-of-squares regression. In some examples the device, method and or system will monitor accumulated growth rates for individual paddocks over the course of the farming season, giving granular detail on each's paddock's respective growing ability.
One step 120 of some examples is to obtaining mob data for at least one group of grazing animals. In some examples there is more than one mob of animals. In some examples the animals are cows. In other examples the animals are sheep. In yet other examples the animals are other domesticated livestock. The mob data may include a number and size of animals and/or a grazing requirement, or production data. In some examples the animals are represented as a model such as described in inventor's other joint works 'Modeling Dairy Cow Variations in Genotype to Determine the Expected Ranges of Response to Supplements' (a copy of which is available at https://www.researchgate.net/publication/267222517 Modeling Dairy Cow Variations in Genoty pe to Determine the Expected Ranges of Response to Supplements), and 'Development and evaluation of a pastoral simulation model that predicts dairy cattle performance based on animal genotype and environmental sensitivity information' (Agricultural Systems Volume 97, Issues 1-2, April 2008, Pages 13-25). The entire contents of both publications are incorporated herein by reference in their entirety.
Another step 140 includes determining a pasture grazing allocation of the group of animals in the two or more target geographical areas based on pasture growth data and mob data. In some examples the pasture grazing allocation includes at least one time and one target geographical area for grazing the group of animals. In some examples the method will calculate based on cow numbers, pasture mass and growth rates a pasture management and mob (animal) paddock rotation/feeding plan.
In some examples the method will utilise Al or ML to learn pasture growth patterns and give recommendations on pasture mass and/or for pasture management. In some examples the method includes establishing a predicted future pasture mass 150 for the paddocks a-p, and based on predicted future pasture mass 150 modify the pasture management and mob (animal) paddock rotation/feeding plan.
Referring to Figure 3, in some examples the method establishes a target pasture mass 320 for the paddocks a-p, and calculates a deviation from the target pasture mass of the obtained pasture mass data for each paddock. In some examples the pasture mass for each paddock is displayed to a user, on a portable device screen, in a graphical representation of the obtained pasture mass data for the paddocks. In some examples displaying to a user a graphical representation of the obtained pasture mass data for the paddocks is ordered according to obtained pasture mass. In some examples the target pasture mass is also displayed. In some examples the obtained deviation is also displayed and represented as an optimisation score (range of 0 to 100%).
In some examples the method includes determining, or predicting, a surplus pasture mass. In some examples a surplus pasture mass includes a positive deviation, or positive second deviation, from the target pasture mass of the pasture mass, or predicted future pasture mass, for one or more of the two or more target geographical areas.
In some examples the method includes determining, or predicting, a deficit pasture mass. In some examples a deficit pasture mass includes a negative deviation, or negative second deviation, from the target pasture mass of the pasture mass, or predicted future pasture mass, for one or more of the two or more target geographical areas.
The system can estimate feed supplementation and suggest paddock fertilisation. In some examples the method will estimate whether a farm will be moving into a surplus or deficit of pasture mass, and suggest what paddocks to drop out of rotation, or how much additional supplement may be required.
In some examples the method utilizes an Artificial Intelligence (Al) or a Machine Learning (ML) process to learn pasture growth patterns and give recommendations on pasture mass and/or for pasture management. In some examples the method utilises smartphone camera optics and an Al or ML process to scan paddocks and obtain an immediate assessment of pasture mass
Figure 4 schematically illustrates another example method of the invention. In a first example step 400 a plurality of paddocks, say paddocks a-p, including all or part of a farm are mapped as previously discussed with reference to step 100 of Figure 1 and Figure 2.
In a next example step 402 pasture mass data and mob data are obtained as previously discussed with reference to 110 and 120 of Figure 1. The farmer also updates mob details, rotation lengths and pasture and supplement intake targets. In some examples the method allows the farmer to complete a 'part farm measurement'. For example, measure, say, 10 of, say, 40 paddocks and use an Al or ML process to estimate/derive the rest. In some examples estimates are based on derived and farmer entered grazing plans, pasture growth rates and learned growth information (see step 405) about each of the paddocks.
In some examples a "pasture wedge analysis" runs at step 403. The method ranks paddocks from most mass to least as a bar graph (see Figure 3) and splits the ranked paddocks into three subsets: longest/high mass 330, middle/medium mass 340 and shortest/low mass 350 paddocks. In some examples the longest subset includes, say, the 20% of paddocks with the highest pasture mass. In some examples the shortest subset includes, say, the 20% of paddocks with the lowest pasture mass. In some examples the middle subset includes, say, the remaining 60% of paddocks with a pasture mass between longest and shortest subsets.
In some examples the method calculates, for each of subset, the deviation of each paddock from a target line 320 and calculates the sum squares of error for each portion of the wedge. In some examples a user can be presented with warnings based on the average deviation of the subset. For example, if there is a positive deviation about level 320 in a section of the pasture wedge it correlates as a 'current or emerging surplus' or cows are under-grazing/level being too much. The method also calculate an optimisation level for the current wedge i.e., 100% optimization equals all paddocks lining up with the target line. This can be communicated to a user. This helps a farmer increase the optimisation by seeking scenarios that maximise the optimisation level.
A farmer may enter or generates pasture mass estimates on distinct days. In step 404 paddock growth rates and farm average pasture mass are calculated based on previous and current pasture measures and days between measures.
In step 405 a squares regression-based approach is used to calculate relative growth rate solutions for each paddock a-p based on historical growth rates calculated in step 404. In some examples this result is stored and used for future pasture growth rate predictions. For example, a particular paddock might be growing at 120% of the average growth rate of other paddocks on the farm. This relative rate is then used as part of an optomisation to grow the paddock at a faster or slower rate in the future. In some examples the calculated relative and average growth rate solutions are also used to calculate a total pasture production per paddock which is then presented to the farmer in a visualisation. In some examples at step 406 a farmer inputs expected pasture growth rates for the next 14 days based on experience. In some examples the system will use the latest pasture growth rate estimated. In some examples an Al or ML method is used to forecaster/predict pasture growth rates.
In step 407 a baseline proprietary paddock and supplement allocation method identifies for each within daytime period and each mob a paddock to graze and for how long and when to move the mob to a new paddock and to which paddock. Each day the paddocks are tracked, grazed, and grown to determine the next day or grazing pasture mass based on data from steps 405 and 406. In some examples the method is used for multiple mobs on the farm and ensures multiple mobs are not grazing the same paddock on the same day.
In some examples where paddocks that don't quite have enough pasture for a full graze (or vice versa) supplement feed is indicated as required for that day to ensure even feed supply. In some examples a paddock is identified as AM grazing only or PM grazing only. In some examples mobs may be excluded/included from paddocks, differing ratios of pasture and supplement intake at AM and PM allocations, fixed and min-max supplement allocation to determine paddock and supplement allocations by mob.
At step 408 a "forecast wedge analysis" is run based on the daily forecast pasture masss calculated in step 407.
If a surplus is forecast in steps 403/408 an optimisation method at 409 identifies the optimum number of paddocks to conserve (i.e., leave out of the grazing rotation), and the suggestion is then delivered to the farmer via the App.
If a pasture deficit is forecast in steps 403/408 an optimisation method at 410 identifies an optimum number of paddocks to apply nitrogen to and the amount of nitrogen needed that will correct the forecast feed wedge and be profitable based on the degree of the deficit, milk price, amount of nitrogen already applied in the season and the cost of nitrogen, and the suggestion then delivered to the farmer via the App, txt, WhatsApp.
If a pasture deficit is forecast in steps 403/408 an optimisation method at 411 identifies the amount of supplement to feed based on the degree of the pasture deficit, and the suggestion then delivered to the farmer or farm worker via the App, text message, WhatsApp, email, or mobile device notification.
In some examples, once the plan is finalised the relevant decisions e.g., paddocks to conserve, amount of supplements to feed, nitrogen to apply, paddocks to graze when and by which mob and projected average feed mass are then communicated to relevant farm staff, such as a farmer or farm worker, via App, SMS, text message, WhatsApp, email or mobile device notification.
In some examples a proprietary paddock and supplement allocation method is used to identify for each within day time period (e.g., AM and PM) and mob (group or herd of animals), a paddock(s) to graze for how long and when to move a mob to a new paddock(s) and which paddock(s). Each day paddocks are tracked, grazed, and grown to determine the next day or grazing pasture mass. In some examples the method works for multiple mobs and paddock(s) on a farm and ensures multiple mobs are not grazing the same paddock(s) on the same day. In some examples each mob is given a colour or indication icon or symbol. Colours or indication icons or symbols can be overlaid on a visual map of the paddocks, such as the visualisation on Figure 2, as a visual indication to a farmer of where and when a mob should be grazing a particular paddock.
In some examples, if paddocks that don't quite have enough pasture for a full break (or vice versa) more (or less) supplement is suggested/allocated for that day to ensure even mob feed supply. In some examples a proprietary optimisation method is used to find the amount of supplement to feed based on the degree of one or more of a pasture deficit, , and the suggestion then delivered to the farmer via the SMS, text message, WhatsApp, email, App or mobile device notification.
In some examples the pasture management plan includes one or more of paddocks to conserve, amount of supplements to feed, nitrogen to apply, paddocks to graze when and by which mob and projected average feed mass. In some examples this is communicated to relevant user and/or farm staff via App, or mobile device notification.
In some examples of the technology a user, typically although not exclusively farmer or farm worker, uses the technology by setting up details of a farm, uploading or providing a map of the farm, entering pasture mass for paddocks or distinct areas of the farm. The technology then uses one or more of proprietary algorithms, proprietary information, artificial intelligence (Al) and/or machine learning (ML) to provide predictions that assist the user to optimise pasture management functions. In some examples, results and/or historical data are used my Al and/or ML processes to continuously learn about the farm to improve future predictions.
In some examples the technology includes, but are not limited to, one of more of the following:
• User entry or import of pasture mass from any existing device
• Use Part farm walk functionality for assessing pasture mass
• Feed Wedge (Figure 3) insights and recommendations to optimise pasture quality
• Generation of grazing and supplement allocation plans on demand
• Forecast of pasture growth down to individual paddock level
• Seasonal and performance-based insights to optimise conservation decisions e.g., to conserve silage or take out of rotation for renovation
• Smart device, or computer, animations and map-based visualisations showing effects of pasture decisions
• Communication to users of grazing and supplement allocation plans
• Entry of pasture mass in an interactive and visual wizard on a smartphone using a pasture measurement technology.
• Ability to carry out part farm walks at the busiest times of the year saving time.
• Access feed wedge insights and recommendations to optimise pasture intake and pasture quality, supplement, and fertiliser use.
• Automatic generation of grazing and supplement allocation plans for each mob on the farm and the ability to tune plans.
• Forecast of paddock pasture mass based on learned knowledge of each of the paddocks so users have the most accurate information.
• Easily identify and track underperforming paddocks and cultivars to optimise pasture renovation decisions
• Test, simulate and visualise short term management decisions and scenarios, and see impact on pasture optimisation, mass and performance against targets.
• Map based and tabular communication of upcoming paddocks to graze users via a smartphone.
Some examples of the technology operate over a single farm. Other examples of the technology operate over multiple farms.
Intense cost pressures and skilled labour shortages are challenging many farmers. While farmers have little control over their cost inputs and pay-outs, farmers do have control over arguably one of the most valuable resources (after the farm herd or mob of course), which is pasture. The present technology allows helps farmers focus on the herd or mob primary feed source and generates insights and plans that have a direct effect on pasture intake, quality and consequently farm profit.
Examples of the technology focusses intensely on finding the best insights and foresight for the next two to three weeks giving you the clarity and confidence of your decisions.
Examples of the technology generate detailed and informed paddock and supplement allocation plans for each mob on the farm that optimise pasture usage and continues to learn the uniqueness of each farm and paddock. All plans have the flexibility of being fine-tuned and re-generated to suit each farmers individual requirements.
Examples of the technology are designed specifically as a digital assistant and coach it can be used directly by farm staff irrespective of experience rather than needing highly trained individuals to use and operate the software.
In some examples individual images with matching pasture mass estimates derived from a rising plate meter are used as training data form an Al or ML process to estimate pasture mass from images. Images are processed and Red:Green:Blue (RGB) statistics derived from each image. Using this data, two machine learning approaches (XGBoost and Random Forest) can be used to develop models that predict pasture mass accurately from new processed images.
Referring to Figures 5 and 6, in some examples of the invention images are obtained by using a mobile device to take video in each paddock. The method processes the images and uses ML models to estimate pasture mass for each paddock. In some examples a GPS stamp on the video is used to match the video to a particular paddock a to p rather than the farmer having to choose the paddock.
In some examples an imaging device may be provided on selected animals in a mob, say mounted on a collar or bridle. At certain times of the day i.e., when it is known that the cows (mobs) are about to enter paddocks or will be in paddocks, video footage of the pasture will be captured and the ML models will be used to estimate pre-grazing and post-grazing pasture mass, and grazing events for each paddock. In some examples an imaging device may be provided on a farm/work bike. As a farmer goes in and out of paddocks or goes up and down a race, video will be captured, and the technology would process the images and use the ML models to estimate pasture mass for each paddock which is processed and calculated in the mobile device or in the cloud.
Figure 7 schematically illustrates yet another example method of the invention. A 240-to-300-degree circular videos of pasture is captured, say via a smartphone. The video is split into individual images and the images are trimmed to remove the sky and horizon, so they include only pasture: 610/620.
Machine vision models were training using XG Boost and Random Forest methods with matching video spectral RGB data and pasture mass measures from a platemeter, or C-Dax Pasture Meter. Feature engineering is used to identify the most informative spectral statistics for seasons and regions: 630/640. In some examples the trained machine vision models are used to estimate pasture mass for each new video and images. In other examples pasture masses are estimated manually using devices such as platemeters, C-Dax Pasture Meters or an experienced eye: 650.
Once the pasture masses are estimated, a 'wedge analysis' is performed: 660. This ranks padlocks from most mass to least mass as a bar graph and into three sunsets: longest (20%), middle (60%) and shortest (20%) of paddock pasture mass - see for example Figure 3. For each of these subsets the deviation of each paddock from the target mass line (minimum = target post grazing mass, maximum = required pre-grazing mass to meet pasture intake and rotation length targets). The sum squares of error for each portion of the wedge is calculated and each paddock is assigned an optimisation score from 0 to 100%; where 100% is perfect alignment with the target mass line and individual paddock bars ranked from highest to lowest pasture mass. Each of these wedge results is match with a predetermined recommendation that is communicated to the user.
Based on previous and current pasture measures and days between measures, paddock growth rates and farm average pasture mass are calculated. A linear regression-based approach to calculate relative growth rate solutions for each paddock based on historical growth rates. In some examples trained Al and or ML method is used to estimate current paddock growth rates and farm average pasture mass based on historical training/trained data.
A grazing / pasture management plan is determined and communicated to a user: 680. In some examples the plan is part based on a pasture mass forecast where all paddocks are grown up based on the user defined forecast pasture growth rate and/or previously calculated growth rate solutions. This relative rate is used to grow paddocks at a faster or slower rates depending on the prior learned knowledge.
In some examples the plan is part based on split AM and PM or AM/PM grazing days for paddock or combinations of paddocks that best matches the desired pasture intake and characteristics of the mob. Considerations include whether paddocks are only grazeable in the AM or PM, or corresponding buddy paddocks that are always grazed together because they share adjacent gates or could be grazed together. In some examples the plan is part based on a pasture mass needed to meet total animal mob pasture demand. In some examples the plan is part based on supplement needs to fill pasture gaps for each mob.
In some examples a production output forecast can be estimated and communicated to a user: 690.
Unless the context clearly requires otherwise, throughout the description and the claims, the words "comprise", "comprising", and the like, are to be construed in an inclusive sense as opposed to an exclusive or exhaustive sense, that is to say, in the sense of "including, but not limited to".
The entire disclosures of all applications, patents and publications cited above and below, if any, are herein incorporated by reference.
Reference to any prior art in this specification is not, and should not be taken as, an acknowledgement or any form of suggestion that that prior art forms part of the common general knowledge in the field of endeavour in any country in the world.
The technology may also be said broadly to consist in the parts, elements and features referred to or indicated in the specification of the application, individually or collectively, in any or all combinations of two or more of said parts, elements or features.
Where in the foregoing description reference has been made to integers or components having known equivalents thereof, those integers are herein incorporated as if individually set forth.
It should be noted that various changes and modifications to the presently preferred embodiments described herein will be apparent to those skilled in the art. Such changes and modifications may be made without departing from the spirit and scope of the technology and without diminishing its attendant advantages. It is therefore intended that such changes and modifications be included within the present technology.

Claims

1. A method of developing a pasture grazing allocation for a group of animals, the method including: obtaining pasture mass data for two or more target geographical areas, estimating pasture growth data for the two or more target geographical areas obtaining mob data for at least one group of grazing animals, and based on pasture growth data and mob data, determining a pasture grazing allocation of the group of animals in the two or more target geographical areas, wherein the pasture grazing allocation includes at least one time and one or more target geographical area for grazing the group of animals.
2. The method of claim 1 wherein the pasture mass data includes an estimate of pasture mass for the two or more target geographical areas, and/or an observation of the pasture mass for the two or more target geographical areas, and/or a partial observation of pasture mass and a pasture mass prediction based on the partial observation and historical pasture growth data.
3. The method of claim 1 or 2 wherein the estimate of pasture mass includes obtaining a video of two or more target geographical areas and based on machine learning processing the video to making a prediction of pasture mass for the two or more target geographical areas.
4. The method of claim 1 or 2 wherein the estimate of pasture mass includes obtaining pasture massage data using one or more of a plate-meter, C-Dax Pasture Meter, or human observation.
5. The method of any preceding claim wherein estimating pasture growth data includes estimating pasture growth rates and mass pasture, and/or obtaining pasture mass data for the two or more target geographical areas at two or more distinct periods of time, and/or a sum-of-squares regression.
6. The method of any preceding claim wherein the method includes establishing a target pasture mass for the two or more target geographical areas, and obtaining a deviation from the target pasture mass of the obtained pasture mass data for the two or more target geographical areas.
7. The method of any preceding claim wherein the method includes: displaying to a user, on a portable device screen, a graphical representation of the obtained pasture mass data for the two or more target geographical areas, and/or displaying to a user, on a portable device screen, a graphical representation of the target pasture mass, and/or displaying to a user, on a portable device screen, a graphical representation of the obtained deviation, and/or
8. The method of any preceding claim wherein the method includes displaying to a user, on a portable device screen, a graphical representation of the obtained pasture mass data for the two or more target geographical areas and organizing the two or more target geographical areas according to obtained pasture mass.
9. The method of any preceding claim wherein the method includes establishing predicting a future pasture mass for the two or more target geographical areas, and obtaining a second deviation from the target pasture mass of the predicted future pasture mass data for the two or more target geographical areas.
10. The method of any preceding claim wherein the method includes determining, or predicting, a surplus pasture mass, and wherein the surplus pasture mass includes a positive deviation, or positive second deviation, from the target pasture mass of the pasture mass, or predicted future pasture mass, for one or more of the two or more target geographical areas.
11. The method of any preceding claim wherein the method includes, or predicting, a deficit pasture mass, wherein the deficit pasture mass includes a negative deviation, or negative second deviation, from the target pasture mass of the pasture mass, or predicted future pasture mass, for one or more of the two or more target geographical areas.
12. The method of any preceding claim wherein the method includes calculating, based on a group of grazing animals, pasture mass and growth rates, a pasture management and mob (animal) paddock rotation/feeding plan.
13. The method of any preceding claim wherein the method includes estimating feed supplementation and paddock fertilization.
14. The method of any preceding claim wherein the method includes estimating whether a farm will be moving into a surplus or deficit of pasture mass, and suggest what paddocks to drop out of rotation, or how much additional supplement may be required.
15. The method of any preceding claim wherein the method includes estimating utilizes an Artificial Intelligence (Al) or a Machine Learning (ML) process to learn pasture growth patterns for one or more of the two or more target geographical areas and give recommendations on pasture mass and/or for pasture management.
16. The method of any preceding claim wherein the method includes estimating utilizes smartphone camera optics and an Al or ML process to scan paddocks and obtain an immediate assessment of pasture mass.
17. The method of claims 15 or 16 wherein the method includes estimating Al or ML process is trained using historical pasture mass data and/or historical pasture growth data for a geographical are.
18. The method of claims 15, 16 or 17 wherein the Al or ML process is trained using accumulated growth rates for individual paddocks over the course of the farming season, and/or a dataset of over 10,000 historical images with matching pasture mass estimates derived from a rising plate meter.
19. The method of any one of claims 15 to 18 wherein the images are processed and Red: Green: Blue (RGB) statistics derived from each image.
20. The method of any one of claims 15 to 18 wherein the machine learning approaches (XGBoost and Random Forest) are applied to develop models that predict pasture mass accurately from new processed images.
21. The method of any preceding claim wherein the method includes providing a digital version of the farm and paddock maps.
22. A method of developing a pasture grazing allocation for a group of animals, the method including: setting up details of a farm, uploading or providing a map of the farm, entering pasture mass for paddocks or distinct areas of the farm using an artificial intelligence (Al) and/or machine learning (ML) to provide predictions that assist the user to optimize pasture management functions. using the results and/or historical data are used my Al and/or ML processes to continuously learn about the farm to improve future predictions.
23. A device for operating the method of any proceeding claim, the device including a processor, a processor readable memory, an input interface, and a display, wherein the processor readable memory includes instructions that when operated by the processor cause the device to perform the method of any proceeding claim.
24. A method of developing a pasture grazing allocation for a mob of animals including obtaining pasture mass data for two or more target geographical areas of a farm, obtaining mob data for at least one group of grazing animals and based on pasture mass data and mob data, determining a pasture grazing allocation of the group of animals in the two or more target geographical areas, and communicating to a farmer or fa rm worker a pasture grazing allocation including at least one time and one or more target geographical area(s) for grazing the mob of animals.
25. A method of developing a pasture grazing allocation for a mob of animals including obtaining pasture mass data for two or more target geographical areas of a farm, obtaining mob data for at least one group of grazing animals, estimating pasture growth data for the two or more target geographical areas and based on pasture growth data, pasture mass data and mob data, determining a pasture grazing allocation of the group of animals in the two or more target geographical areas, and communicating to a farmer or farm worker a pasture grazing allocation including at least one time and one or more target geographical area(s) for grazing the mob of animals.
EP24789130.2A 2023-04-13 2024-04-15 Device, method and system for pasture estimation and management Pending EP4694675A1 (en)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
NZ23799052 2023-04-13
NZ23800589 2023-06-02
NZ23800891 2023-06-14
PCT/NZ2024/050040 WO2024215212A1 (en) 2023-04-13 2024-04-15 Device, method and system for pasture estimation and management

Publications (1)

Publication Number Publication Date
EP4694675A1 true EP4694675A1 (en) 2026-02-18

Family

ID=98518263

Family Applications (1)

Application Number Title Priority Date Filing Date
EP24789130.2A Pending EP4694675A1 (en) 2023-04-13 2024-04-15 Device, method and system for pasture estimation and management

Country Status (1)

Country Link
EP (1) EP4694675A1 (en)

Similar Documents

Publication Publication Date Title
CN118735202B (en) An intelligent decision-making management system for pig farming based on multimodal information monitoring
Jerhamre et al. Exploring the susceptibility of smart farming: Identified opportunities and challenges
KR101934961B1 (en) Method and apparatus for predicting profitablilty
US20180059691A1 (en) Predicting Crop Productivity via Intervention Planning on Small-scale Farms
Shalloo et al. A review of precision technologies in pasture-based dairying systems
Rathod et al. Precision dairy farming: Opportunities and challenges for India
JP7653681B2 (en) Information processing device, information processing method, and program
US20210256631A1 (en) System And Method For Digital Crop Lifecycle Modeling
US11096375B2 (en) Smart cattle reproduction management and digital farm to market transparency metrics
WO2024215212A1 (en) Device, method and system for pasture estimation and management
Hütt et al. UAV LiDAR-based grassland biomass estimation for precision livestock management
CN116579508A (en) Fish prediction method, device, equipment and storage medium
Eckhardt et al. Insights into thermal stress effects on performance and behavior of grazing cattle via multimodal sensor monitoring
Freier et al. Dynamic interactions between vegetation and land use in semi-arid Morocco: Using a Markov process for modeling rangelands under climate change
EP4694675A1 (en) Device, method and system for pasture estimation and management
Hofmann An evaluation of GPS technology as a tool to aid pasture management
Ricci et al. Does diverse grazing behavior of suckler cows have an impact on predicted methane emissions?
Das et al. An overview of data analytics in smart agriculture
Sallam et al. Smart livestock farming: present status, opportunities, and future trends
WO2022115916A1 (en) Livestock monitoring and management
Derakhshan et al. Body rumen fill Scoring of dairy cows using digital images
US20250166091A1 (en) Systems, devices and methods for managing a herd of livestock
Casten Carlberg et al. Artificial intelligence in agriculture: Opportunities and challenges
Burman et al. Digitalization in Indian agriculture
CN121352398A (en) Grassland-livestock balance monitoring methods and devices, electronic equipment, and storage media

Legal Events

Date Code Title Description
STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE

PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE

17P Request for examination filed

Effective date: 20251112

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC ME MK MT NL NO PL PT RO RS SE SI SK SM TR