Precision dairy part

  • Created by: Cat
  • Created on: 09-05-18 19:25

Precision farming

Precision farming:  Gathering and processing information to improve the precision of resource management • Biggest impact so far has been in the arable sector: • Rather than apply e.g. fertiliser at a standard rate • … we can map crop growth and apply it only where it is needed

Livestock production has been intensified to help us control production (at the group level) • Precision livestock farming (PLF) is changing this:

Gather data from individual animals so we can then manage them as individuals •

Much closer monitoring and control 

Increased use of robotics 

Greatest initial uptake is in the dairy sector

Killer app = AI requires reliable oestrus detection 

Easier return of investment due to the value of milk over a cow’s lifetime

1 of 10

How precision farming works

 External, manual and sensed data ->integrated data ->semi automated decision making - improved manual control External control, Semi-automated decision making-> Integrate/monitor/document outcomes

Electronic identification (EID) 

Whilst GNSS (Global Navigation Satellite System e.g. GPS) is arguably the ‘backbone’ of precision arable farming • EID is arguably the backbone of Precision Livestock Farming as it enables individuals to be identified, monitored and receive targeted management •

Radio-frequency ID (RFID) is the most widely used EID with livestock

• A transponder on the animal wirelessly transmits a unique code to the RFID reader

2 of 10

Types of RFID transponders

Active: Transmitter includes its own power (battery)

• Larger and more expensive than passive

• Can store data • Longer read range compared to passive so more reliable

• Greater use in future?

Passive receiver • Very small, low cost • Cannot store data • Reduced read range compared to active • Currently the main type used in livestock EID • Wireless energisation of the RFID transmitter by the reiceiver • Very small, low cost • Cannot store data • Reduced read range compared to active • Currently the main type used in livestock EID

3 of 10

Attaching an RFID

  S/C tags unsuitable for meat animals

Ear tag

• Less likely to injure the animal when attaching tag • Can be accidentally lost • Easier to fraudulently ‘swap’ or remove tags • Easy to recover at slaughter • Main type used in UK

Rumen bolus

Improper dosing may seriously injure the animal • Unlikely to be lost • Very difficult to fraudulently remove from animal • More difficult to recover at slaughter

Importance of oestrus detection • Oestrus detection is economically important in dairy farming because: • Inseminate → Pregnant → Calf → Milk • Cost of wasted semen from ‘good’ bulls • Although disease is important, sick cows produce some milk

4 of 10


Main aim goes from finding food to finding a mate

• Activity levels increase on the day of oestrus: • 2.3 times higher (Schofield et al., 1991) • 4 times higher in cows in free stalls (Kiddy, 1977)

• Detect using pedometers and/or accelerometers

• On the day of oestrus i.e. the day of successful insemination, Halli et al. (2015) found: • Number of visits to feeder declined 9.1% • Fresh matter intake declined 10.3% • Time spent feeding declined 20.8% Intake rate increased

Activity changes during oestrus

• First generation activity monitors were simple leg-mounted pedometers only capable of determining activity level

• Second generation activity monitors included triple-axis accelerometers • As well as activity level, these can determine the number of steps and whether or not the cow is lying or standing, making oestrus detection more reliable

5 of 10

Challenges in measuring activity

 • Activity-based sensors for oestrus detection in dairy cows were largely developed (and work best) for housed cows

• They are less reliable in cows out at pasture as they exhibit a lot of activity (moving) when grazing (compared with housed cows that are largely stationary when eating a TMR) 

• Traditional tail paint or mounting detectors (e.g. Kamar) work better at pasture as the cows exhibit more mounting behaviour (cf. indoors)

Detecting oestrus at pasture

• Tailpaint and Kamars traditionally relied on visual inspection by a human • LIC (New Zealand) have developed a camera-based system that can automatically sense when a heat detection patch has been activated • Cows can be automatically drafted after milking for AI

Using a bull to detect oestrus • Arguably the best detector of a cow in heat is a bull! • MooCall HEAT uses an RFID reader in a collar on a vasectomised bull to detect an RFID tag on each cow • The pattern of bull-cow interactions can determine which cows are in oestrus • Algorithms are designed to work with animals at pasture

6 of 10

Recording eating and intake

 • Foraging animals are engaged bites, chews, chew-bites and body movements: • Head movements at a feed face • Head and body movement when grazing • Consequently, accelerometer signals from animals that are eating are very complex • Cows are generally stationary when ruminating (lying down or standing still), and they only perform one type of jaw movement (chewing)

• In contrast to eating, rumination is comparatively easy to detect • Dairy tech companies have gone for the “low hanging fruit” when looking to use a ‘foraging’ measure to detect the reduction in feed intake during oestrus

Rumination and DMI • Rumination time (RT) is linked to recent intake: • RT increases 4h after a period of high DMI (Schirmann et al., 2012) • RT made a significant but small contribution in a DMI prediction model (Clement et al., 2014) • After controlling for between-cow variability, RT can predict DMI (Johnston and DeVries, 2015) • Halli et al. (2015) found on the day of oestrus : • intake declined 10% but feeding time declined 21%

7 of 10


automatic cluster removal, robotic feeder, automatic gate(integrating pasture access and automatic milking), automatic milking systems, robotic feed pusher, robotic electric fence, robotic slat scraper, rumination monitoring PH (detect acidosis, monitor rumen function), cow location-track movements, find cows, assess social structure, feed intake and weight gain to select FCR efficient cows to breed, footbath-automatic maintenance, leg movement pedometer- detect oestrus, locomotion recording-lameness, milk yield-individual cow yield and quality data, neck-mounted accelerator- oestrus 

Milk analysis

• Robots can now incorporate milk chemistry analysis, including yield, conductivity etc. • DeLaval Herd Navigator: • Progesterone (oestrus) • Lactase dehydrogenase (mastitis) • Beta hydroxybutyrate (ketosis) • Urea (feed protein balance)

BCS camera with 3D monitoring 

In future need integrated methods- not just lots of single PLF gadgets. farmer needs to pick and choose what suits their farm 

8 of 10

Lecture 2

Precision livestock farming (PLF) is changing this: - Gather data from individual animals so we can then manage them as individuals - Much closer monitoring and control - Aim is to improve the efficiency of resource use by only using them when needed 

Grazing management Measure the available herbage (kgDM Ha-1) • Match this to the intake requirement of the animals to be grazed • Control access e.g. using ***** grazing

time release gates, robotic electric fencing 

Sensor platforms- animal, cameras(static), in future roving? Drones?

Electronic identification • Electronic identification (EID) is the core technology in PLF • Also known as RFID • Passive tags are energised by the reader causing them to transmit their ID code • Active tags have their own power and have a longer read range but cost more • Each animal carries its own unique ID, either in an ear tag or a rumen bolus • Allows identification of individual animals

9 of 10

static sensors

It can be difficult to accurately match lambs to their mothers • The ‘Pedigree Matchmaker’ system uses EID to record which lambs are associated with which ewes as they pass an EID reader • Typically as they pass through a gate as they access water • Gives reliable dam pedigree information

Walk Over Weighing • EID can be used with a weigh platform to automatically weigh animals as they cross it (WOW) • Typically locate it between two resources (e.g. pasture and water) • Challenges around weighing and reliable EID from moving animal • Potential to give growth rates: - Monitor production - Detect sick animals - Target e.g. wormer use

Cameras • Cameras can be used to automatically monitor animals that pass near them • EID used to identify the individuals in view • Vision processing used to derive management information • e.g. DeLaval uses a 3D camera to body condition score cows

Accelerometers everywhere! • The development of cheap tripleaxis accelerometers is revolutionizing the capture of animal behaviour data • Includes human behaviour: • Nintendo Wii Remote (games) • Smart phones (e.g. VR apps) • Smart watches (fitness)

10 of 10


No comments have yet been made

Similar All resources:

See all All resources »See all Precision dairy resources »