Sie sind auf Seite 1von 73

Value Creation through Precision Livestock

Farming systems
Daniel Berckmans

M3-BIORES KU Leuven - Belgium


Measure, Model & Manage Bio Responses
Conference on State-of-the-Art Technology to Drive
Agriculture Productivity in the Next Quarter of the Century
28-30 June 2016
Tokyo, Japan

Overview
Challenges for Livestock Production
What is PLF?
Basic principles of PLF
Examples of PLF systems
Business Models

Conclusions
M3-BIORES KU LEUVEN

M3-BIORES team

253 A-Publications
359 Conference papers

17 products
2 spin-off companies

15 patents
3

Challenges for livestock production

Problem of monitoring animals

Livestock farming in the past

Farmer had the time to use audio-visual scoring


5

Worldwide Individual Meat Consumption

Source: FAO (2010)


6

Numbers of animals
Number of farmers

Resulting in

High number of animals per farm


Less available time per individual animal
More welfare and other problems
8

Challenges for livestock production


Over 60 billion animals are slaughtered every year, increase
with up to 40%?
Health: Relationship between animal health and healthy food
Animal welfare (e.g. EU)
Environmental Issues
Social importance
Economic importance including Valorisation of knowledge

3 approaches in Europe with focus on


the animal

10

First approach: Iceberg Indicator (1)

M3-BIORES KU LEUVEN

11

Second approach: Welfare Quality (2)

M3-BIORES KU LEUVEN

12

Europe has invested in a methodology to quantify


Animal Welfare
Welfare Quality

Procedure: Experts do audio-visual scoring by visiting


farms and looking to (behaviour) of animal.

13

TomorrowAutomated Systems
Technology can help to quantitatively measure behaviour,
health and performance of animals.

M3-BIORES KU LEUVEN

14

Welfare Scoring/Monitoring
Iceberg
indicators
Welfare Quality
once a year

slaughterhouse

Time
Time

PLF-Continuous animal based


management during growth
period
15

What is Precision Livestock


Farming (PLF)?

Precision Livestock Farming

Management of livestock by continuous automated real-time


monitoring of production/reproduction, health and welfare of
livestock and environmental impact.

17

Internationalisation
7 European PLF Conferences

4 Smart Sensors Workshops


2000: Silsoe
2002: Bremen
2004: Leuven
2006: Gargnano

EC-PLF
CIGR
ASABE
EurAGENG
ISAH
...

EU Projects:
-Bright Animal
-ALL-Smart-Pigs
-BioBusiness
-EU-PLF

M3-BIORES KU LEUVEN

18

The basic principles of PLF

Position Steering
wheel

Process

Direction

MEASURE

MEASURE

PREDICTION

1
FEEDBACK

Desired direction
MODEL-BASED
CONTROLLER
M3-BIORES KU LEUVEN

20

Methodology
Position and balance
on the board

MEASURE

Direction

MEASURE

PREDICTION

FEEDBACK
3

Desired direction

Prediction-based
controller
21
M3-BIORES KU LEUVEN

1991: Modern control theory applicable to living organisms?


Process
MICROENVIRONMENT

DYNAMIC
BIORESPONSE

MEASURE

MEASURE

PREDICTION
MODEL
FEEDBACK

3
DESIRED PROCESS
OUTPUT

MONITOR/
REGELAAR
M3-BIORES KU LEUVEN

22

Model-based Monitoring/Improvement of
Bioresponses (1991)

Improvement

Monitoring

Process
MICROENVIRONMENT

MEASURE

DYNAMIC
BIORESPONSE

MEASURE

Algorithms
FEEDBACK

3
Controller
DESIRED PROCESS OUTPUT
M3-BIORES KU LEUVEN

23

A living organism is a CITD system

M3-BIORES KU LEUVEN

24

A living organism:
Complex

M3-BIORES KU LEUVEN

25

Living organisms are ...

M3-BIORES KU LEUVEN

26

...individually different

M3-BIORES KU LEUVEN

27

A living organism:
Individual

Heartbeat (bpm)

Complex

Time (s)

M3-BIORES KU LEUVEN

28

A living organism:
Complex

Individual

Time
N

population
Individual

Individual

Response
variable

Av. population
29

A living organism:
Individually different

Identical

M3-BIORES KU LEUVEN

30

A living organism:
Complex

Individual

Time-Varying

17

13

16

12

HEAT PRODUCTION (W/KG)

HEAT PRODUCTION (W/KG)

Example: Heat production of broiler chickens


5 days old
30 days old

15

14

13

12

11

10

8
11
0

TIME (HOURS)

7
0

TIME (HOURS)
MEASURED
MODELLED (1ST ORDER)
MODELLED (2ND ORDER)

MEASURED
MODELLED (1ST ORDER)

M3-BIORES KU LEUVEN

31

A living organism:
Complex

Individual

Time-Varying

Dynamic

Living organism = CITD - system

1. Measure
2. Model

3. Manage &
Monitoring

In an on-line way

M3-BIORES
M3-BIORES KU LEUVEN

32

Examples of PLF
Technology:
What is possible today?
Fully automated monitoring

M3-BIORES KU LEUVEN

33

Infection Monitoring by
On-line Pig Sound Analysis
i.c.w. University of Milan, SoundTalks NV, Fancom BV

M3-BIORES KU LEUVEN

34

Health monitoring by on-line sound analysis:


On-line cough recognition algorithm in pig stables

35LEUVEN
M3-BIORES
M3-BIORES
KU LEUVEN
KU
- PLF Course

35

On-line cough recognition algorithm in pig stables

M3-BIORES KU LEUVEN

36

PCM: Results

Pigs ill again

Pigs ill upon entering

Animals again ill


Animals treated

Animals
treated

37
M3-BIORES KU LEUVEN

M3-BIORES KU LEUVEN

38

MAIN FUTURE APPLICATIONS


REDUCING THE USE OF ANTIBIOTICS
sound
V(t)

micro

T
Q(t)

Sound
analysis

infection

Therapeutic
decision

Climate
controller

Antibiotics

M3-BIORES KU LEUVEN

39

Example: Monitoring of
drinking behaviour of pigs
(health)

M3-BIORES KU LEUVEN

40

Monitoring of drinking behaviour


in pigs (i.c.w. Ughent, Fancom BV)
Monitoring water usage as indicator for health
status
Estimate hourly water use in a pig pen by
analysing hourly duration of drink nipple visits

M3-BIORES KU LEUVEN

41

Model-based monitoring of water use


Images
Water flow
measured
Detection of
visits

Water use
from water
meters

Compare

Water use
estimated
from image
M3-BIORES KU LEUVEN

Transfer
function
modelling

Duration of
visits

42

Model-based detection of visits


Drink
nipple

Threshold

T = duration of the visit

Play
M3-BIORES KU LEUVEN

43

Results

Hourly water use can be estimated with an accuracy of


92% or 200 ml over 13 days
M3-BIORES KU LEUVEN

44

Continuous automated monitoring


of feed intake of broilers by sound
technology

Play

Play

Continuous recording of sounds (top) and individual pecking


sounds (bottom) as extracted by the algorithm.
46

Results and Discussion


Data Set

Number of
pecking
(Algorithm)

Number of
pecking
( Video
Labelling)

Accuracy of
Algorithm
(%)

True
Positive

False
Positive

113

105

93

105

99

95

96

95

109

106

98

106

34

107

101

94

101

35

98

91

93

91

36

95

88

92

88

TotalAverage

3707

3447

93

3447

47

More Results

Minutes

Number of
peckings per
experiment

13

1193

12

759

18,98

0,198

18,78

0,025

10,3

895

24,17

0,222

23,94

0,027

0,92

15

1250

32,50

0,236

32,26

0,026

0,73

13,5

1283

30,79

0,365

30,43

0,024

1,19

15

1460

35,04

0,348

34,69

0,024

0,99

7,04

651

16,28

0,168

16,11

0,025

1,03

4,35

468

12,17

0,111

12,06

0,026

0,91

7,26

533

13,33

0,124

13,20

0,025

0,93

6,54

583

13,99

0,145

13,85

0,024

1,04

7,43

654

16,35

0,165

16,19

0,025

1,01

6,65

573

15,47

0,155

15,32

0,027

1,00

300,10

25285

633,26

6,22

627,04

0,025

0,98

Exp.
Chickens
1
2

feed uptake Feed loss Feed intake Feed Intake


per
per
per
Per Pecking
experiment experiment experiment
(g)
(g)
(g)
(g)
28,63
0,325
28,31
0,024

12

TotalAverage

Feed loss
per
experiment
(%)
1,14
1,04

48

Feed Uptake Per Experiment


(g)

35.00

Feed Intake Per Experiment


(g)

35.00

R = 0.9858

30.00
25.00
20.00
15.00
10.00
5.00
0.00

R = 0.9856

30.00

500

1000

1500

Number of Peckings per Experiment

25.00
20.00

15.00
10.00
5.00
0.00
0

500

1000

1500

2000

Number of Peckings Per Experiment

49

50

Example : Real time monitoring of


problems in a broiler house
i.c.w. Fancom BV

M3-BIORES KU LEUVEN

51

eYeNamic monitor tool

Image
pre-processing

Farm manager
Camera network

Farm network

Monitoring software

M3-BIORES KU LEUVEN

52

Vision-based Early Warning System for


Broiler Houses
Solution?
Farmers can use automatic tools to continuously monitor the welfare and
health of their broilers

53

Experiments ground plan

Detecting malfunctioning in broiler houses


Produce alarms in real-time when malfunctioning happens
(in feeder or drinker lines, light, climate control, etc.)

55

Birds and housing


Experiment rounds: 42 days
Initial broiler weight: weight of 405
grams
Broiler type: ROSS 308 broilers
House capacity: 28000 broilers
Climate control: Fancom FUP1EA2

Farmer logbook and manual video


observation as references

Measured vs. modelled


animal distribution
Measured distribution

Distribution (%)

Predicted distribution

Prediction window: 1 light period = 5 hours


26/10 30/10 03/11

07/11

11/11

15/11 19/11 23/11


Date (dd/mm)

27/11

01/12

05/12

Event detection
Feeder line

Defect Feeder line


Measured values
Smoothed values within 25% range
Smoothed values out of 25% range
Predicted values

Problem in
feeding lines

Distribution
(%)

Normal situation

Date(dd/mm)

Detected events in the validation


experiment over 42 days

Electricity
failure

Light
problems

Water supply
problems

Farmers
inspection

Climate control
system problems

Feeding system
problems

Unloading

Conclusion: Events in a broiler house


could be detected using top-view image
analysis with an accuracy of 95.24 %

Vaccination

Some other examples

61

Cow lameness monitor: i.c.w. Volcani, DeLaval, Wur Aggression monitor: Umil, TIHO, Fancom BV

Play

Play

Scratching behaviour: Ughent, ILVO

Weight estimation: Fancom BV, Agrifirm

62

Play

Play

Real Time Stress Monitoring

Stress Monitoring @ Racing


Joint Development: BioRICS
Problem: high stress levels cause race car pilots to drive at higher risk

Low arousal
High arousal

Play

BioRICS NV

64

Stress@work: positive negative stress


Mental energy

Performance

Optimal
Performance
Undermotivation
Laid back
Relax

Anxious
Distress

Eustress
% Stress

burn-out = no return from distress zone when stressor is gone!


BioRICS NV

Real-time stress monitoring for livestock

Existing BioRICS Stress Level


Monitoring

Existing Sensixa eAr-sensor

New technology Delivers:


Realt Time Stress Level Monitoring
Real Time adaptive Algorithms
Wireless connection & recharging
M3-BIORES KU LEUVEN

66

Business Models

67

Value for
farmer:
Welfare
Health

Environmental
Impact

Euros

Labour and Time

Consumer
Social
Recognition

68

PLF is a tool that helps farmers and stakeholders

Veterinarians
Researchers
Consumer

Press
Citizens

Governments
Companies

69

The PLF Business Model?


Cost of PLF investment & operation shared along the value
creation chain by payment for access to data pool
Tech. Prov.

Breeding

Vets.

Others

PLF Service Provider

Consumer

(funding, set-up, service, data)

Feed

Farmer

Slaughter

Retailer

70

General Conclusions
PLF offers fully automated continuous real time detailed
monitoring and management of animals.
PLF brings the farmer to the individual animals that need
his/her attention, active management tool.
PLF is a tool that helps farmers and stakeholders.
PLF will allow the animals to drive the system.
Efficient implementation of PLF needs collaboration between
researchers, farmers, industry and stakeholders!
71

Acknowledgments and Disclaimer


This project has received funding from the
European Unions Seventh Framework Programme
for research, technological development and
demonstration under grant agreement n 311825
The views expressed in this presentation are the
sole responsibility of the author(s) and do not
necessarily reflect the views of the European
Commission.

72

Thank you
For more information you can check our website:
http://www.m3-biores.be

Contact: daniel.berckmans@biw.kuleuven.be

73