Product B: automatic detection of lameness in dairy cows

1 - Introduction

Lameness is one of the most common health issues in dairy cattle. Lameness is associated with a decrease in milk yield, a negative impact on reproductive performance, an increase in the risk of culling, an increase of the production costs and it negatively affects animal well-fare. The detection of lameness is one of the most important tasks for its control, however existing subjective methods for its diagnosis, demands time and could suffer from poor reliability within and between observers. So it can be concluded that the development of an objective, reliable and accurate method for lameness detection is necessary for the management of the modern dairy farm.

2 - Objectives

The overall goal of the "Product B" group of the BioBusiness Project is to create a business solution for the lameness detection problem in dairy farms according to the Precision Livestock Farming (PLF) scheme.

3 - Methodology Overview

The goal of a PLF scheme is to create a feedback information cyclus in order to obtain a predictive mathematical model with the purpose of better control of the productive system. The predictive model of the PLF requires a continuous (real time) and accurate input of information. In the specific case of the "Product B" group on BioBusiness project, the inputs of the model will be chosen after analysis of video-images of the cows' gait when walking through an alley after milking, and from physiological, behavioural and production information from the routine database of the farm. The video-images and the production data will be analyzed in order to determine which of them (or both) could be useful as lameness indicators and could serve as input data for the predictive model. In order to validate the model for lameness detection, the model outputs will be compared with the analysis of gait/posture evaluated by means of a subjective method for lameness detection (the golden standard). To increase the range of the data input of the model, the database that will be used in the experiment of the "Product B" group, will come from two farms with two different milking systems:

  • "farm A" with a Voluntary Milking System (VMS);
  • "farm B" with a rotary milking parlour (conventional).

4 – Expected Results

The main result consists of a reliable, accurate and valid automatic system based on predictive mathematical model approach to detect lameness in dairy cows in an early stage.

Possible lameness product ideas Using visual techniques to develop an automatic system for continuous on-farm detection, prediction and control of lameness in dairy cows based on the measurement of different gait/posture characteristics.

5 - Partners and Functions

Katholieke Universiteit Leuven

  • Experimental design
  • Development of the predictive model

Agricultural Research Organization

  • Experimental setup
  • Data gathering

ASG Wageningen UR

  • Validation of the model
  • Labelling

DeLaval International AB

  • Equipments and support
  • Business solution, product commercialisation