Project Description

COMPUTER VISION AND MACHINE LEARNING IN THE FARMING INDUSTRY

FARMER PIET

BRIEF

Music can affect human emotions. But who would’ve guessed that pigs are also influenced by country music or a Coldplay song? Our customer “Farmer Piet” is a pig farmer in Belgium who was eager to find out whether music affects pigs’ behaviors. If so, how can it improve their overall health?

Using our expertise in computer vision and machine learning algorithms, we got the answer! Through a mobile and desktop app, as well as a strong intuitive dashboard, “Farmer Piet” is now able to monitor and control the effect of music on his pigs. This Mediaan Conclusion solution enables him to manage different types of music for each stable, watch the live detection of the pigs and analyze their behaviors in correlation to the music through the dashboards. How cool is that? Using our Mediaan Conclusion expertise, we guided and supported our customer in an Agile fashion through numerous areas:

  • Computer Vision, Object Detection and -Tracking

  • Machine learning algorithms
  • Front-end development
  • UX design for dashboards and front-ends
  • Microsoft Azure cloud deployment

  • Scaled training of the models in the cloud

THE CONCEPT

Our customer “Farmer Piet” got in contact with a student from the University of Gent with whom he researched how music correlates to the pigs’ behaviors. Results show that pigs recognize sounds which are expressed in their moods and behaviors. Farmer Piet did not stop here and wanted to utilize the information strategically. Can music help him improve the overall health of his pigs? If so, which music do they like? Which music increases their appetite? And how can he automatically manage the music or analyze their behaviors? He has all these burning questions and it was time for Mediaan Conclusion to step in to add some technology that supports finding some answers.

CUSTOMER CHALLENGE

Our Mediaan Conclusion team developed an end-to-end solution that runs both on mobile and desktop. This application allows our customer to manage the music of each stable separately and watch the live detection of his pigs. Our Mediaan Conclusion team also developed a powerful intuitive dashboard where all the data is analysed. This part of the project required the use of computer vision and machine learning algorithms which allows Farmer Piet to accurately track the pigs’ behaviors in correlation to the music. The challenge during this project was that it was quite hard to detect the right movements of the pigs. Mediaan Conclusion made sure to choose the right algorithm to make sure that it’s suitable for the type of data we have.

MEDIAAN IN ACTION

The full solution was developed by a multidisciplinary team of four Medianers; two Data Scientists, a Data Architect and a Front-end Developer. Our team uses the following technology stacks to provide the best possible solution:

  • JavaScript
  • Angular
  • YOLO network
  • OpenCV
  • Microsoft Azure
  • Web API
  • Azure functions
  • .NET
  • Scrum/agile

RESULTS

Farmer Piet is now able to manage the music to each pig stable, watch live detection of his pigs and collect data regarding their movements and behaviors according to the played music. Only after a proper analysis correlations can be determined and question like “how can music affect the overall health of the pigs?” be answered. At this moment the setup is done, so the next step would be to monitor and analyse the data using the intuitive dashboard. Using this dashboard, observing and analyzing the data can be done easily, fast, factually and is not labor intensive anymore. One of the upcoming challenges is to accurately determine the (evolution) weight of the pigs through utilizing the current camera images and positions.

This project also opens the door to other exciting projects leveraging computer vision and artificial intelligence in all different kinds of areas like tracking people, animals and other types of objects and products.

WANT TO SEE MORE?

OTHER PROJECTS