HomeCourse

BIN302 Advanced Analysis for High Throughput Phenotyping and Precision Farming

Credits (ECTS):10

Course responsible:Gareth Frank Difford

Campus / Online:Taught campus Ås

Teaching language:Engelsk

Limits of class size:20. If there are fewer than 10 students the course will not run.

Course frequency:Anually

Nominal workload:The total workload of 250 hours can be divided as follows: Lectures and practical data labs 60 hours, mini assignments 60 hours, 130 hours independent work on final report and course readings.

Teaching and exam period:This course starts in Autumn parallel. This course has teaching / evaluation in Autumn parallel.

About this course

Rapid technological developments are providing researchers and farmers alike with new digital tools. Coupled with automation, this provides an opportunity to rapidly and cheaply acquire large numbers of phenotypes required for precision farming and breeding. Skills are needed in research and industry alike to manipulate these diverse data types and extract the most useful phenotypes possible.

This course will give students the skills to manipulate, analyse and interpret the diverse data streams such as image, video, drone imagery, sensors in time series and vibrational spectroscopy on animals and plants into useful phenotypes. Furthermore, this course gives students theoretical and practical skills in machine learning for unique combinations of different data types to produce novel phenotypes as well as method comparison and validation analyses to determine the merit of newly computed phenotypes. All data-lab work will be conducted using either R-statistical or Python software.

Special practical excursions are planned to demonstrate automated phenotyping platforms such as operation of drones for field measurements of crops, industry visits to see rapid online conveyor belt measurements as well as a virtual tour of robotic milking systems.

The overall objective of the course is to develop students’ practical skills in manipulating data, programming, and analysis towards the emerging fields of precision phenotyping by computing or modelling new phenotypes which best serve a more sustainable plant and animal breeding and farming.

Learning outcome

Knowledge: Students will gain a theoretical understanding of data collection from different digital technologies and be able to evaluate the merit of new phenotypes from different digital technologies.

Skills: Students will extend their data analysis skills to include image, video, vibrational spectra and sensors in time series. Students will learn statistical analysis skills for formal method comparisons and validation studies and gain skills in combining different data types with machine learning.

Competence. Students will be able to optimize and generate new phenotypic measurements for toward precision farming and breeding.

  • Learning activities
    • Weekly lectures will cover theoretical concepts behind the course modules: Image analysis, vibrational spectroscopy, sensors in time series, machine learning and method comparisons
    • Weekly data labs will have focused data manipulation and analysis in R or Python with mini assignments afterwards
    • A final written data analysis report on the student's choice of given datasets towards the generation of a new phenotype
    • Oral examination on the students' reports and the course content
    • Visits to operate drones for in field measurements, virtual tours of new phenotyping platforms and visit to see online methods of food measurement
  • Teaching support
    Teachers will be available during the in person teaching and data labs/tutorials. Active guidance of learning activities will be provided.
  • Prerequisites

    Basic programming skills in R or Python statistical software, essentially the ability to read in data (eg .xlsx .csv .png .dat), manipulate data (eg sort data, delete missing values, subset data etc) and an understanding of linear regression .

    Statistics courses including at least one of the following: STAT 200, STAT210, DAT121, STIN300

  • Recommended prerequisites
    Quantitative genetics courses in plant, animal and aquaculture (such as HFA200, AQB270/AQB250/AQX250, HFX315) are recommended but not required.
  • Assessment method
    The examination is a combination of a written report (50%) and oral exam (50%). Grading AF

    Oral exam Grading: Letter grades Written report Grading: Letter grades
  • Examiner scheme
    Censor
  • Mandatory activity
    Data labs and weekly mini assignments. A final written report assignment.
  • Teaching hours
    Lectures: 2 hours per week. Datalab / exercises: 2 hours per week.
  • Preferential right
    The course is intended for MSc students with basic knowledge in relevant fields of plant and animal science, biology or computational / data science. Masters students will have priority, but PhD students are welcome provided we have space.
  • Admission requirements
    The course is intended for masters and PhD students with basic knowledge in relevant fields of precision farming and plant or animal phenotyping.