Home > MATLAB data modelling tool supports agricultural research at University of Melbourne

MATLAB data modelling tool supports agricultural research at University of Melbourne

Supplier News

The MATLAB data modelling software from MathWorks is being employed by students at the University of Melbourne to acquire accurate and meaningful information on crop growth, water status and quality of agricultural produce.

As the effects of climate change on Australian agriculture become more apparent, the importance of monitoring changing weather conditions and their diverse impacts will grow to paramount importance. Flexible and scalable processes for data analysis and modelling, particularly image and sensor data, are required to monitor and respond to the changing environment. It is also important to foster a new generation of scientists and engineers who possess not only the technical skills to analyse this data, but the critical thinking and innovative aptitude to turn it into more sustainable outcomes for the entire planet.

Monitoring and analysing changes of climate and weather, and their effects on agriculture for adaptation purposes, involves analysing huge volumes of sensor and image data, which can become highly complex, laborious, and expensive. However, modern data modelling tools can be easily adjusted to quickly process and analyse photographs and visual scans, even more than 2000 images at a time, which is a standard volume for small-scale climate analysis projects.

Teams of students at the University of Melbourne are using modelling software to acquire accurate and meaningful information on crop growth, water status and quality of agricultural produces. The students sought to create an automated program whereby large data sets of images could be consolidated and analysed quickly, allowing them to draw broader correlations between climatic and weather conditions and crop growth, water status, and yield and quality parameters. Those data sets included RGB images and video of plant material, scanned images and infrared thermal photographs, soil moisture and weather data, all of which were attained via short-range remote sensing and sensor technology.

Requiring modelling tools that were comprehensible and accessible to them where they could make changes according to discrete processes, rather than having to trawl through hundreds of lines of code, the students found MATLAB to be an effective platform for repeating and varying analysis patterns; its ability to scale up to larger volumes of data was also kept in mind for future extensions of the students’ pilot projects.

By developing automated image and video analysis codes, the students were able to quickly analyse extensive data sets of still images and video to understand plant physiology activity and changes across a vast area. If performed manually, this process may take as long as five minutes per image; yet by automating the process through MATLAB, students were able to process up to 2000 images in five minutes, allowing them to quickly and easily assess correlations between various environmental factors with plant physiology, including plant growth, canopy architectural parameters, leaves and fruit development, and plant water status.

The potential applications of data modelling for agriculture and climate science are immense. The ability to rapidly repurpose process code from MATLAB is already allowing the students to develop ‘mini-lab kits’, which apply similar data modelling methodologies to different areas of environmental science.

As the work conducted by the University of Melbourne proves, problem-solving at the tertiary level can have some fantastic results that resonate into real-world thinking and agricultural practice. But to deliver these benefits, the STEM skills gap that continues to persist in Australia needs to be closed, which can be done by fostering greater collaboration between academics and businesses.

Project-based learning should be incorporated into all tertiary curriculums to engage students in the investigation of science and real-world engineering problems. This form of project-based learning not only allows students to embrace technologies they will encounter in the workforce, it also assists in solving real-world problems.

Similarly, the skills base that employees require from STEM graduates should resonate more strongly within the curricula of Australian universities, resulting in more engaged students who are better equipped for the workforce. Fostering technical skills on industry-standard platforms such as MATLAB, which already benefits from a strong body of academic resources, will also provide students the groundwork on which to develop their own processes and innovations.

Written by Dr Sigfredo Fuentes, Research Fellow – University of Melbourne

Newsletter sign-up

The latest products and news delivered to your inbox