While doing research in Agricultural Plant Science to complete her PhD, Yingisani Chabalala currently works at the Department of Environmental Sciences, at the University of South Africa, as a Lecturer in Remote Sensing.
She is covering a project on “Integrating Copernicus Sentinel-1 and Sentinel-2 data and machine learning models, for mapping fruit tree species in heterogenous landscapes of the Limpopo Province, in South Africa”.
While doing research in Agricultural Plant Science to complete her PhD, Yingisani Chabalala currently works at the Department of Environmental Sciences, at the University of South Africa, as a Lecturer in Remote Sensing.
She is covering a project on “Integrating Copernicus Sentinel-1 and Sentinel-2 data and machine learning models, for mapping fruit tree species in heterogenous landscapes of the Limpopo Province, in South Africa”.
Yingisani has found that Sentinel-2 data in particular have been a game-changer for her research—let’s see how they can make a difference.
Born in Mavambe village on the outskirts of Malamulele town, in the Limpopo Province of South Africa, Yingisani grew up at Mabayeni (then known as Shisasi), under the previous Thulamela local Municipality, now known as Collins Chabane Local Municipality.
She completed her first degree in Environmental Sciences from the University of Venda (UNIVEN). She furthered her studies at the University of South Africa (UNISA), where she obtained a BSC Honours in Geography, specialising in Geographic Information Systems (GIS) and Remote Sensing/Earth Observation (RS/EO).
She later enrolled for a Master’s in Remote Sensing and GIS at the University of Witwatersrand (WITS), still in South Africa, where she was also working as a Senior Technician in Remote Sensing and GIS within the institution. After completing her Master's, she immediately enrolled for a Ph.D. in Agricultural Remote Sensing within the same institution (WITS), where she applies EO data for modelling horticultural crops in smallholder agriculture.
ESA: How and when did you decide you wanted to focus on Earth Observation?
Yingisani: I started developing this interest when doing my honours degree in Geography, at UNISA. Most of my assignments for the GIS module were EO-based. In this module, students were required to use Terreset software to do their assignments and through dedication, I ended up mastering the EO skills by myself, since UNISA is a long-distance learning institution.
This is where I realised that I had potential in the EO field. I then enrolled for an MSc in Remote Sensing and GIS, at WITS University. One day we had visiting lecturers who were in need of students to test the application of Copernicus Sentionel-2 data for mapping grass species in a protected environment.
My lecturer (Prof Elhadi Adam), asked if I could be interested and I accepted the offer. In my Master's research thesis, I applied Sentinel-2 data to model the distribution of grass species at Telperion Nature Reserve, in the Mpumalanga Province of South Africa. This project gave birth to my first publication, where I modelled animal movement in relation to forage quality.
ESA: How did this agriculture project come to life?
Yingisani: Growing up in a rural area, being involved in agriculture was the order of the day as it was a main source of food. So I would say this was a strategic placement because I happen to fall into this field after struggling to get a calculator to use in mathematics in grade 11. I started doing agriculture in grade 12 and went on to obtain a distinction.
My Grade 12 agricultural teacher (Mr Nemauluma), at Shithlangoma High School, encouraged me to pursue a career in agriculture. However, having been born to a traditional father, and being a girl, I knew that I was not going to study further; hence I did not apply to continue my studies. However, due to my passing rate, the teachers at my high school rallied together and called my brother, Noel Mathonsi, asking him to take me to University. Noel then spoke to my father and to my surprise, he agreed that I could go to tertiary education.
I wanted to study horticulture but unfortunately, the college that was offering the course only worked with applications. Luckily, the University of Venda allowed walk-ins, so I decided to go there. First I wanted to study Agricultural Sciences, but I could not as they required mathematics which I abandoned in Grade 11, due to lack of resources. Hence, I ended up enrolling for a BSc in Environmental Sciences, majoring in Geography and Environmental Management. This is where I was first introduced to Remote sensing and GIS. So I continued with GIS and RS until Master’s level, where I did very well on remote sensing.
ESA: How are Copernicus Sentinel data used within your PhD?
Yingisani: For my PhD, I decided to merge my two favourite fields (i.e., EO and agriculture). The advancement and free availability of EO data with high spectral, temporal and spatial resolution present possibilities to model smallholder agriculture with accuracy comparable to commercial satellites. Thus, this project explored the integrated use of Copernicus Sentinel-1 and Sentinel-2 in modelling heterogenous horticultural landscapes in a sub-tropical region. The approach has enabled the development of EO-based solutions relevant to evidence-based decision-making and innovation in the horticultural industry of South Africa.
ESA: How have the various data made a difference?
Yingisani: Conducting research in sub-tropical smallholder agriculture is challenging due to the prevalence of clouds, which prevent extracting quality fruit traits during their critical growth stages. The ability of Synthetic Aperture Radar (SAR) satellites to provide data regardless of the weather, and structural information about crops allows easy detection of fruit canopy information. This enhances the ability of the models to detect fruit tree species in heterogeneous sub-tropical landscapes that are intercropped and operating under different management strategies (crop calendar, irrigation scheduling) with high accuracy. Furthermore, smallholder crops have overlapping phenologies that are difficult to unravel using optical data and single-date images. In agriculture, optimisation is key; and that equates to the timing of optical image acquisition.
To leverage this, my research integrated ground truth data, data sampling, and fruit phenological information to model their spatial distribution phenomena using Machine Learning (ML ) and Deep Learning (DL). Furthermore, the project identified optimal temporal windows, in which fruit trees can be modelled with high accuracy.
ESA: Any last thought on your overall experience, especially as a woman seeking a STEM (Science, technology, engineering, and mathematics) career?
Yingisani: Breaking into the STEM industry as an African woman was never easy, since the field is male-dominated. I had to be firm and consistent in my interests. There were a lot of obstacles along the journey, but I conquered them by being persistent with my dreams.
It takes support for a woman to reach this level. I juggled work, and tertiary education from honours to Ph.D. studies, while growing a family. This would have never been possible without the support of my husband. I am also grateful to my supervisor, who sustained me and motivated my application to be permitted to halt my studies, while I was growing my family. Nurturing the nation is women’s nature. Thus, as women, we need to be persistent by inserting ourselves and taking up roles in these spaces in this industry.
Finally, food systems are complicated and STEM unleash the transformative power of innovating food systems, through the implementation of sustainable integrated management approaches.