After the thesis: Aspexit, a start-up project in R&D for precision agriculture

Last December, Corentin Leroux completed his #DigitAg PhD thesis. He looks back at his experience as a PhD student and talks about his post-thesis project: the creation of his company, Aspexit, which will provide support for the agricultural sector in its precision agriculture projects.

With his multidisciplinary interests in biology, mathematics and physics, Corentin opted for an agricultural engineering school. In his second year at Montpellier SupAgro, he discovered the AgroTIC course. Very quickly, precision agriculture and its innovative topics became a multidisciplinary area that interested him. He completed this specialist Masters with work experience at Telespazio Bordeaux, focusing on aerial imagery processing applied to viticulture. He then followed this with a CIFRE thesis at Smag.

I’ve always wanted to work in applied fields, it’s what really motivates me. I didn’t yet have any clear idea of a job, but I wanted to do R&D and this thesis came along at the right time.

Can you sum up your subject and your results?

During my thesis, I worked on spatial information processing in precision agriculture, especially intra-field yield data, which is a pioneering, relatively symbolic type of data in precision agriculture. One thing I particularly liked was being able to mobilise skills in computer science, statistics, geomatics and agronomy.

The original goal of my project was to set up an operational web service within SMAG to process members’ yield maps and to use them in operational agronomic models, such as the implementation of plans for basal dressing fertilisers, phosphorus and potassium.

I mainly worked on methodological developments to facilitate and automate yield data processing with, of course, underlying research questions. For example, I focused on improving the accuracy of yield maps, on the representation/mapping of data in space (zoning), or on the description/characterisation of intra-field variability.

This work raised several research questions, especially on how to remove variable noise/bias in space, how to take into account operational constraints and expertise when dividing a field into homogenous zones, etc. Since the goal was to develop an operational web service, it was important to integrate into the process the fact that methods needed to be automated and robust.

At the end of my thesis, an operational processing chain was set up to use and process yield maps. However, work is still needed to enable the use of these yield maps to fine-tune the recommendations of an operational basal dressing service.

Thinking early on about the post-thesis stage

Over those three years, I was constantly thinking about the post-thesis stage: carrying on in the world of research, was that what I wanted to do later on?

Personally, during my work experience and thesis, I was a little frustrated to have not worked enough with the final beneficiaries of the services I was developing, in other words the actors in the field: cooperatives, farmers, etc. Despite an applied subject, I did very little co-construction with those actors. In a thesis, there is always a balance between [A1] research work, where it’s important to propose a new way of looking at things, to formalise approaches, to publish, etc., and operational work. I would perhaps have liked to go further with the latter. In fact, I always thought that it was not quite right for me, that I needed to position myself differently. After this very instructive experience of research, I wanted to be able to retain this aspect, but to continue with the operational side.

I’m interested in data processing, it’s exciting, and I want it to serve a purpose. I want to work more closely with cooperatives, service companies, and actors in the field. To be able to say “What do you need? We can propose tailor-made solutions and co-construct a service”. It’s essential to understand how actors work and what’s wrong in the field.

This, along with the fact that I’ve always planned to be an entrepreneur, is why very early on I started thinking about a business project that corresponds to my expectations and values. I always wanted to try, and in my mind, there’s nothing more rewarding than setting up your own project, defending your own ideas. It’s important to step back and think about what you’ve learnt and what you want, before setting out. It’s great, I’m not taking a huge risk by trying. And my goal is not to carry on alone, I’d like to employ others, but if it doesn’t work, I’ll have learnt a great deal.

How do you go about creating your own company?

By building a network through meetings. Don’t hesitate to contact people. More people respond than I would have thought, and are happy to take time to talk. My thesis supervisors also continued their support, by reading my applications, for example.

Next, by using as many support mechanisms as possible, entrepreneurship networks, and business accelerators, such as BGE, French Tech, BIC, etc. Pôle Emploi can also help, as it’s not always obvious what’s out there, whether before, during or after setting up your company.

And also by entering competitions, which increases your visibility but also teaches you to present yourself, to clarify your services, and to draw up your plan. For example, the Doctorat à l’Entrepreneuriat en Occitanie competition, run by the Occitanie region, with the i-site MUSE and the accelerator AxLR, for which I received the people’s choice award. I also entered the Graines d’Agro competition, which gave me the opportunity to participate in workshops on how to set up a company: business plan, project management, pitch, etc.

I also began communicating early on, with personal support: my flatmate helped me to build a visual identity by creating a logo and a graphic charter, my mother took some “professional” photos, and I launched the website and an information blog.

 

 

In Latin, “Aspexit” implies searching, examining, looking. What R&D services will your start-up provide?

The idea is to accompany actors in the agricultural sector, such as cooperatives, technical institutes and service companies, in their precision agriculture projects.

In my opinion, taken alone, any data collected has little value. It needs to be placed in a particular production context, processed using high performance, made-to-measure algorithms, and must absolutely be discussed with the actors concerned: farmers, advisors, cooperatives, etc. I have a real desire to generate specific decisions and actions in the field, to work with multidisciplinary approaches in agronomy, pedology, computer science, geomatics, statistics, etc., and to build bridges between public and private actors in the agricultural sector.

My value proposition is really made-to-measure support for these actors, through data processing, monitoring, training and project management services in precision agriculture. For example: customised analysis of experimental data in micro-plots or strips; use and processing of high-resolution vegetation maps; audit of an information processing chain in precision agriculture; QGIS training adapted to the needs of the agricultural sector; increased accuracy and mapping of intra-field yield maps, etc.

I’d like to rapidly be able to take on an IT developer to manage the side involving structure, architecture and data exchange with my potential clients, and later on, if possible, statisticians and agronomists to support me in data processing and monitoring services.

Three months after completing his thesis, Corentin is a winner of the 10th edition of the Graines d’Agro competition, winning the Graine d’Excellence prize. He now needs to join a business incubator to set up his company and to approach future Aspexit clients. Should you wish to meet him, Corentin will be present at SIMA 2019.

 

 
Process, Analyse and Make value of your data
www.aspexit.com – cleroux [AT] aspexit.com

See also : PhD thesis & publications