FlowerPower4

Project description

As a next step is also the realization of a software system that integrates the existing AI models, stores the data in the cloud, is able to generate georeferenced maps from stitched images, and a UI presenting statistics. We also need to integrate a labelling tool (right now we use the open source labelimg tool), a module to train the AI and that monitors the performance of the model. This toolbox should support researchers in the first place (and in the near future also other stakeholders, such as provinces and citizens).

Context

Biodiversity (both flora and fauna) has decreased dramatically, especially over the last two decades. The research group “AI & Big Data” of FHICT wants to contribute – in line with the SDG program – to innovations that foster “people and planet”. Among other things, we co-create innovations and prototypes for monitoring biodiversity at a large scale by means of ICT-enabled solutions. One of these initiatives is the FlowerPower project. Having a standardized method to estimate biodiversity helps to evaluate the impact of both urban and rural landscape interventions and conservation efforts.

Results

For the final system, we succesfully deployed eight services inside of a virtual environment.
We created our own AI model with a 80% accuracy. This model is completely unique for our system which makes it
really optimised. Our system also integrates this AI alongside with a lot of security. In the span of just six months we created a
secure and safe environment for researchers and citizens to enjoy.

About the project group

We are 10 people in our group from 6th semester Software engineering(Arvin, Luuk, Mattheo, Matthijs, Jelle, Antonio, Wondimu, Moni, Fadi, Colin). We have been working in Scrum methadology. We had standup meetings every Tuesday and Friday at the university where we were supposed to be. We divided the project in smaller parts so everyone had something to do. We think that we have one satisfactory end result.

Video