Alleviating world hunger by turning to Nature - Team 158

Eradication of poverty in all its forms is one of the SDGs. One of the most striking indicators of poverty is hunger, which we hope to partially address in this project's prototype.

Comment

Inspiration

With the push towards sustainable lifestyles and climate action, many have seen the protection of forests and conservation of nature as something that cannot benefit us in the present. However, a number of people in emerging economies go hungry due to crop failure, and other disasters, natural or otherwise.

Instead of going hungry yet people are surrounded by greenery, it is best to investigate the crop that is in the vicinity, and establish whether or not they are poisonous, first of all, and if they are edible, how the preparation of the same can be effected.

Such a task would require immense investment in both equipment (in terms of servers that host the catalog of all known plants and their nuances, as well as enough juice to train industry-scale models on identification of plants from just images) and human resource (the domain knowledge). IBM's Z systems have the perfect mesh of the above

What it does

The presented prototype allows users to capture images with their mobile devices, or they can upload them. It then runs a classification model to determine the identity of the plant, before looking into the database for the plant's qualities. Whether or not it is poisonous, if it is, what is the antidote to exposure, and if it is edible, how to prepare it for consumption.

How we built it

The model was built using tensorflow, and the application was built using Python's Flask microframework. IBM's Linux One OS was used to both train the model, as well as serve as the central collaboration point of the team members, as we are in different locations.

Challenges we ran into

Internet connectivity. We had intermittent connections, which resulted in our ssh sessions being disrupted quite a number of times. Furthermore, there were some hitches at the start of the hackathon, which meant that we had to start late.

Accomplishments that we're proud of

Finally getting an opportunity to work on one of the projects I always wanted to do, but never seemed to get round to it. Learned how to capture an image using a phone camera, without any complicated code. Just html

What we learned

Never give up

What's next for Alleviating world hunger by turning to Nature

Refining the product, definitely, and building a much more elaborate data store of plants and their various applications. Perhaps going green can be directly beneficial to us too, in the here and now

Built With

Share this project:

Updates