Honeybee conservation with Python
a talk by Thiago da Silva Alves

In the apidologie research area, there is one task that obliges the researcher to classify and count the contents of each comb cell in each frame. With this task is possible to control the progression of brood, bees, and food reserves. Since each frame can have thousands of cells, in most cases this task is done by a human in an approximate way, making it error-prone. The automation of this process, using image analysis represents an evolution in this field.

The honey bee is the world’s most important pollinator of food crops. Almost one-third of the food that we consume each day relies on pollination done mainly by bees. So the creation of software that helps the preservation of this species has a direct impact on our lives.

I am going to show you a few challenges we have faced, from creating comb cell detectors using OpenCV and Shapely, to developing models based on Deep Learning to classify the cell's content using the Caffe framework. With these models, we have obtained accuracy above 98% within eight different classes and solved the proposed problem.

This talk is aimed at advanced Pythonistas. (While it might be interesting for beginners we recommend them to choose another talk.)

Thiago da Silva Alves

I like academic environments, so I am currently studying Computer Science at UTFPR and Masters in Information Systems at the Polytechnic Institute of Bragança - Portugal.

Among my main interests are artificial intelligence, bioinformatics and any other area with difficult problems to solve computationally.

I am currently doing research with CIMO, where I use python to create tools that help researchers and beekeepers to preserve bees and increase their productivity. In my free time I also contribute with an AI group that I recently helped to create in the institution.