Skip to content

bayer-science-for-a-better-life/beehive

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Beehive

This is a side-project I was working on a few years ago. I created an algorithm that can take pictures of honeycombs and find out how many cells of each type are on it: small/medium/big larvae, capped/empty cells, cells with eggs/nectar/pollen.

I splitted the task into 2 problems. First, segment the whole honeycomb picture into small pictures of each cell. Then classify each cell based on its small picture.

  • segmentation a segmentation algorithm using classic computer vision
  • classification a convolutional neural net for classification

My results are so so: >80% accuracy in both detecting cells and classifying them. However, if you are reading this now I would recommend DeepBee. They essentially did the same, but their results are much better. If their code would have been public earlier, I probably would have just used DeepBee.

segmentation

classification confusion matrix

About

Side project for helping beekeepers. Segment and classify honeycomb pictures to find out how many larvae, eggs, nectar, etc are in the brood.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published