Beescanning
| Reference number | |
| Coordinator | Fribi Holding AB |
| Funding from Vinnova | SEK 300 000 |
| Project duration | May 2018 - March 2019 |
| Status | Completed |
| Venture | Innovative Startups |
Purpose and goal
Goal: "develop the functionality of the Beescanning app" has been achieved. The app analyzes indicators in a colony of bees for health diagnosis with proprietary artificial intelligence, AI. It has high accuracy, delivering results and advice on appropriate action in a matter of seconds. The app´s performance, features and collection of metadata (race, time, location) has been developed. The ”Tagger” database has evolved into a hub for continuous AI development. Video sequences have begun to be collected. Initial planning for analyzing video in real time has started.
Expected results and effects
The most important results are that the app now works as it should, that the data set was increased from 4000 to over 10000 and thus the quality in AI, and that infrastructure for step 2.0 is in place in the app, databases and know-how. And in the longer perspective, we now have a tool to assess the health of the bee population and to find bees that are healthy without chemical treatment. The purpose of further developing the technology with analyzes of video and sound has begun. Video has been collected in 2018, collaboration for AI analyzes of acoustic signals has begun.
Planned approach and implementation
Beescanning builds the world´s largest database for images of bees http://tagger.beescanning.com. for training AI. Now more than 10,000 items. Grows through automated transfer when users take pictures in the app. Video sequences have begun to gather in field 2018. For sound analyzes, collaboration has been initiated. ”Tagger” is becoming a platform for collaboration between different actors for new functions, in research, education, pollination industry. The potential of the app has been expanded to become a universal tool for increased knowledge of bees and beekeeping.