Hives can be ‘optimized’ using data and models

Bees make important contributions to biodiversity. Pollination is actually an essential step in the life cycle of plants, and bees contribute entirely to it. However, the threats to biodiversity today weigh heavily on pollinating insects and especially honey bees (e.g.Apis melifera Living in Europe). This disappearance would be catastrophic for humans: it is estimated that about 35% of our diet depends on the pollination services provided by bees!

Colony collapse syndrome Colonial collapse disorder Or CCD, meaning systematic abandonment of hives) is a phenomenon that is frequently seen in bee colonies. A significant decline in bee colonies is thus reported worldwide; It can cause damage up to 90% of hives. Regarding honey bees, the causes can be multiple: poisonous, parasitic, viral, and even predatory, including the appearance of Asian horns in recent years (Vespa Velutina) Then from the east horn (Vespa OrientalisHexagon.

By keeping bees, beekeepers play a key role in saving the species. To help them, they now have less invasive solutions to monitor and predict the state of hives. In particular, this article highlights two of our research work, which uses the Internet of Things technology. The thing is the internet Or IoT) associated with artificial intelligence, computer models and simulations, to assist beekeepers in their business practice.

Bee behavior at the entrance to the hive

The first key to monitoring and predicting bee health is modeling bee behavior; This makes it possible to identify the strengths or weaknesses of bees caused by disease, famine or predators. In this way we can understand what is happening inside the hive by observing what is happening outside. For this, images and videos are rich sources of information that are usable, non-destructive, and which constitute a powerful scientific and environmental challenge. Nevertheless, trajectory modeling has not yet developed, which inspired our work.

Our first task focuses on identifying bees, as well as modeling their trajectory. The idea was to capture the bee’s flight paths individually, then mark the rhythm of the overall activity in front of the beehive so that the observations could be combined and made available to scientists for cross-reference with beekeeping data. (Famine, predators, drought …).

In practice, this involves capturing the movement of bees with a stationary, non-invasive camera, filming the entry and exit of insects. Using a high-resolution sensor and high acquisition frequency, image processing techniques make it possible to isolate bees from the background.

Figure 1: The center of the body (green dot) and the displacement of the two bees (red, blue, purple and yellow line), calculated from different images – Gregory Zacarevich and Baptist Magnier

Next, the bee counters are removed from each frame of the video, so that the center of each bee can be identified (the green dots on the bee Figure 1) Then, the orientation of each bee is represented by an ellipse (related to the shape of a bee). The orientation and size of the ellipses of the image makes it possible to calculate the bee’s movement between different images in a video (blue and red lines for this. Bee 1 And for the yellow and purple lines Bees 2)

Figure 2: Plot of asymmetric (left) and consistent (right) trajectory
Figure 2: Asymmetrical (left) and consistent (right) trajectory plots – Gregory Zacarevich and Baptist Magnier

In fact, it’s easier to follow these objects of constant size and similar orientation in videos than distorted objects. This technique avoids confusing bees and calculates confusing pathways, e.g. Figure 2.

Various traces are thus recorded. The Picture 3 The result of a video with 1755 pictures. It shows the direction of entry of the bee hive (green line), leaves (red), or simply passes through the front (blue). Wrongly marked trajectories are also shown in blue. From this information, it will then be possible to study and classify the behavior of bees.

Figure 3: Observation of the movement of bees in front of the hive, based on 1,755 images.  Green Line: Bees enter the hive;  Red line: Bees are leaving the hive;  Blue line: Bees are passing in front of the hive and poorly marked trajectory
Figure 3: Observation of the movement of bees in front of the hive, based on 1,755 images. Green Line: Bees enter the hive; Red line: Bees are leaving the hive; Blue Line: Bees pass in front of bees and poorly marked trajectories – Gregory Zacarevich and Baptist Magnier

In the future, bee behavior could be further explained by supplementing studies with machine data learning and a semi-supervised AI methodology.

Physical characteristics of bees

The second key decision support for beekeepers is the analysis of the internal health of the bee. Our team work here uses scales connected to data from different sensors (such as relative humidity and internal temperature) to analyze the evolution of the weight of each hive, as well as videos of activity on the take-off board (as shown above).

The BeePMN project – led by our team in partnership with USEK in Lebanon, ConnectHive in France and l’Atelier du miel in Lebanon – contributes to the physical properties of the bee kingdom using apiaries data from the shared database. We have proposed a method based on the recognition of characteristic patterns in the weight data recorded by the hive scale. One reason may be weight gain, then a plateau followed by weight loss which will coincide with the departure of the bees (for example by swarms).

Subsequently, the collected data was evaluated and processed using algorithms, which made it possible to discover repetitive patterns related to the events occurring in the hive.

Connected bee keeping

Combined with other information, these two examples can eventually be combined into a common beekeeping monitoring system. Computer model (presented in Figure 4) Is triggered automatically by a series of predefined warnings, which invite the beekeeper to take action. For example, a beekeeper’s information on the need to feed bees may be triggered by a passage of weight below a reference value within a certain period of time in autumn or winter.

Specifically, since beekeeping activities cannot be automated and human intervention is mandatory, the proposed system will help and guide beekeepers to plan and perform a number of more relevant tasks more specifically: breeding and breeding new colonies, feeding colonies, adding super ( The upper part (honeycomb for honey collection), planning of sanitation operations, pest control (e.g. verro), planning of operations like hibernation etc.
The beekeeper will thus be able to easily observe its colony, perform its routine tasks, respond to warnings of possible errors, and predict its future supply needs.

Figure 4: Hives, sensors, AI and smartphones / tablets form a common monitoring system.
Figure 4: Hives, sensors, AI and smartphones / tablets form a common tracking system – Gregory Zacarevich and Baptist Magnier

These models are based on business rules created with the help of domain experts. It has also been speculated that these business rules may later be developed using the system thanks to the contribution of the beekeeping community.

Finally, based on the principles of gamefiction, all of the above will be orchestrated and presented in a user-friendly interface on a smartphone or tablet.

This contribution will improve the experience of amateur and professional beekeepers, reduce the risk of operating an apiary, and open the door to other available inputs (detailed weather events, flower maps, humidity, bee colony behavior, etc.) to increase efficiency. Simulation. Finally, beekeeping should be paved with more precision to reduce the contribution of recent generation digital techniques, and especially simulation, invasive and synthetic treatments.

This analysis was written by Gregory Jakarevich and Baptist Magnier, both professors at the Ales Institute of Mines-Telecom.
The original article was published on the site Conversation.

Declaration of interest
● Gregory Zacarevich receives funding from Campus France for IMT Minus Aless

Leave a Comment