AI to recognise AH at bait traps?

[JamezF]

I don't think this is actually that hard to do. I've been considering putting something similar together to scare off deer from my veggie plot and I've seen one described that scares foxes and badgers out of someone's garden. All the software exists and is open source, and I think runs on a Raspberry Pi. It's just a question of plugging it all together and training it.

Could be more interesting to get it to work in the dark as well, but do AH even fly in the dark? (I know European Hornets do.)

https://www.bbc.co.uk/news/uk-england-devon-68717793

James

 

[Sutty]

I was looking at this based on the same badger & fox system. I haven't built anything yet. It can run on a few platforms. Obviously training not possible yet in the UK.

 

[roche]

I was looking at this based on the same badger & fox system. I haven't built anything yet. It can run on a few platforms. Obviously training not possible yet in the UK.

I had got as far as thinking about it but not doing any research. Could it be trained using still pictures?

 

[Sutty]

I had got as far as thinking about it but not doing any research. Could it be trained using still pictures?

That's what I was thinking for a start. Could then operate some form of actuator to direct VV into an enclosure for later eg FC treatment

 

[warsaw_hive]

https://www.pollenize.org.uk/asian-hornet-ai-detection

 

[Sutty]

https://www.pollenize.org.uk/asian-hornet-ai-detection

I thought/hoped someone would already be developing it - just surprised I haven't heard of other systems eg in mainland Europe, and trapping rather than monitoring.

 

[JamezF]

I thought/hoped someone would already be developing it - just surprised I haven't heard of other systems eg in mainland Europe, and trapping rather than monitoring.

Divert captured AH into a separate box and leave them until they start eating each other, then release the last one left?

James

 

[Sutty]

Divert captured AH into a separate box and leave them until they start eating each other, then release the last one left?

James

Divert into a separate box, then chill it, daub them all with fipronil custard & release.

 

[warsaw_hive]

I hope somebody is looking into adding AI to the Asian hornet app. That would drastically reduce the workload.

(Possibly showing my ignorance with this comment since I don't have the app downloaded)

 

[Dodge]

This is a brilliant idea and hopefully it cab be adapted easily so a Resberry Pi can be used.

 

[Jules59]

I don't think this is actually that hard to do. I've been considering putting something similar together to scare off deer from my veggie plot and I've seen one described that scares foxes and badgers out of someone's garden. All the software exists and is open source, and I think runs on a Raspberry Pi. It's just a question of plugging it all together and training it.

Could be more interesting to get it to work in the dark as well, but do AH even fly in the dark? (I know European Hornets do.)

https://www.bbc.co.uk/news/uk-england-devon-68717793

James

afaik, AHs do not fly at night unlike our european friends.

 

[Karol]

Divert into a separate box, then chill it, daub them all with fipronil custard & release.

Time to KISS things I think.

FC isn't warranted in the UK at present because velutina is not established in the UK.

IF and when the time comes to use FC it'll be the velutina visiting your hives that will need treating not those that visit bait stations.

There's no guarantee that worker velutinas visiting bait stations come from a hunting phase nest. Treating such hornets might simply be releasing FC into the environment rather than targetting the queen back at the nest.

You would not treat and release a queen. That would be absurd.

You would not want to chill worker velutinas before treatment as the calorific stress may result in them not making it back to the nest.

For now just monitor and report back to the NBU.

 

[Sutty]

Time to KISS things I think.

FC isn't warranted in the UK at present because velutina is not established in the UK.

IF and when the time comes to use FC it'll be the velutina visiting your hives that will need treating not those that visit bait stations.

There's no guarantee that worker velutinas visiting bait stations come from a hunting phase nest. Treating such hornets might simply be releasing FC into the environment rather than targetting the queen back at the nest.

You would not treat and release a queen. That would be absurd.

You would not want to chill worker velutinas before treatment as the calorific stress may result in them not making it back to the nest.

For now just monitor and report back to the NBU.

I certainly wasn't suggesting this is needed in the UK currently, and if we are lucky it may never be!
Training machine recognition will need large numbers of images of both velutina and possible bycatch species, so would need at least part of the model generation to occur in an area where vv is endemic, eg the south coast of France. However I think there is potential to create exceptionally specific traps for use in areas where they may be needed.

I'm curious what you mean by a "hunting phase nest", do they have a phase when they don't need to hunt?

 

[Karol]

I certainly wasn't suggesting this is needed in the UK currently, and if we are lucky it may never be!
Training machine recognition will need large numbers of images of both velutina and possible bycatch species, so would need at least part of the model generation to occur in an area where vv is endemic, eg the south coast of France. However I think there is potential to create exceptionally specific traps for use in areas where they may be needed.

I'm curious what you mean by a "hunting phase nest", do they have a phase when they don't need to hunt?

Once the sexual progeny have left the nest the queen will stop producing brood so the need to hunt for protein vanishes. Similary, if a queen is killed within a nest, for example as a consequence of FC treatment, the remaining workers will stop hunting once there is no more brood.

Bait stations are nevertheless essential tools for monitoring purposes because velutina has to supplement carb intake in low prey environments such as in the UK. It's just that there is ambiguity as to what may drive velutina to bait stations so FC may be less effective. Hunting around hives is far less of an ambiguous behaviour with much higher chance of trojan success. Hope that makes sense.

 

[Sutty]

Once the sexual progeny have left the nest the queen will stop producing brood so the need to hunt for protein vanishes. Similary, if a queen is killed within a nest, for example as a consequence of FC treatment, the remaining workers will stop hunting once there is no more brood.

Bait stations are nevertheless essential tools for monitoring purposes because velutina has to supplement carb intake in low prey environments such as in the UK. It's just that there is ambiguity as to what may drive velutina to bait stations so FC may be less effective. Hunting around hives is far less of an ambiguous behaviour with much higher chance of trojan success. Hope that makes sense.

Thanks, though that seems like nests are hunting except at the very end of the season, or if the queen has died. In that case it would seem reasonable, unless at very end of season, to treat them all as hunting.
Do let me/us know if this is wrong though.
As always, only in an area, or if, velutina is established as a problem!

 

[Karol]

Thanks, though that seems like nests are hunting except at the very end of the season, or if the queen has died. In that case it would seem reasonable, unless at very end of season, to treat them all as hunting.
Do let me/us know if this is wrong though.
As always, only in an area, or if, velutina is established as a problem!

Depends on what you mean by the end of the season. Nest maturation is governed by a host of variables that only vespines (sic. velutina) understand. Factors that come into play are prey density and weather. So in the UK the probability of velutina coming to a bait station during the hunting phase will be higher than say in Southern France but it is not a given. If for example the NBU treat a nest during the day, there is a high chance of solitary workers surviving that will subsequently attend bait stations where the risk from such isolated sweet feeding workers is negligible because they won't hunt and are simply subsisting until their demise.

The point I'm trying to make here is that there is an environmental risk to using FC which has to be reduced as much as possible in order to justify the use of FC on a risk to benefit ratio. The risk of using FC is related to the vector failing to return to its nest posing an environmental threat to non target species vs the benefit of killing the queen in an enclosed nest with very little risk to non target species. This risk to benefit ratio is vastly improved by targetting velutina that hunt at hives with FC. Monitoring is about calling in the NBU.

 

[Dodge]

Is it me or, given the warm weather, it is possible good news there are no Asian Hornet sighting rumours so far this week / weekend?

 

[hemo]

In a way no sightings may be a good sign though it isn't something we can take for granted. Vigilance is still needed and required.