Welcome, Guest

Author Topic: Mobile app for beekeepers - Varroa Detection  (Read 1537 times)

Offline Bartek_ziBees

  • New Bee
  • *
  • Posts: 4
  • Gender: Male
    • ziBees
Mobile app for beekeepers - Varroa Detection
« on: March 29, 2018, 08:27:09 am »
Hi,
I want to introduce you to ziBees. A free application for a phone that uses artificial intelligence to improve the process of managing an apiary.
We analyse the pictures of the bee frame and detect the number of bees in the picture and the number of bees that have varroasis. In the future, we want to detect other important for beekeepers elements.
We want to know your opinion about the application. Your involvement can contribute to the introduction of new functionalities in the future.

In April we are starting with closed beta tests. We want to test the application and collect photos of bee frames, to constantly improve the algorithm.

You can sign up for beta tests through our website www.zibees.co or directly via the link: http://www.smartsurvey.co.uk/s/zibees/

Offline Robo

  • Technical
  • Administrator
  • Galactic Bee
  • *******
  • Posts: 6739
  • Gender: Male
  • Beekeep On!
    • Bushkill Bee Vac
Re: Mobile app for beekeepers - Varroa Detection
« Reply #1 on: March 29, 2018, 08:21:16 pm »
How do you account for varroa on the underbelly of the bees?
"Opportunity is missed by most people because it comes dressed in overalls and looks like work." - Thomas Edison



Offline Acebird

  • Galactic Bee
  • ******
  • Posts: 6097
  • Gender: Male
  • Practicing non intervention beekeeping
Re: Mobile app for beekeepers - Varroa Detection
« Reply #2 on: March 29, 2018, 09:34:42 pm »
Sounds too good to be true.  A hand held phone could have limitless angles and depth or focal distance.  Detecting varroa with all these variables would be some trick.  How does one validate it's accuracy?
Brian Cardinal
Just do it

Offline moebees

  • House Bee
  • **
  • Posts: 193
  • Gender: Male
Re: Mobile app for beekeepers - Varroa Detection
« Reply #3 on: March 29, 2018, 10:55:25 pm »
I'm skeptical.
Bee-keeping is like raising Martians  - Isabella Rosselini

Offline sawdstmakr

  • Global Moderator
  • Universal Bee
  • *******
  • Posts: 11386
  • Gender: Male
Re: Mobile app for beekeepers - Varroa Detection
« Reply #4 on: March 30, 2018, 07:45:39 am »
Since a computer can find the face of one person in a crowd of hundreds I suspect it can detect the mites that are visible on the bees in a picture. Sounds possible.
Jim

Offline Bartek_ziBees

  • New Bee
  • *
  • Posts: 4
  • Gender: Male
    • ziBees
Re: Mobile app for beekeepers - Varroa Detection
« Reply #5 on: March 30, 2018, 07:56:26 am »
Hi,

thanks for the questions!

We are aware that from picture analysis we can't count all mites. We have plans to perform research in a cooperation with one of the Universities in Poland. We want to compare how many mites we can find by using an app and then compare it to other diagnostic methods. We want to create an algorithm which will show more accurately how many mites are on the frame.

The question is: amount of varroa mites is important for You? Do You want to track the infestation level?


Offline little john

  • Queen Bee
  • ****
  • Posts: 1483
Re: Mobile app for beekeepers - Varroa Detection
« Reply #6 on: March 31, 2018, 08:07:43 am »
Of course it would highly desirable to know the level of infestation - but duff information would be worse than having no information at all, because this would pose the risk of generating false levels of confidence. 

As Robo has already commented - how would you go about detecting those phoretic mites which have inserted themselves between the lower abdominal plates - a choice location for puncturing the soft tissue which these protect ?

I would suggest that arithmetically extrapolating solely from the numbers of mites detected on the upper surface would be so much guesswork - but - it will be interesting to see what you come up with nevertheless.
LJ 
A Heretics Guide to Beekeeping - http://heretics-guide.atwebpages.com

Offline Bartek_ziBees

  • New Bee
  • *
  • Posts: 4
  • Gender: Male
    • ziBees
Re: Mobile app for beekeepers - Varroa Detection
« Reply #7 on: April 03, 2018, 09:17:22 am »
Little John - Thank You for Your comments. We will make everything we can to help a beekeeper do his tasks faster and smarter. What are in Your opinion other important features You would wish to have and will be very useful?

Offline Acebird

  • Galactic Bee
  • ******
  • Posts: 6097
  • Gender: Male
  • Practicing non intervention beekeeping
Re: Mobile app for beekeepers - Varroa Detection
« Reply #8 on: April 03, 2018, 01:59:57 pm »
Since a computer can find the face of one person in a crowd of hundreds I suspect it can detect the mites that are visible on the bees in a picture.

Nothing is as reliable as CSI I am afraid.
Brian Cardinal
Just do it

Offline Robo

  • Technical
  • Administrator
  • Galactic Bee
  • *******
  • Posts: 6739
  • Gender: Male
  • Beekeep On!
    • Bushkill Bee Vac
Re: Mobile app for beekeepers - Varroa Detection
« Reply #9 on: April 03, 2018, 02:53:08 pm »
I think this will be a great tool for the masses.  I know many new beekeepers in my group that will benefit from it.   Regardless of the accuracy concerns,  it will make it simple for the average beekeeper to get an indication of varroa infestation.  Yes there are better methods to get an accurate count,  but there are many many beekeepers not doing accurate measurements now because of the effort. 

Will you provide analytics to get a hive perspective from a series of frames?   How about DWV bees?

I signed up for the beta, is there a timeframe that this will happen?
"Opportunity is missed by most people because it comes dressed in overalls and looks like work." - Thomas Edison



Offline Bartek_ziBees

  • New Bee
  • *
  • Posts: 4
  • Gender: Male
    • ziBees
Re: Mobile app for beekeepers - Varroa Detection
« Reply #10 on: April 05, 2018, 03:15:55 am »
Thanks Robo!

Yes, we want to link frame analysis with specific hives. We start beta test in April (for Android users) so Thank You that You have signed up.


Offline hydrocynus

  • New Bee
  • *
  • Posts: 15
  • Gender: Male
    • Beekeepers Association of Southwest Florida
Re: Mobile app for beekeepers - Varroa Detection
« Reply #11 on: December 16, 2018, 01:45:40 am »
Hello, I am a Ph.D. researcher and I have undergrads conducting bee research. We could test your system if the university of Poland has not tested it yet. Let me know. I might even get funding from my university.

Online TheHoneyPump

  • Field Bee
  • ***
  • Posts: 938
  • Work Hard. Play Harder.
Re: Mobile app for beekeepers - Varroa Detection
« Reply #12 on: December 17, 2018, 05:42:06 am »
To be honest, I am skeptical. Here are my reasons:

By the time the mites can be seen on the backs of the bees in order to be able to count them, the infection level would be so high it is not worth bothering to count them.

Do not quote these numbers as specification as I have read so much literature, observing that the numbers vary between articles and am fuzzy as to which numbers come from which articles, nor verified any validity of them. That said, if nothing else one can bring them all together and get a ball-parking range of such values which should help get you thinking pragmatically about this.
 - 95% of mites on the bees (phoretic or parasitic is current debate) will be on the underside of the abdomen.  Your eyes, your phone, a camera, the AI app will not see those when looking at a frame of bees.
 - depending on the time of year; typically 15% - 20% of the mites will be on the bees. The rest, 80%-85%, of the mites in a hive are hidden beyond view in the combs under the brood caps.
 - For one mite to be seen on the back of a bee, there are likely another 2 to 3 more mites on the underside of that same bee which cannot be seen.
...... doing really simple math on those %ages, the mites that can be seen on the backs of the bees by the eye or a camera being analyzed by AI app represents less than 0.75% of the mites that are present in the hive.  This is why I say that by the time you are seeing mites on the backs of the bees, the situation is so far advanced and dire that counting them is pointless. The infestation is so high that the bee and the colony are already dead. They just do not know it yet.

I appreciate the thought, the concept, and the project as being interesting from a R&D development exercise perspective. However, to be pragmatic about it I fail to see any practical use the app as proposed will provide to the beekeeper. Certainly agreed, varroa is seriously a bad thing and we all really want more tools to deal with them.  But you are not going to discover nor fix a varroa problem looking at a picture of bees.

As myself being an innovator, I know it takes at least 10 bad ideas to discover 1 good idea.  To be frank, the varroa counting app by phone camera and AI idea needs to go into the bad idea box.  Not because of technical feasibility, but because there is no practical usefulness of it after one considers the points listed above.

All that said. Please do not be discouraged by my comments. Come up with at least 9 more ideas to discover a good one worth pursuing ...  For example; instead of looking for varroa such an app could be developed and used to detect other visually apparent ailments such as;  deformed wing virus in static photos, acute/chronic bee paralysis virus in short videos of quivering bees on the frame or the tell-tale hairless black shiny bees in still photos, sneaky hive beetles, worms, excessive dysentry on the frames, and/or brood indications such as AFB, chalk brood, sac brood, brood pattern quality and density, ... other ... 

Therefore, my recommendation is to dump the varroa counting idea and instead steer this project towards something useful by focusing on those many other colony ailments that have clear indications which are easy to identify upon seeing them but which are too often obscured from the untrained/inexperienced eye looking at a frame full of bees and their stuffs.

Hope that helps ... In some way.
« Last Edit: December 17, 2018, 02:42:03 pm by TheHoneyPump »
Bees Buzz Bees Buzz Bees Buzz Bees Buzz Bees Buzz Bees Buzz Bees Buzz Bees Buzz Bees Buzz Bees Buzz Bees Buzz Bees Buzz Bees Buzz Bees Buzz Bees Buzz Bees Buzz Bees Buzz Bees Buzz Bees Buzz Bees Buzz

Online Ben Framed

  • Galactic Bee
  • ******
  • Posts: 4924
  • Mississippi Zone 7
Re: Mobile app for beekeepers - Varroa Detection
« Reply #13 on: December 21, 2018, 10:29:22 pm »
I have to agree with TheHoneyPump. The following video was placed here a while back by Live Oak and after watching it, (more than once), I am thinking it to be accurate. Perhaps some of you more experienced bee folks will think differently. If so please chime in.  But as I stated, I believe this to be accurate.
Phillip Hall "Ben Framed"

« Last Edit: December 21, 2018, 11:12:40 pm by Ben Framed »
For this people's heart is waxed gross, and their ears are dull of hearing, and their eyes they have closed; lest at any time they should see with their eyes and hear with their ears, and should understand with their heart, and should be converted, and I should heal them.

Offline Acebird

  • Galactic Bee
  • ******
  • Posts: 6097
  • Gender: Male
  • Practicing non intervention beekeeping
Re: Mobile app for beekeepers - Varroa Detection
« Reply #14 on: December 22, 2018, 09:09:48 am »
However, to be pragmatic about it I fail to see any practical use the app as proposed will provide to the beekeeper.

My suggestion would be to only look at returning foragers that have mites on board as a trigger to look further manually.  One would want to know if a healthy hive is robbing out a mite bomb or in contact with mites hitching rides from other foragers.
Brian Cardinal
Just do it

Offline Live Oak

  • House Bee
  • **
  • Posts: 97
  • Gender: Male
    • Tractor Farm and Family
Re: Mobile app for beekeepers - Varroa Detection
« Reply #15 on: December 25, 2018, 01:41:05 pm »
Hi,
I want to introduce you to ziBees. A free application for a phone that uses artificial intelligence to improve the process of managing an apiary.
We analyse the pictures of the bee frame and detect the number of bees in the picture and the number of bees that have varroasis. In the future, we want to detect other important for beekeepers elements.
We want to know your opinion about the application. Your involvement can contribute to the introduction of new functionalities in the future.

In April we are starting with closed beta tests. We want to test the application and collect photos of bee frames, to constantly improve the algorithm.

You can sign up for beta tests through our website www.zibees.co or directly via the link: http://www.smartsurvey.co.uk/s/zibees/

I would think it may be more productive and accurate for your software algorithm to incorporate looking for evidence of varroa mite infestation in the comb.  The tell tale small white flecks of varroa mite fecal material would be a good starting point.  The algorithm would not likely spot many attached/phortetic varroa mites since the vast majority of varroa mites attach to the honey bee's abdomen.  Another tell tale symptom would be small tears or openings in the brood comb cappings the nurse bees make to check a suspected brood cell for varroa. 

Adding these features may help make your software package more effective but for most fairly experienced beekeepers, these symptoms are not difficult to spot by eye ball frame inspection.  You have an interesting concept though.  Have you discussed your concept with Randy Oliver?  Randy I am sure could probably provide you with the best feedback and suggestions you could possibly get.