• Font Size:
  • A
  • A
  • A

ASK THE EXPERTS

Vision Software Questions

Question Asked:

I would like to collect data on small plants in the greenhouse and correlate that to later plant phenotype in the field. This data would be use to build a model for early elimination of plants from the production process that are unlikely to have the desired phenotype. What type of system would be used to do this?

4 Answers

  • = Company Profile
  • = Answered Questions

Andrew Long - CEO
Cyth Systems, Inc.
andy [dot] long [at] cyth.com
(858) 342-6795 cell

Greg, the only systems that are capable of doing this are AI based systems. If you use a traditional machine vision approach this involves a large number iterations of software development which can become unmanageable very quickly. We are doing a lot of work in this space, if you would like to give me a call, we can talk through how you would implement something like this. We typically do this in a phased approach where we will acquire images, then train a model and then test to see if the solution is valid, before implementing a solution. Andy

Tom Bartoshesky - Member
Vichine LLC
tom [dot] bartoshesky [at] vichine.com
(734) 276-2260

Greg, Are you hoping to image a large area at one time or individual plants? We might have a device that would work for individual plants. If that is what you are looking for please contact me.

Phalguni Raut - Applications Engineer
Uniforce Sales and Engineering
phalguni [at] uniforcesales [dot] com
(510) 657-4000

Hi Greg, you could collect the data by imaging the greenhouse in multiple spectrum. You would be able to extract different data from the plants at different wavelengths. The unconventional plants can then be located by acquiring the coordinates from the images. These plants can then be correlated to their later phenotype and the algorithm could be used to decide whether they should be eliminated or kept.

Robert McConnell - President
WAY-2C Color Machine Vision
rkm [at] way2c [dot] com
(781) 641-0605

Greg, we are doing some work with flower petal age stage recognition that may be relevant. The process consists of: 1. Obtaining spectral images of all classes of interest. 2. Applying our newly patented method to determine an optimum combination of spectral bands to differentiate those classes. 3. Collecting only information from those bands to differentiate the classes. There's more information on our website, and I'll be speaking on the subject at the upcoming Vision Conference in Boston. (For more information see: http://www.way2c.com/w2upcpre.php and surrounding pages and/or contact me as how it might apply to your problem.)

Search AIA:


Browse by Products:


Browse by Company Type: