I am performing end-to-end deep learning in ArcGIS Pro 2.5, following a similar procedure to the one set out in this tutorial: https://learn.arcgis.com/en/projects/use-deep-learning-to-assess-palm-tree-health/

I am trying to train the model to recognise bats in thermal orthomosaics taken from a drone. The bats are much warmer than their background, so are detectable in the images as bright circular objects with a diameter of approximately 5 pixels, against a dark background (see image 1).

So far, I have been defining training samples by drawing a circle from the center of each bat to just outside the edge of the animal.

However, this is not giving me great results.

Is anyone able to offer advice on the best way to define a bat in my training images?

Should I exclude all of the dark area surrounding the bat?

Should i use the square training sample drawing tool instead?

enter image description here

  • If you didn't manage to solve the problem, can you provided a picture of the training (like, the bat and the circle around) so we can have a better visualization of what you are doing ? Also, how many bats are you using from the training dataset ? Jul 2 '20 at 9:33
  • Hi Caroline, thanks for your interest here! I've attached an image showing what the bats look like, as well as how i define them in training sample chips (brown circles). Jul 7 '20 at 0:36
  • You said the bats are bright objects against the dark background, but on the last image you provided, there are some long and thin part of the image that are also bright. I'm guessing that this is no longer bat ? You also said that you use thermal image. I'm not familiar with this kind of data, but aren't thermal image different shade of red/violet color ? Maybe using that instead (if you have it) would improve the result Jul 8 '20 at 6:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.