1

I am using eCognition Developer Trial 8.0. I have rgb Drone Otho foto. We want to classify this image for various biotope types exist in that region. We have already an existing classification as a thematic layer that contains at least 100 of classes.

My primary idea was to classify based on TTA mask. So that I donot need to assign it by hand. I have genererated TTA mask using this solution.

I do not have any idea if it is possible to classify the whole area/image based on TTA mask or how it works. You can take a look of my data screenshots: enter image description here

  • 1
    Yes, you can. When you load the TTA mask, classes will be added Class hierarchy with samples assigned by TTA mask. You can select them to perform a classification. Obviously, you need to perform a multi-resolution segmentation or another kind of segmentation first. Also, select object properties to extract response and so on – aldo_tapia May 9 '18 at 11:44
1

Now I got the process diagram of doing image classification using TTA Mask in eCognition.

step 1: Multiresolution segmentation to generate objects

step 2: loading TTA mask for classification:

  • Classification--> Samples--> Load TTA Mask
  • You can control viewing TTA Mask or image layer using view settings.

step 3: create samples for NN Classification:

  • Classification--> Samples--> Create Samples from TTA Mask

step 4: Select classes for NN: - Classification--> Nearest Neighbor--> Apply Standard NN to Classes--> all/some you want

step 5: process tree--> Classification-->level1--> filter(none)--> used class->all including unclassified

I hope it would also help others who have a similiar question.

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.