Relational operation in eCognition

Question

Is there anyone who has a experience with relational operation in eCognition. I'd like to ask 2 questions related to eCognition. 1. How can I classify the pixel values between 0.23 and 0.46 as a class 1, the pixel values between 0.46 to 0.64 as a class 2 etc. How can I achieve this purpose in eCognition. 2. Is it possible to use two variables for classification of one class. For example, can I use given ranges of temperature and precipitation to classify class 1. How should I build rule set for this purpose?

Thanks a lot

Answer 1

Solution for your first question:

1. Create your interested class by right clicking in the 'Class Hierarchy' window and select 'Insert class'. Give the name and select color as you want.
2. Double click on Contained >> and(min). The 'Insert expression' dialog will open.
3. Double click on feature you supposed to give a value; for example Object features >>> Layer values >>> Mean >>> NIR. The 'Membership function' window will open.
4. Initialize the fuzzy diagram; in your case, select the last one which covers the minimum to maximum and give the values in the box given.
5. Do the same from the first step for the creation of second class.
6. Instead 'Assign class' algorithm, use 'Classification' algorithm. Select class1 and class2 in the 'Active classes' field. Execute the process.

This is for your second question

1. Do the same from step one to step four; again double click on and(min) to define the next variable.

You can define different feature values to define a class. There is no limit. Even you can give thresholds by clicking 'Thresholds' in the 'Insert Expression' dialog box.

Regards, Sandhya