using ecognition software. Regarding the use a segmentation process; I am seeking some Ruleset for change detection analysis, any idea ?
Dear Mikwa,
you can join the eCognition Community. There are some stored rule set and I saw a good one dealing with change detection. It processes on three maps to calculate the change.
Regards,
Benjamin
Hi Mikwa
I found this information see if it can of any help to you "Rule classification is the process of taking a segmented image and grouping similar pixel clusters into groups called classes. A class contains one or more rules that you can build based on your knowledge of certain features. Each rule contains one or more attributes such as area, length, or texture, which you constrain to a specific range of values. For example, we know that roads are elongated, some buildings approximate a rectangular shape, and trees are highly textured compared to grass
Build New Rules
Each class will contain one or more rules (which contain a combination of various attributes) that best define the class. For example, if you want to extract roads from an image, you could build a rule that will use a combination of Elongation, Length, and Area spatial attributes:
The rule logic for this class is as follows:
To meet the conditions of the Roads class, either one of the two rules must apply (i.e., the OR operation). The first rule has only one attribute definition. To satisfy the conditions of the second rule, both the Length and Area attribute definitions must apply (i.e., the AND operator). Each rule is equally weighted, which works well in most cases.
However, if you discover an attribute that effectively identifies the feature of interest, you could override the default weight and promote that rule above other rules. See Rule Classification Background for details.
If you are building a rule set for the first time, the Rule Classification panel appears with an empty folder called All Classes. This folder will contain all of the feature types (classes) that you define.
Click the Add Class button. A new class appears with one undefined rule and one default attribute.
Select the new class, and edit the name and color as needed within the Class Properties table. The Class Threshold property is an advanced setting; the default value is acceptable in most cases; see Rule Classification Background for details.
Select the new rule name that was created for you (Rule [1.0]), and add attributes to the rule as described in Select Attributes.
To add another rule to a class, select the class name and click the Add Rule to Class button. An OR logical operation is used to combine multiple rules within a class. To remove a rule from a given class, select the rule name and click the Remove Rule button.
To add a new class, select All Classes and click the Add Class button.
To save your rule set to disk (as a .rul file), click the Save Rule Set button. Click the Restore Rule Set File button to later restore the file.
"
- Badges
- Science topic
- Similar topics
- Analytical Chemistry
- Extraction