10. In Raster Object Operators (ROO), add three operaors- Size Filter, Dilate and Probality Filter.
The Size Filter
.. filters out raster objects that are too small, too large, or both. It allows you to restrict the pool of raster objects to those of an appropriate size.
I have set the Minimum Object Size to 100.
..expands raster objects. It considers all pixels near the raster object. If any pixel of the raster object is adjacent (horizontally, vertically or diagonally) to the pixel being considered, it is added to the raster object.
No change here.
The Probability Filter
.. removes all raster objects whose zonal probability mean is less than the specified Minimum Probability.
I set the Minimum Probability to 0.70 or 70%.
11. At Raster to Vector Conversion (RVC) select Polygon trace.
12. Like before we will add three operators at Vector Object Operators. These are
This operator removes all islands or holes of an object.
Set Max Object Size. Mine is 100000000.00. Remember to hit enter whenever you change a value.
Generalize, with tolerance 0.30 m.
This operator simplifies polylines and polygons by removing unnecessary vertices. It helps remove noise and improves the performance of downstream processing.
Smooth, with only 20 as Smooth factor.
This operator smooths polygon objects in a vector layer.
13. We are almost finishing. At Vector Object Processor (VOP), select Geometry:Area and Geometry:Compactness.
.. computes the area of each polygon shape in the input shapefile. The area of islands within the other polygon is subtracted from the outer polygon’s area. The output shapefile will have an Area attribute containing the result of this metric.
..measures the compactness of an object. The result is based on the ratio of the perimeter of a circle with area equal to that of the object and the perimeter of the object. The maximum result is 1 for a circle and the minimum result approaches 0 for thin and complex objects.
Select Area subsection and go to Distribution tab. There are Min, Max, Mean and Standard Deviation area you have to fill. To know about the pixel statistics we will use the Measure Point, Line and Area tool at the menu bar. Hit it! Select Main View as data source.
There is a new window. We will be needing the area in sq. meter, so change the unit as per. Now draw an area around your subjected feature to know about its area in sq. meter. You’ll need to study the average size of the feature, what could be the minimum and maximum area size, and ofcource the mean and SD. See how I did it.
You’ll need to add the same at Compactness properties.
14. The Vector Cleanup Operators is the last stage. We add two operators,
Island Filter, with same property as previous one,
Probability Filter as the same as before. This is my whole tree.
You finished with adding filters. Now see how the objective quickly finds out your desired feature in a flash. Hit the red bolt sign at the menu bar and see!
15. What if the Run button do not appear to be active, there must be something you forgot to mention earlier.
Hit the bolt with ? mark, “Why Isn’t The FEature Model Ready?” button. You’ll meet a window describing the problem, just follow the instruction.
I hope your Model is clean enough to perform the task. The Process List will show you the detailed processing while Erdas performing the classification.
THAT’S IT! This is your classified image. If you want something more go play with the numbers and sliders of the operators. You can also add more operators from the list or can delete them as well. Just right click at any operator and you will see the options. You can also halt the whole process selecting Stop Here, and see whats the particular operator do. You can have the output anytime from the I/O tab of the property area.
This is my result.