First import the image you want to run OBIA tool on. The window will look like :
Now when you click on the OBIA->Read_data option.
Close the initial window and browse for the data directory ("OA_data" in the repo).
Now click on "Image Segmentation" after changing the parameters (spatial bandwidth, range bandwidth and minimum region area) as you like.
It will save the segmented image in the temp folder by name seg_OpticksImage.ppm. Now click on Object Attribution, it will make the assign the features to the objects made in image segmentation.
Now we have to provide training samples for classification.
Enter the number of classes you want and click get training samples.The input image will be displayed, click at any point on the image and save the sample to its class.
Push as many training sample as you want. Then click on classify. It will do mahalanobis distance classification to the rest of the objects and display a pruning window.
Change the attribute values and see the output in "generate Image".
Image displaying all the objects of class 1 an in area between 1 and 100 units.
Image displaying all the objects of class 1 an in area between 200 and 300 units.
Play around with all the attributes and see the outputs.
Now to export all the objects to a shape file click on "Generate Shape file" button in the main window. It will ask you to provide the name of the output file(which should be unique).
Now the output file will be generated in the data folder that you mentioned.
This is the output file opened in QGIS.
Looking at the attribute table.