top of page

AT 319 Week 13 - More Object Based Classification

johnpkeough
May 7, 2025
2 min read

This weeks lab continues the previous week's work on object based classification. The area of interest in this lab is a county park in West Lafayette, Indiana. I started by working with the larger dataset. The smaller dataset will be involved later. The instructions for the large dataset are as follows:

  • Engage in object classification of the image where you create a minimum of 8 classes. Examples of classes might include:

    • Roofs

    • Concrete

    • Asphalt

    • Forest Deciduous (no leaves an brown)

    • Forest Coniferous (green)

    • Grass

    • Bare ground

  • Run the following two geoprocessing tools to clean up your data:

  • Add a field and generate a field calculation to determine the square meters of each land cover.

  • Generate a cartographically correct map of your classes. Include on the map a table showing the square meter coverage of each class (Figure 1).

  • Now reclassify and create two classes from the ones you have: Permeable and impermeable.

    • Add a field and calculate the square meters for each of the classes.

  • Generate another cartographically correct map of your classes. Include on the map a table showing the square meter coverage of each class (Figure 2).



Figure 1: Supervised Classification, 8 classes with areas
Figure 1: Supervised Classification, 8 classes with areas

Figure 2: Supervised classification, 2 classes with areas
Figure 2: Supervised classification, 2 classes with areas

Part 2 involves performing object based classification on a smaller area containing cracks in the pavement. The instructions are as follows:

  • Use the Crackdetection mosaic and generate 3 classes:

    • Pavement

    • Cracks

    • Vegetation

  • Run the following two operations to clean up your data:

    • The majority tool helps to fix problems

    • The boundary clean tool also helps to clean up the image

  • Add a field and generate a field calculation to determine the square meters of each land cover.

  • Generate another cartographically correct map of your classes. Include on the map a table showing the square meter coverage of each class (Figure 3).

Figure 3, crack detection with class areas
Figure 3, crack detection with class areas

Comments


Contact me

bottom of page