Object Based Classification Part 1

Introduction

Classifying objects is a great way of being able to identify different types of objects based on their properties. One such example of being able to do this is in regards to classifying pervious and impervious surfaces. Pervious surfaces allow for the absorption of water. Many of these types of surfaces are nature-related such as that of grass and trees. These surfaces can be labeled as pervious because they absorb water. However, the opposite of being pervious is being impervious, which means objects cannot absorb water. Some examples of this are roofing on houses as well as other man-made objects such as cars. These surfaces do not absorb water hence making them impervious.

Methods

For this lab, I was tasked with going through an ESRI tutorial on object-based classification, specifically that of identifying pervious and impervious surfaces. After that, I needed to re-create the tutorial as a way of helping guide myself through future lab assignments where I would be tasked with other classification activities. Through the usage of ARC-GIS pro, I was able to conduct the proper processing steps to complete object-based classification and gather the correct results.

Deliverables and Discussion

Segmenting the Imagery: This allows for like pixels of an image to be grouped together. It is also the first step necessary when doing object based classification.

Extract Spectral Bands: Allows for the reordering of your bands so that the color combination you use matches the type of information you are wanting to display (Example: Using Infrared to show negation growth). This is the second step in object based classification.

Figure 1: Completion of the Spectral Band Extract.

Classification Wizard Configuration: This is where you create your different types of classifications and label them throughout the picture.

Figure 2: Each color of the shape corresponds to a different classification class that you created in the classification wizard.

In addition from here you are also able to combine these classes into sub classes such as that of pervious and impervious.

Figure 3: After running the classification wizard based off of your created classes.

Output Classifier: This is when we are going to take all of our classes that we created and then label them as either pervious and impervious. Once you run this you will be able to create a map like the one below that shows which surfaces are pervious and which are not.

Figure 4: Map identifying pervious and impervious surfaces.