Object Based (Oriented) Image Analysis

[Norman_Nash]

I am currently using ArcGIS 10.3 and new to OBIA.  My goal is to segment out portions of dead trees in the Lake Tahoe Basin and not sure how to go about this.  I have searched extensively online and had no luck as to any tutorials or steps to follow.  I am using Landsat 8 OLI (30m) and WorldView-2 (2m) spatial resolution imagery for this task.  Here is my plan and please correct me if I am wrong. (For Landsat 8 OLI)

1.  Create a Natural Color Image using Bands 4,3,2.

2.  Establish training areas of dead trees and healthy lodgepole and red attack trees in the training sample manager.

3.  Create a signature file

4.  Under ArcToolbox>Spatial Analyst Tools>Segmentation and Classification>Classify Raster

Are these steps suitable to segment the landsat image?

Any help would be greatly appreciated

Norm

[TravisSaladino]

Here is a good link to a blog on using the Segment Mean Shift tool...  Pass the classification but hold the salt and pepper! | ArcGIS Blog

[MarkRomero]

Hi Norman,

I wrote this demo/tutorial with the workflow and steps to segment your imagery. Its a draft, but it should get you through the necessary steps. It was written for 10.3. There have been some improvements to 10.3.1 with additional tools to collect training samples.

Mark

[KarlZimmer]

Thanks for the demo. Just wondering if you run this on the raw DN values of the multispectral imagery or did you convert it to Top of Atmosphere reflectance prior to the segmentation/classification? Thanks. 

[ArthurCrawford]

Hello Norman,

You might want to identify where the trees are first.   You could use LiDAR from Data | OpenTopography ,​, imagery and model I wrote  http://www.arcgis.com/home/item.html?id=d280e1c023084992b98088d9f14d3252 to pick out the trees as points.   Once you have the points, use the buffer for radius as polygons.   Clip the imagery with the polygons and then use this for your image classification to determine if the trees are dead or not.   I ran a quick area through using lidar from Open Topography and the California Imagery from 2012 NAIP.    Here's an image attached of the trees in ArcGIS Pro and a shapefile of the trees zipped.

Arthur Crawford