Detecting change with AI parcel data
One method for reducing errors is to look at simple changes of your parcel data. For example, when a property goes from having no solar panel to having one, a swimming pool is removed, or a new building appears on a vacant lot.
This is robust to a number of errors – slight shifts in location or extent of a detected object due to imagery shift, parallax error and lighting. You can make it even more robust by only looking at changes with high confidence, as shown in the example.
Note that this simple form of change detection can only find major changes such as objects being created or demolished. More subtle changes will be ignored, such as when a solar array is expanded from 10 to 20 panels, or an extension is added to a dwelling.
Extending this workflow
You can extend this approach by taking advantage of the two tiers of AI content you can export from MapBrowser:
- The total area of objects of a particular type (and measuring change) from the spreadsheet file, or
- Counting the number and area of individual objects through geospatial calculations on the rich geospatial file.
While still being robust to reasonably large shifts due to parallax error and on-the-ground accuracy, looking at variations in area in the spreadsheet file can reveal changes such as the installation of a new shed, an extension, or an expanded PV array.
You can use a workflow similar to the workflow described here to sort the magnitude of change (difference in object counts or area), where the largest area of change is a useful proxy for the confidence that the change is real. This workflow can determine the point at which you can ignore changes below a certain threshold, depending on your requirements.
