Introduction
This blog provides guidance for users who have existing oriented imagery catalogs created using oriented imagery classic and need to migrate to the formally supported oriented imagery dataset and oriented imagery layer format.
If you are using oriented imagery today, you already understand the core concepts. This post focuses specifically on migration paths, supported options, and required changes. Deprecation of Oriented Imagery Classic was announced in January 2025, and it has now been retired with the February 2026 ArcGIS Online update.
Why Migrate to the New Format?
Oriented Imagery Classic was a prototype implementation of concepts for non-mapping imagery, including the oriented imagery catalog data model. However, it was never released as a formally supported product.
Oriented imagery is now fully integrated into the ArcGIS system. Oriented imagery layers—which replace oriented imagery catalogs—are supported as a first‑class layer type across ArcGIS applications.
Classic vs. Current Format
An oriented imagery catalog (Classic – now retired) consisted of:
- A point feature layer defining camera locations, with feature attributes defining orientation and image metadata
- An optional vector tile layer showing imagery extents
- A separate JSON file defining layer‑wide properties
These components could be shared locally or published as three separate items in ArcGIS Online or ArcGIS Enterprise.
An oriented imagery layer (Current – fully supported across ArcGIS) is a type of point feature layer where the point geometry defines the camera location, the feature attributes define the orientation and other key image metadata, and the layer properties define key layer-wide metadata.
The data source for an oriented imagery layer can be an oriented imagery dataset, created and managed in ArcGIS Pro, or a feature service that includes an oriented imagery layer.
An optional coverage feature layer showing imagery extents can be defined in the oriented imagery dataset properties, or published as an additional layer in the feature service.
What’s Required to Move Forward?
At a minimum, all future use of oriented imagery must be based on an oriented imagery dataset and/or oriented imagery layer.
You may need to make updates to your imagery, metadata, and/or applications, depending on:
- Where your images are hosted
- How your Classic catalog was created
- Which applications consume the imagery
The sections below walk through these decisions step by step.
Step 1: Choose a Supported Image Hosting Option
Before creating a new oriented imagery dataset, you must ensure your images are hosted via a supported method.
Image Hosting Options
Users are able to host images using one of the following supported options:
- Public cloud storage (for example, Azure Blob or Amazon S3),
- Secure feature attachments within the oriented imagery layer, hosted in ArcGIS Online or Enterprise, and
- Secure 3rd-party storage accessed via a Custom Data Feed (ArcGIS Enterprise 11.5 or later)
Users can also host images using a local web server (such as IIS). Local files are supported directly for desktop-only workflows.
Note: In Oriented Imagery Classic, secure image access in 3rd party storage was possible using a custom token server. This approach is not supported in the current oriented imagery capability. We are working on adding support for secure image hosting using ArcGIS Enterprise, which will become the recommended solution for securely hosting images once available.
Image Hosting Scenarios and What to Do
Images already in public cloud storage or a local web server
No changes to image storage are required. You can reference these images directly when creating your new oriented imagery dataset.
Images hosted as feature attachments in a Classic catalog
You have three options:
- Start over with your local copies of the source images, or
- Download the images (see here) and re‑upload them as feature attachments when publishing the new oriented imagery layer, or
- We will make a Python script available soon that can read and reformat published Oriented Imagery Classic layers in place. You can check back here for updates when that becomes available.
Images hosted securely
If you need to host images in a secure 3rd-party storage in the near term, use a Custom Data Feed in ArcGIS Enterprise to access imagery and metadata. Refer to the Custom Data Feed documentation for configuration details. Unless you need the flexibility of a custom data feed, Esri recommends migrating to secure image hosting using ArcGIS Enterprise (mentioned above) when it becomes available.
Step 2: Create a New Oriented Imagery Dataset
Esri strongly recommends returning to your original source data to create a new oriented imagery dataset. Use ArcGIS Pro to run the Create Oriented Imagery Dataset geoprocessing tool to create an empty dataset, then run Add Images To Oriented Imagery Dataset to populate the metadata.
Your input to the Add Images tool depends on how your Classic catalog was originally created.
If Your Classic Catalog Was Created by Reading EXIF/XMP tags in each image
- Provide a list of image URLs, file paths, or a local folder
- Image names are not required
- Images must be accessible (for example, public cloud storage)
When you run Add Images To Oriented Imagery Dataset, the tool will read the EXIF & XMP metadata and populate required fields automatically.
If Your Classic Catalog Was Created from a CSV table of metadata
- Create a new table formatted for the oriented imagery dataset
- Your original table (if available) should be reusable as a starting point
- If needed, export the attribute table and point geometry from the Classic catalog
- Recommendations:
- Follow the field names in the current oriented imagery schema (some field names from Classic have changed)
- Use separate values for the orientation angles (heading/pitch/roll or omega/phi/kappa) in the table instead of the legacy CamOri string where possible
- Verify geometry and orientation values before publishing
If Your Classic Catalog Used an Image Service or Mosaic Dataset as the source
- Image services are not supported as an image hosting option.
- If you used a mosaic dataset, you may use it as a data source to build a new oriented imagery dataset. You must still choose a supported image hosting option for the underlying images.
- If you do not have access to the original mosaic dataset, you can start from the beginning, or export the attribute table and point geometry from the Classic catalog.
Step 3: Building Coverage and Publishing
Once your dataset is populated:
Step 4: Update Client Applications and Web Apps
Once new layers are published, all client applications must be reviewed and updated. Any application referencing Classic catalog URLs must be updated.
Experience Builder
If your app uses the Oriented Imagery Classic widget, it must be rebuilt using:
- Oriented Imagery Viewer widget, or
- The Sidebar or 3D Viewer Instant App templates available in Experience Builder within the Online/Enterprise portal
The Classic widget has been removed from Experience Builder with the February update to ArcGIS Online.
Special Cases
Oriented Imagery Catalogs Created with ArcGIS QuickCapture
If your catalogs were generated by ArcGIS QuickCapture, refer to this blog to use QuickCapture designer to update your existing Classic layers to the current oriented imagery format.
Features Not Yet Supported
The following features are not yet supported in oriented imagery datasets and layers, but are planned for later in 2026:
- Video files as source data
- Superimposing images in a 3D scene
If these capabilities are critical, refer to this blog for options to continue using Oriented Imagery Classic in the short term.
Final Recommendations
- Test migration workflows with a small subset of imagery first.
- If you no longer have your original images, perform thorough validation before deleting Classic catalogs or services.
- Update documentation and user instructions once new layers are live.
