ArcGIS Pro 3.0.1: How to preserve the visual quality of a tiff format imagery as it’s added to a mosaic?
ArcGIS Pro 3.0.1: How to preserve the visual quality of a tiff format imagery as it’s added to a mosaic?
The screenshots below show that the visual quality of a tiff format imagery gets changed as it’s added to mosaic.
What could be the issue here? how to get the imagery and its mosaic identical in terms of visual context?
A few things come to mind. Check to see if the mosaic layer and the raster layer are using the same Stretch Type, and make sure the statistics have been calculated for the mosaic dataset.
For our rasters that are aerial imagery, I set the source type to Processed so that the Stretch Type is set to none when adding it to ArcGIS Pro. I do the same for any mosaic datasets that use the aerials. You can use the Set Raster Properties geoprocessing tool in batch mode to set the correct source type for multiple rasters at the same time.
With Aerial imagery, the Bilinear resample type seems to work best.
Make sure you have both set to the same for accurate comparison:
I have seen issues with this before when the mosaic pyramid/overviews are created with default Nearest Neighbor type, Until I zoom past the source resolution of the dataset, they can appear different.
So, I make sure the statistics,pyramids,overview settings all say to use Bilinear, then once added to the map, often have to import symbology from the original tiff to view the mosaic one properly.
R_
I couldn’t manage to get the mosaic and the imagery identical in terms of visual context.
As you have already recommended, I applied:
- Statistic
- Bilinear
- Processed data source type
Nevertheless, they never got visually identical as per the screenshots below
I see the statistics are the same so that is not causing the difference. One thing to try to troubleshoot the issue is to set the Explore tool to "Visible Layers", set the Stretch Type to "None" for both raster and mosaic, and with the raster and mosaic both displayed click on the map and compare pixel values.
If the pixel values in the popup are the same, then they are most likely not being stretched the same way when a stretch is applied.
If the pixel values are different then something in the configuration of the mosaic dataset is modifying the image. Some things to check could be the mosaic projection different from the source projection causing resampling, the overviews using a different compression and resampling than the raster's pyramids, or the mosaic transmitting with a compression applied.
I use mosaic datasets to manage my organizations imagery acquisitions. Our images are already color-balanced so we do not use a stretch or color-balance the mosaic dataset. I keep the mosaic settings as simple as possible (same projection, product definition is "None", manually set the number of bands and pixel depth, use the raster's existing statistics and pyramids on raster import, calculate the mosaic's statistics, define the overviews, build the overviews, and set the mosaic's data source type to "Processed"). With that setup my mosaics match the original rasters pixel for pixel. Using a transmission compression method of None or LZ77 will not affect the pixel values. JPEG definitely will change the pixel values and LERC might change the pixel values depending on the tolerance set.
- It appears that the “color balance mosaic dataset” tool doesn’t get the imagery and its mosaic visually matched
- Right, the RGB values are matched in both the imager and its mosaic as per the screenshots
Not sure what other approaches can get the imagery and its mosaic matched
