Issue with RapidEye large mosaic dataset

439
1
06-04-2014 06:32 AM
by Anonymous User
Not applicable
Original User: hiranzani

Hi all �?? I have been trying to build a large mosaic (> 400 scenes) with the mosaic dataset. The major issue so far is with the Color Balancing function.
I´m aware that color balancing can only take place if the following is true about the data:
               All the bands have their statistics calculated.
               All the bands have their histograms built.
               All the raster datasets have the same number of bands. -
               All the raster datasets have the same pixel type and pixel depth. Only 8-bit unsigned and 16-bit unsigned bit depths are supported.
               None of the raster datasets have an associated color map
However, the statistics result seems distorted and this may be compromising the color balance. For example, the digital numbers mean are close to the maximum value (65536) and this s certainly not true.
My dataset comprises of RapidEye as received from the distributor with 16bits.
I really appreciate if anyone has some thoughts on my problems.
Regards
Hiran
0 Kudos
1 Reply
by Anonymous User
Not applicable
Original User: hlzhang301

Hi all �?? I have been trying to build a large mosaic (> 400 scenes) with the mosaic dataset. The major issue so far is with the Color Balancing function.
I´m aware that color balancing can only take place if the following is true about the data:
               All the bands have their statistics calculated.
               All the bands have their histograms built.
               All the raster datasets have the same number of bands. -
               All the raster datasets have the same pixel type and pixel depth. Only 8-bit unsigned and 16-bit unsigned bit depths are supported.
               None of the raster datasets have an associated color map
However, the statistics result seems distorted and this may be compromising the color balance. For example, the digital numbers mean are close to the maximum value (65536) and this s certainly not true.
My dataset comprises of RapidEye as received from the distributor with 16bits.
I really appreciate if anyone has some thoughts on my problems.
Regards
Hiran


Hiran, share as follows:

Firstly, ArcGIS raster models (including Mosaic Dataset) are still facing �??operational�?? challenges to produce an �??on-the-fly�?? color mosaic containing massive sensor-based images like RapidEye, even though having a collection of GP tools to help users conduct statistics computation and color-balancing.

For your given project (with over-450-scene RapidEye), advisable to pre-process (color-balance, orthorectify with/ without GCP) by remote sensing packages (ERDAS Imagine/ ER Mapper, PCI, or ENVI) and then manage those pre-processed images within Mosaic Dataset, which make use of some advanced functions like edge detection within MD.  Pls refer to http://forums.arcgis.com/threads/71343-From-imagery-processing-perspective-the-mosaic-dataset-model-...


++++++
Personally, the use of those color-balance GP tools in ArcGIS is strictly �??problem-oriented�??. In other word, image enhancement (color-balancing) is more an art than a science in practice, and the definition of �??properly enhanced�?? images / mosaic is highly based on subjective criteria, like feature/ edge clarity and beautiful color in whole mosaic, etc.

In operation, we couldn�??t ignore those subjective criteria, because of basic requirements from customers and applications.
0 Kudos