/

Pan-sharpening

Pan-sharpens Pleiades, SPOT, and Sentinel-2 images using the panchromatic band.


Introduction

For more information, please read the block description.

Block type: PROCESSING

This block pansharpens images of the Pleiades, SPOT or Sentinel-2 sensors. It creates a single high-resolution color image from high-resolution panchromatic and lower resolution multispectral image bands. For detailed information on how Sentinel-2 data is pansharpened see the Advanced section below.

Supported parameters

  • bbox: The bounding box to use as an AOI. Will clip to scenes that intersect with this box. Use only bbox or intersects or contains.

  • intersects: A GeoJSON geometry to use as an AOI. Will clip to scenes that intersect with this geometry. Use only bbox or intersects or contains.

  • contains: A GeoJSON geometry to use as an AOI. Will clip to scenes that intersect with this geometry. Use only bbox or intersects or contains.

  • clip_to_aoi: When set to true, the area that defined in bbox, contains, or intersect for previous data block will be clipped for processing. Note that by default this parameter is false which means that the whole scene will be processed.

  • method: Method used in the pansharpening procedure. Default is SFIM (Smoothing Filter-based Intensity Modulation) as described in [Liu2000]1.

  • include_pan: Include the panchromatic band in the output pansharpened image.

Example parameters using the SPOT DIMAP download block as data source, returning the pansharpened multispectral product appended with the panchromatic band:

{
  "oneatlas-spot-fullscene:1": {
    "ids": null,
    "bbox": [
      13.405215963721279,
      52.48480326228838,
      13.4388092905283,
      52.505278605259086
    ],
    "time": null,
    "limit": 1,
    "order_ids": null,
    "time_series": null
  },
  "pansharpen:1": {
    "include_pan":true
    "bbox":null
    "contains": null,
    "intersects": null,
    "clip_to_aoi": false,
  }
}

Another Example parameters using the SPOT DIMAP download block as data source, returning the pansharpened multispectral product appended with the panchromatic band which is clipped to the specific AOI:

{
  "oneatlas-spot-fullscene:1": {
    "ids": null,
    "bbox": [
      13.415594100952148,
      52.491560852691116,
      13.430356979370117,
      52.49992172845934
    ],
    "time": null,
    "limit": 1,
    "order_ids": null,
    "time_series": null
  },
  "pansharpen:1": {
    "include_pan":true
    "bbox": [
      13.415594100952148,
      52.491560852691116,
      13.430356979370117,
      52.49992172845934
    ],
    "contains": null,
    "intersects": null,
    "clip_to_aoi": true,
  }
}

Another Example parameters using the ESA Sentinel-2 L2A Analytic (GeoTIFF) as data source, returning the pansharpened multispectral product with 13 bands (including the panchromatic band) which is clipped to the specific AOI:

{
    "esa-s2-l2a-gtiff-analytic:1":{
       "ids":null,
       "bbox":[
          13.415594100952148,
          52.491560852691116,
          13.430356979370117,
          52.49992172845934
       ],
       "time":null,
       "limit":1,
       "order_ids":null,
       "time_series":null
    },
    "pansharpen:1":{
       "include_pan":true,
       "bbox":[
          13.415594100952148,
          52.491560852691116,
          13.430356979370117,
          52.49992172845934
       ],
       "contains":null,
       "intersects":null,
       "clip_to_aoi":true
    }
 }

Advanced

Synthetic panchromatic band Sentinel-2

Sentinel-2 provides a high range of multispectral bands with different spatial resolutions (10, 20 and 60 m). Since there is no panchromatic (PAN) band in Sentinel-2 images, we use a synthetic panchromatic band to increase the spatial resolution of the 20 m and 60 m bands to 10 m. The synthetic panchromatic band is generated using the average value of the visual and the near infrared bands. Read more about this process in the paper by [Kaplan2018]2.

Methods

In [Vivone2014]3 an extensive review of pansharpening procedures was performed, with results being assessed on the geometric detail of the final result and additionally the spectral correspondence of the pansharpened result with the input multispectral imagery.

In this paper, SFIM, or Smoothing Filter-based Intensity Modulation (based on [Liu2000]4, has one of the top performances in all of the metrics assessed and because of this we have selected this method as the default pansharpening procedure.

Additionally, two other methods have been implemented, Brovey or Weighted Brovey and Esri, as described below.

SFIM

SFIM has been developed based on a simplified solar radiation and land surface reflection model. By using a ratio between a higher resolution image (panchromatic band) and its low pass filtered (with a smoothing filter) image, spatial details can be modulated to a lower resolution multispectral image without altering its spectral properties and contrast. An additional (optional) parameter has been added to control the blurred edges that appear in the pansharpened result (edge_sharpen_factor) - setting this factor to 1.7 (the default) removes most of this effect. Read more about this procedure in the paper from [Liu2000]5.

Example of parameters to use in the pansharpening block with the SFIM method:

{
  "pansharpen:1": {
    "edge_sharpen_factor": 1.7
  }
}

Brovey

The Brovey transformation is based on spectral modeling and was developed to increase the visual contrast in the high and low ends of the data’s histogram. It uses a method that multiplies each resampled, multispectral pixel by the ratio of the corresponding panchromatic pixel intensity to the sum of all the multispectral intensities. It assumes that the spectral range spanned by the panchromatic image is the same as that covered by the multispectral channels. Read more about this here. The weight parameter can be set to a value between 0 and 1 (default is 0.2).

Example of parameters to use in the pansharpening block with the Brovey method:

{
  "pansharpen:1": {
    "method": "Brovey",
    "weight": 0.2
  }
}

Esri

The Esri pan-sharpening transformation uses a weighted average to create its pansharpened output bands. The result of the weighted average is used to create an adjustment value that is then used in calculating the output values. The weights for the multispectral bands depend on the overlap of the spectral sensitivity curves of the multispectral bands with the panchromatic band. The multispectral band with the largest overlap with the panchromatic band should get the largest weight. A multispectral band that does not overlap at all with the panchromatic band should get a weight of 0. By changing the near-infrared weight value, the green output can be made more or less vibrant. Read more about this here.

Example of parameters to use in the pansharpening block with the Esri method with Pleiades or Spot imagery:

Pleiades weights

{
  "pansharpen:1": {
    "method": "Esri",
    "weights": [0.2, 0.34, 0.34, 0.23]
  }
}

SPOT weights

{
  "pansharpen:1": {
    "method": "Esri",
    "weights": [0.24, 0.2, 0.24, 0]
  }
}

It's not recommended to use the Esri pan-sharpening method with Sentinel-2 data.

Processing

Additional local interpolation of outlier values in the panchromatic bands of Pleiades and Spot data ensures a consistent pansharpened multispectral image.

Optional parameters

  • edge_sharpen_factor: Used only for SFIM method. Factor to reduce blurring of edges in pansharpened result.
  • weight: Used only for Brovey method.
  • weights: Used only for Esri method. The weights in sequence for each multispectral bands that depend on the overlap of the spectral sensitivity curves of the multispectral bands with the panchromatic band. For Pleiades the default weights are [0.2, 0.34, 0.34, 0.23] while for SPOT weights are [0.24, 0.2, 0.24, 0].

  1. Liu, J. G. (2000). Smoothing filter-based intensity modulation: A spectral preserve image fusion technique for improving spatial details. International Journal of Remote Sensing, 21(18), 3461-3472.
  2. Kaplan, G., Avdan, U. (2018). Sentinel-2 Pan Sharpening—Comparative Analysis. Proceedings 2(345).
  3. Vivone, G., Alparone, L., Chanussot, J., Dalla Mura, M., Garzelli, A., Licciardi, G. A. & Wald, L. (2014). A critical comparison among pansharpening algorithms. IEEE Transactions on Geoscience and Remote Sensing, 53(5), 2565-2586.
  4. Liu, J. G. (2000).
  5. Liu, J. G. (2000).