Use Georeference Data to georeference Level 1B AVHRR data, calibration results, and sea surface temperature images using information embedded in the original data file.

If you are going to use AVHRR data for quantitative analysis, you should calibrate the data before georeferencing. Likewise, if you plan to calculate a sea surface temperature (SST) image, you must do this before georeferencing.

Note: It is recommended that you use the more accurate techniques Build Geometry File and Georeference from Input Geometry, instead of Georeference AVHRR Data.

  1. Open an AVHRR file by selecting File > Open As > Optical Sensors > AVHRR >data_type from the menu bar.
  2. Select Geometric Correction > Georeference by Sensor > Georeference AVHRR from the Toolbox. The Enter AVHRR Filename dialog appears.

  3. Select an input AVHRR file and perform optional spatial and spectral subsetting, then click OK.

    If the input data file is not the original AVHRR format file, another Input File dialog appears for selection of the original AVHRR file.

  4. Click OK. The georeferencing information is extracted from the Level 1B AVHRR file. The Georeference AVHRR Parameters dialog appears.

  5. Select the desired output map projection from the list and enter any necessary parameters.

    You should carefully consider which map projection is appropriate for the location and extent of your image. Some projections are more appropriate for AVHRR scale images than others. It is recommended that you not use UTM and State Plane projections for images that cover areas larger than UTM or State Plane zones. If you choose either of these projections, then the resulting registration will be geographically correct but the map coordinates in areas furthest from the origin will be well beyond normal maximum values for these projections.

    With images projected to Geographic Lat/Lon, pixels have constant degree units but not necessarily length units, because the size of a longitude degree varies with latitude. This results in extreme distortion across most AVHRR-scale images.

    Map projections that are appropriate for AVHRR-scale images include stereographic, transverse mercator, or conic projections. Stereographic projections are commonly used for images with equal N-S and E-W extent. Transverse mercator projections are often used for images with greater N-S extent. Conic projections are often used for images with greater E-W extent.

  6. Enter the Number of Warp Points to use for X and Y.

    The Level 1B file contains 51 geolocation points for every line of the image. The Number of Warp Points fields specify the number of embedded geolocation points that ENVI will extract and use as ground control points (GCPs) in the georeferencing. The more warp points you choose, the more accurate the georeferencing will be, but the longer the process will take.

    This utility does not just georeference the image, it actually georectifies it. This is equivalent to performing a full registration warp in ENVI. If you are concerned about achieving the most accurate registration possible, it is recommended that you choose all 51 warp points in the X direction, and perhaps one tenth the number of image lines in the Y direction. If you are georeferencing large AVHRR images (for example, 2000 x 2000 x 5) using thousands of GCPs, be prepared to wait. The process can take hours for such large files.

  7. Click OK. The Registration Parameters dialog appears.

  8. Select the warping and resampling methods.

    9. For large coverages, the Delaunay triangulation warp method produces significantly more accurate results. If you choose the Polynomial warp method, it is recommended that you restrict the degree to 1 (or perhaps 2 in special cases), as this type of warp model often introduces unrealistic features.

  9. Select output to File or Memory.
  10. Click OK.