Creating Spectral Indices ( NDVI , NDBI , NDWI etc ) in ArcGIS Pro (example using Landsat 8 Imagery)

Поделиться
HTML-код
  • Опубликовано: 10 сен 2017
  • How to make spectral indices like NDVI , NDBI , NDWI , NDSI or some other indices in ArcGIS Pro
    The procedures are fairly simple, thanks to raster function implemented in ArcGIS pro.
    You must save the result in order to make it become permanent dataset.
  • НаукаНаука

Комментарии • 31

  • @GEO2004
    @GEO2004  Год назад

    these Indices can be combined into a color composite that integrates the unique and rich spectral information of the indices, here is a tutorial exammple for geology/mineralogy mapping application: ruclips.net/video/XLmYqKpxsTs/видео.html

  • @kanruchen8194
    @kanruchen8194 5 лет назад +1

    Hi I saw your comments and downloaded the Surface Reflectance package again, I'm sure the ZIP file size is correct, but I cannot find MTL.txt file...I want to calculate the NDVI, is there some other way to do that? Thanks a lot!

    • @GEO2004
      @GEO2004  5 лет назад +1

      Actually ndvi can be calculated without involves the MTL, you can do that using map algebra/raster calculator or built in ndvi tool available in the image analysis toolbar. Because there is no MTL file, you must add the red band and NIR band image manually using add data button

  • @2516killer
    @2516killer 6 лет назад

    Hi ! I need to make an urban expansion analysis and I need to count all pixels containing roads, buildings, etc. What index should I use ? NDVI, False Color, NDBI ? Thanks in advance !

    • @GEO2004
      @GEO2004  6 лет назад

      Martin Anchov combination of NDBI and NDVI will work best for you mate, and also multispectral classification

  • @coolvidyos
    @coolvidyos 3 года назад

    Hi, thank you for the video. If the NDVI scaling produced is not from -1 to 1, how do i change it so it produces this scale? Thanks again!

    • @GEO2004
      @GEO2004  3 года назад

      I think you better repeat the NDVI generation that resulting directly to -1 to 1 scale, rescaling the signed integer NDVI usually wont give exactly same range values compared to the directly -1 to 1 (for example water surface doesnt have value close to zero)

  • @geospatialworld4180
    @geospatialworld4180 6 лет назад +1

    This video has been helpful.

  • @ellenirogers
    @ellenirogers 6 лет назад +1

    Hi - How did you get the metadata from Landsat? I am trying to use NDBI for calculating urban change and all i get when i order a scene from Landsat 8 is the individual bands and empty metadata files. There is nothing in my MTL.txt files

    • @prem3569
      @prem3569 6 лет назад +1

      Elleni Rogers meta data files will be downloaded along with satellite images. Open that text document in ms word for better visualization

    • @GEO2004
      @GEO2004  6 лет назад +1

      if there is nothing in the MTL files, there is a possibility that it is happened due to downloading error, USGS server sometimes send finished download command to browser (or downloader software), which resulted in completed ZIP files which are the contents inside can be extracted, but it is actually is only partly completed. My advice is (1) check the ZIP file size prior to downloading (or use checksums if you are computer geeks), so If it is done halfway, you will notice it by the file size differences, or (2), get a better internet bandwidth,

  • @dannicaroseaquino881
    @dannicaroseaquino881 4 года назад

    Hello, I am conducting a supervised classification using Landsat bands. I want to improve the classification by including ancillary data such as DEM and the indices you’ve discussed in this video. For instance, I can infer from DEM the areas which may be misclassified and I can use NDVI to differentiate the 5 vegetation classes I have for my classification. My question is how can I simulataneously use the Landsat bands, indices and DEM in the classification? Any suggestions please, if simultaneous utilization of all the data are not possible. Thanks!

    • @GEO2004
      @GEO2004  4 года назад

      for simultaneous classification using ancillary data such as DEM, you can stack the DEM into the imagery multibands dataset using composites bands tool (resources.arcgis.com/en/help/main/10.1/index.html#//00170000009p000000)
      for the other way (easier to do), you can use the Image Analysis Toolbar:
      1. windows - Image Image Analysis
      2. Select all layers which you want to stack
      3. Go to processing tool (Stack).
      the DEM will be incorporated into the imagery bands, and you can directly perform the supervised, unsupervised or even image segmentation using the DEM as ancillary data

    • @dannicaroseaquino881
      @dannicaroseaquino881 4 года назад

      @@GEO2004 Thanks a lot! I am using ArcGis Pro and as far as I know there is no Image Analysis toolbar in Pro. I'm assuming that the stack tool can be found in Raster Functions pane. Is this right?

    • @GEO2004
      @GEO2004  4 года назад +1

      Yes, image analyst toolbar has been replaced by raster functions, the composites band tool is also available in arcgis pro from data management tools>raster

    • @dannicaroseaquino881
      @dannicaroseaquino881 4 года назад +1

      @@GEO2004 Thanks again!

  • @yorkshirepeatpartnership6638
    @yorkshirepeatpartnership6638 6 лет назад

    Very nice. Can you recommend an indicies to determine soil moisture?

    • @GEO2004
      @GEO2004  6 лет назад +1

      one of the best method to determine soil moisture as far I know is Tasseled cap transformation. because it transform all the band into three bands, brightness band, greeness band, and wetness band. Brightness and wetness band could be related to soil moisture as explained in many researchs. Raster functions in ArcGIS has been implemented this tasseled cap transformation, unfortunately it is only working on limited sensor which are landsat (mss to etm), quickbird and IKONOS. If you want to use tasseled cap, make sure you dont perform any preprocessing that can alter the pixel values

    • @yorkshirepeatpartnership6638
      @yorkshirepeatpartnership6638 6 лет назад

      Thank you for the quick and informative response. Do you know which bands are required to perform this function? Don't worry if not I'll look it up online. I'm wondering if this will work with the Sentinal Satellite data so I can examine soil moisture at a finer scale.

    • @GEO2004
      @GEO2004  6 лет назад

      all the bands available in a sensor can be used, for example, in landsat TM and ETM, it uses all bands except thermal bands. Tasseled cap is similar with principal component analysis (PCA) where many bands can be compressed into three bands which store all the variance of every band involved.
      about sentinel-2, yes it will works, however as far I know, there is no 'ready to use' tool for that, so you must perform it manually by learning first how the algorithm works

  • @user-qz9ic9ve7s
    @user-qz9ic9ve7s 8 месяцев назад

    Thanks for this video...but how can i do that in landsat9...metadata when i opend is embty!

    • @GEO2004
      @GEO2004  8 месяцев назад +1

      Use the latest version of arcgis pro.

    • @user-qz9ic9ve7s
      @user-qz9ic9ve7s 8 месяцев назад

      @@GEO2004 can you give me url to downloaded?

    • @GEO2004
      @GEO2004  8 месяцев назад

      Of course not. You must purchase them from esri.

  • @eLTIMIMI
    @eLTIMIMI Год назад

    How to combine spectral indices?
    To create classes

    • @GEO2004
      @GEO2004  Год назад +1

      Spectral indices can't be combined, however, you can combine the classification result of the indices, for example, you can combine the classification result of NDVI and NDBI to generate urban footprint with better precision and accuracy instead of only using NDBI

    • @eLTIMIMI
      @eLTIMIMI Год назад

      @@GEO2004 I am trying to look it up right now. I should combine the classification as you said using Raster calculator right? To create a new raster. I need to use some equation

    • @GEO2004
      @GEO2004  Год назад +1

      Yep, right, you can use raster calculatoe or simply mathematical raster operators available in gis software to do that

    • @eLTIMIMI
      @eLTIMIMI Год назад

      @@GEO2004 How can I send you an image, I think my previous comment was removed because I had a link maybe?