Modelling NDVI, NDWI and Other indices in ArcGIS

Indices are very good techniques for analysing remotely sensed data. Especially the indices like NDVI, NDWI, NDBI, NDSI etc. are quite often used in Geospatial Science.  This tutorial explains the procedure to model these indices in ArcGIS.

Once the concept and the method are understood any indices can be modelled if the formula is known.  Watch the video till end to understand the compete concept and to understand the places where caution is to be exercised to avoid errors.

View other posts in the blog to model these indices in ERDAS Imagine

See other posts to carryout these indices in ERDAS Imagine.

Comprehensive Guide to Geo-Referencing

The fact that you can pin point the location of everything is the most vital part of geospatial technologies. Though the satellite images that we download from internet (say earthexplorer or bhuvan) have a predefined coordinate system, there many other things things that are not georeferenced. We can quote examples like scanned toposheets, guide maps, Images obtained from Google Earth etc. Here is a comprehensive set of videos to carry out the georeferencing process with different cases.

Georeferencing a Satellite Image Using another Georeferenced Image:
Say you want to georeference an image, and you have another image that is already georeferenced (say a satellite image downloaded from earthexplorer) and that covers the first image either wholly or part.

Now you will be taking the reference from the second image (that is already georeferenced) and georeference the first image.

Carry out this process in ERDAS

Same Process in ArcMap

Georeference Satellite Image Using Available Latitude and Longitude (GCPs):
It is quite possible that there is no reference image available or the resolution of the two images are so different that you find it difficult to precisely identify common points. In such a case if you have GCPs (latitudes and Longitudes of any point on the image), may be collected from Google Earth or using a GPS or any other means, you can correlate those GCPs with the points on the image and georeferenced the image.

Carrying out this process in ERDAS Imagine

Same Process in ArcMap (This process is also a method of downloading and Georeferencing images from Google Earth)

Georeference an Image from Google earth:

For a free high resolution image, Google Earth is a very good solution. However whenever the image downloaded from Google Earth will not have spatial references. So it is virtually unusable for any geospatial project. Following is a method to download and georeference the image from Google Earth. Here both image and reference is taken directly from Google Earth.

Carrying out the process in ArcMap

Another Method

You can try the similar process in ERDAS Imagine by following the similar idea and using the technique shown in the previous section (georeference using Latitude and Longitude in ERDAS Imagine)

Georeferencing A vector File:
There are cases where we end up digitizing an image and then realise that the image is not georeferenced. In such a case we will be needed to georeferenced the vector file, which is not as straight forward as a Raster file. Here we will be needed using spatial adjustment tool to carry out the process. Here we will be using another georeferenced file as reference.
Carrying out georeferencing of Vector File in ArcMap.

Changing the Projection of A georeferenced file:
There are many projection system, each suitable for various areas and various projects. Dpending on the requirement, there might be need to change the projection from one to another. Here is the tutorial in ArcMap to change the projection of a georeferenced file.

Increase or Enhance the Spatial Resolution of DEM using Spatial Interpolation Technique in ArcMap

Digital Elevation Model is very important dataset for carrying out many geospatial analysis, especially in processes like Hydrology. There are free DEMs available online for the benefits of researchers thanks to organizations like NASA and ISRO. However the freely available DEMs generally have a spatial resolution of 30 meters or more. For many applications we might need higher resolution.

Interpolation predicts values for cells in a raster from a limited number of sample data points. It can be used to predict unknown values for any geographic point data: elevation, rainfall, chemical concentrations, noise levels, and so on.

Here I have just interpolated the values of the DEMs and created a higher resolution DEM. In this example I have increased the resolution of the DEM from 30 meter to 10 meter. You can follow the video and get your high resolution DEM. However there are certain disadvantages, to do watch the video completely to understand the theory behind it and also the disadvantages.

Here is the video

Note and Tips:

• The accuracy of the result will not be as good as the accuracy of the original 10 meter resolution DEM.

• Try to understand different types of interpolation techniques, so that you can choose the best interpolation technique suitable to you.

Gapfilling or Destriping Landsat 7 Image for Scientific Analysis - ERDAS Imagine

Almost everyone in the field of Remote Sensing and GIS would have felt thezcpinch of stripes in Landsat 7 image one or other time in their career. In 2003 Landsat 7 has developed a technical snag which called SLC (Scan Line Correlator) off, which has resulted in stripes or gaps throughout the image. Using these images for Scientific Analysis were virtually impossible.

NASA has developed a few techniques to destripe these images. Here is a method using ERDAS imagine to Gapfill Landsat 7 image as suggested by NASA for scientific analysis. Though this problem can never be addressed perfectly as the problem is with the sensor, this is probably the best available solution.

Here you will need a historical images. Basically means to say, you will have an image with stripes for which you will considering for gapfilling, and another image (either without gaps or with gaps) from which the data will be extracted and filled in the gaps of the first image. 

A simpler method for gapfilling using Focal Analysis method in ERDAS imagine is also available in my channel, however that method is best suited for display purpose and not for scientific analysis.

Do watch video completely as I have also briefly explained as to how to select the images for carrying out this work. If not understood completely, you might end up selecting wrong images resulting in accurate and unsuitable result.

Following is the video.

.

Notes and Tips:

• While selecting historical images for gap filling, select images which are temporally and seasonally close to each other.

• Do remember to layer stack all the bands once you have completed the gapfilling process for individual bands

• If you are looking for an easy process for display purpose, you can use focal analysis in ERDAS Imagine. Look for the other post for the video tutorial on this.

• Do try to select cloud free data to the extent possible.

.

Download Satellite Images and DEM for Free

Most vital part of any GIS and Remote Sensing work is satellite data. Buying a satellite image for research purpose is quite expensive. However there are many online resources those provide data for this purpose free of cost. All that you may need is to credit them for using data from their archive. 

earthexplorer: Is a place where you will find satellite images, DEM and many other related data for any area around the globe. You get data right from 1972. You get data from various sensors like LANDSAT series, ASTER, SRTM etc. This is an amazing service provided by NASA. Do remember to credit NASA in your research work.

bhuvan: If you are from India, you would consider NRSC (National Remote Sensing Centre) data first. It is also observed that, bhuvan data scores better than earthexplorer for Indian region. You get data primarily from LISS and CAROSAT. But note that you get data only for Indian subcontinent. Also to mention that data is available after 1990s.

Following is the tutorial illustrating the procedure to download satellite Images and DEM for free.

Notes and Tips: 

• You will need to register to both sites as a user first in order to download data.

• Your any query related to NASA products can directly be sent to the mail ID provided. You can expect an early and prompt reply (personal experience)

• Both in earthexplorer and bhuvan, do explore all the options for all available data, you will get many more data products that would be useful for your work. So just do not log out once you get the data, you are looking for, explore more.

• Download natural look like data before downloading the image to check for quality of image. Primarily to have a look at the cloud condition.

Calculation of NDWI (Normalised Differential Water Index) in ERDAS

The normalized difference water index (NDWI) is uses similar principles as Normalized Difference Vegetation Index (NDVI) to enhance the reflectance values of surface water bodies. NDVI was introduced in 1973 by Rouse et al.  Differences of two bands, red and near-infra-red (NIR) is carried out suing the formula. NDVI=  ((NIR-RED))/((NIR+RED)), As a result the Spectral Reflectance values of the image will be highlighted. Using similar principles in 1996 GAO derived NDWI technique. The primary idea was to assess the liquid water content of vegetation canopy. Formula used is NDWI=  ((NIR-SWIR))/((NIR+SWIR)) . Where SWIR is Short Wave Infrared. This was further modified by McFeeters in 1996 to enhance the spectral reflectance values of surface waterbodies in an image. The modified formula is NDWI=((Green-NIR))/((Green+NIR)) .

There is readymade model available in ERDAS on NDVI. All we need to do is just replace the relevant bands as required. Primary concept involved is the arithmetic operation as per the formula on each pixel of one band with the corresponding pixel in the bother band.

Here is the video on calculating NDWI in the simplest way using ERDAS Imagine.

Notes and Tips:

• There will be slight changes in carrying out the process depending on the version of the software. 

• Logically and Theoretically the pixel values of NDWI and NDVI result shall be varying from -1 to +1. In ERDAS, it is stretched to 255 values for easy comprehension. So don't be under doubt on the accuracy of the result seeing the values beyond the range of -1 to +1.

• NDVI is explained in another post.

• You can try these only with proper satellite images or UAV images having Infra Red band. This can not be done without IR band as you see in the formula, so do not try with Google Earth images. 

Changing Coordinate System and Projection

Coordinate system and Projections are most important part of any Map. This helps in in determining the location of any object on the map. There are various coordinate systems and projections. Even if the files are having geographical reference and they are of same region, they may not overlap due to different projection.  Simplest way to definition as follows.

Coordinate System: A reference system with which every location(point) on the surface of the earth can be located using intersection of a latitude(easting) and a longitude(northing). 

Projection: It is the method of representing 3D earth surface on 2D paper.

You can further study about Coordinate System and Projections at your free time to get a detailed understanding.

It is very likely that various maps, satellite images, DEMs, GPS coordinates be in different projections and hence they do not overlap. So it is always advisable to have all the files of a project in the same Coordinate System and Projection. So there would be a need to change Coordinate system or Projection of one file into another. Below is the video which illustrates on how to change projection from one to another.

Notes and Tips:

• Just try to understand the basics of the projection and coordinate system you are going to retain in your project, that will help you in taking the best decision to select the projection and coordinate that suits best to your work.

• It is always better to import the projection from another file than defining on your own, as there are possibilities of human errors while you define on your own.

• in ArcMap version above 10 importing Projection from another file is slightly different than in 9.3. In the window there is a small down arrow next to the globe symbol in the search row. When you click on that, you will get the option "Import" 

Download Very High Resolution Georeferenced Satellite Images

High resolution satellite images are of utmost importance in many RS and GIS studies. However high resolution images are not available for free of cost. Hence it is often beyond reach for student and research community. While using google earth images for this purpose, often we end up in accuracy issues while carrying out georeferencing. This tutorial explains a method using Elshayal Smart GIS Application to download already georeferenced Very High Resolution Google Earth Image.

Watch till end to understand how to do it for larger region of interest and also to understand certain disadvantages.

The software directly takes the coordinates from Google Earth to georeference the image. However there are a few points to note,

• As the georeferencing is done automatically, you will not find any issue with accuracy or edge matching (However read the note and tips below)

• As the Images are from Google Earth, you will find only RGB bands. Since there will not be Infra Red band, these images have limited uses for applications like Normalized Differential Indeices, getting an FCC image etc.

• However these images can be used in applications, like Urban Planning and Development, Flood Monitoring etc. which need high resolution natural look images

• You can use these Images along with free Landsat or other images to use in any application

Below is the video on Downloading High Resolution Georeferenced Satellite Images using Elshayal (Smart GIS).

Note and Tips:

• In some cases edges may not match if caution is not exercised. While saving the image in Google Earth as shown in the video, take care that there is no shift in the image. Sometimes while saving the image, you will feel a small shift in the screen up or down (due to whatever reason), in which case edge matching will be an issue. So take care that there is not shift, if at all there is a shift you may have to redo the process.

• Make sure that Google Earth is connected to Elshayal (Smart GIS) before you save the image, else it will return an error.

• In some cases software says, Server is Busy. It is a small problem with Google Earth. What you all need to do is, close the GE, reopen it. Zoom to the level you want, make sure that the area you are looking forward is loaded completely. i.e, just pan your screen all over the place and make sure that all the area are in high quality(not blurred). Then do the usual process as shown in the video. It will work. It should do it automatically, but somehow in some versions of GE, as the next screen is loaded, the screen keeps buffering and doesn't automatically show up in High Resolution.

Supervised Classification in ERDAS Imagine

Classification is one of the very basic and important parts of Goespatial Technologies. Any satellite image will generally have 256 discrete values. Hence talking from layman’s point of view, every image will have around 256 classes. This implies that vegetation might be covered in 50 classes based on their DNs (As different type and stages of vegetation will have different spectral reflectance values).

Classification is basically a method which puts all these 50 values into 1 class of vegetation. So in general classification can be defined as the process of grouping all the pixels of an image into a specified number of classes.

Primarily there are two types of classification,

Unsupervised Classification: Discussed in unupervised Classification video in the blog.

Supervised Classification: This is type of classification that requires quite a bit of human intervention. Here the user will define something called signature set, which are primarily samples of the classes user is going to define. User will digitize a sample portion of a feature and then save it under s specific name. For example user will digitize a small potion of a waterbody and save it under the name water. Classifier will check the entire image and groups all the pixels with similar spectral properties of a specific spectral signature. This is the basics of Supervised Classification and as you can realise there is a need for the user to supervise the entire process. Resampling method selected will also affect the classification result to some extent.

Following is the video on Supervised Classification Using ERDAS Imagine.

Below is the video on classification if an image using ERDAS Imagine.

Notes and Tips:

• Accuracy of the classification only depends on the accuracy of the signature set. So take extra care while you define any signature
• If you are classifying the image into only a few classes say less than 5, may be unsupervised classification is the better solution
• If you feel there is mis-classification, you can edit the signature set and define a new class (or a class that is already existing, but you feel that the mis-classified portion should belong there) selecting the portion which you feel is mis-classified.  With this, you may end up having two classes of vegetation but likelihood of mis-classification will be addressed to some extent. While preparing map where you have more than one class for any feature, you can assign same colour to both the classes and keep only one class in the legend in the map

Unsupervised Classification using ERDAS Imagine

Classification is one of the very basic and important parts of Goespatial Technologies. Any satellite image will generally have 256 discrete values. Hence talking from layman’s point of view, every image will have around 256 classes. This implies that vegetation might be covered in 50 classes based on their DNs (As different type and stages of vegetation will have different spectral reflectance values). 

Classification is basically a method which puts all these 50 values into 1 class of vegetation. So in general classification can be defined as the process of grouping all the pixels of an image into a specified number of classes.

Primarily there are two types of classification,

Supervised Classification: Discussed in Supervised Classification video in the blog.

Unsupervised Classification: This is the simplest way of classifying an image, where human intervention is minimum. Here the user will just define the number of classes and there after we will not do any sort of supervision. System will classify the image based on the DN of the pixels into the number of classes defined by the user. A general comment may be made that, the DNs having same and close by values will be clubbed into one class. However the entire process that happens in the background is not so simple, it also varies slightly based on the re-sampling method selected.

Below is the video on classification if an image using ERDAS Imagine.

Notes and Tips:

• There are possibilities of mis-classification, for example, a portion of vegetation getting classified under the class water. If you find such a case, one method of addressing it to some level is to increase the number of classes. With this, you may end up having two classes of vegetation but likelihood of mis-classification will be addressed to some extent. While preparing map where you have more than one class for any feature, you can assign same color to both the classes and keep only one class in the legend in the map.
• Use unsupervised classification when you are looking for less number of classes. If you are looking for more number of classes, like more than 8 or 10, it is advisable to used supervised classification.

Generate Contours and DEM from Google Earth

Vitality of elevation information in the field of Geospatial technologies is a well known factor. This may be in the form of Spot Heights, Contours or DEM. Unfortunately the open source DEM we get is generally of spatial resolution of 30 meter and up. How many aware of the fact that the Elevation Information available in Google Earth can be utilized to get Spot Heights, Contours and DEM? Yes, the 3rd dimension information available in Google Earth can be manipulated to create Spot Heights, Contours and DEM in various GIS software using and small third party app. The app primarily gives you the elevation and all related information of each point that you have digitized on Google Earth in Excel format. Spot Heights can then be interpolated to get contours or DEM. Below is the link for the tutorial to create Contours and DEM from Google Earth.

Watch the video completely and understand the theory behind as well. If you are not looking for very high spatial resolution DEM, you can download ASTER DEM directly from earthexplorer or if you are from India you can download it from BHUVAN (I find that for Indian Region, CARTO DEM is better than ASTER).

Notes and Tips: 

•  Few people have asked me that altitude column is not getting updated. That is either problem with the installation or operating system. So try to re install tcx and try. Make sure that internet is connected then you do it. If it still doesn't work, try to do it in different operating system or different computer. Most of the people told it worked after doing this.

•  In the tutorial method is explained using SURFER to get final output. To do the process in ArcMap, follow the tutorial till getting excel file with elevation values, and then follow the other tutorial where it is explained how to create contours using spot heights (already available in the channel.)

•  Elevation information available in Google Earth is also interpolated values, so DEM or Contours extracted are as accurate as Google Earth Values.

Geo-Spatial Tutorial

Hello Geospatial Enthusiast,

Many of you visiting this blog might already be familiar with my Geo-Spatial Tutorials youtube channel. Remote Sensing and GIS is that vital part of Earth Science that can be applied to anything, anywhere and anytime. Though the spread of this technology is across the spectrum, the availability of the resources to effectively apply these are very scarce. Geo-Spatial Tutorials is a genuine effort to put many aspects and techniques of Remote Sensing and GIS together, free of cost, for the benefit of Earth Science fraternity.

I hold doctoral degree in remote Sensing and GIS. Here you will find the video tutorials on Remote Sensing, GIS and Photogrammetry.  Tutorials are primarily practical in nature, with small explanation on the theoretical part. However in the blog, along with every video I will be providing small write up on the theoretical part related to that video. Also few points and tips to note before you start trying, based on the E-mails I received from subscribers from youtube channel. Going through them will help in addressing the problems or questions those were faced by other fellow enthusiasts even before you face them.

Feel free to contact me on the E mail address provided for any question that you have. Generally I reply within 24 hrs. However please note that due to the large number of mails received, there might be a delay in response. Being a working professional, I make it a point to dedicate some amount of time on a daily basis to reply all mails. However due to the significant increase in the number of mails from both students and professional communities, I find it difficult to provide solutions to the problems that requires considerable amount of time and involvement. I will indicate the same in the reply to such mails. However for the queries those consume significant time and efforts, I will be happy to provide solution/consultation at mutually agreed consultation charges.

Do note that it is advisable to read the note provided along with the and also watch the video completely video prior to trying. Doing so will automatically answer many of the questions that would arise. Many of the mails that I receive are either due to not reading the points or not watching the video completely. Also do go through all the videos before asking for any video as it would already be present in the website.  These practices would help in effective and efficient use of time and resources.

Do subscribe to get automated updates.

HAPPY MAPPING