Digital Image Processing of a Satellite Image
For image processing, need satellite images, which is also called digital Image/data.
Introduction to Digital Image/Data (Satellite Image)
Introduction
- An image is an array, or a matrix, of square pixels (picture elements) arranged in columns and rows.
- Digital images consist of discrete picture elements called PIXELS. A number is associated with each pixel represented as DN (Digital Number).
- DN depicts by positive numbers, which have the average of electromagnetic energy captured by satellite sensors.
- The pixel size is reduced more scene detail is preserved in digital representation.
- The digital numbers are in the form of ‘bits’ which vary from 0 to a selected power of 2. Such as if a 8 bit data then it becomes 28 and represent 256 variance recorded by sensor of reflected radiance from object.
There are two type of satellite image
- Black and White, which is also called panchromatic image.
- Color composite, which can be displayed or analyze in the form of natural color composites and false color composites.
Black and White Image
It is also called, grey scale image in which picture element has an assigned intensity that ranges from 0 to 255 in 8bit data.
A grey scale image is normally call a black and white image. As the name defines that such an image includes many shades of grey.
Color Composite
Color composite is displaying the different bands of a multi-spectral data set. Color images obtained by using different bands, which represents reflected energy received in different wave length of electromagnetic spectrum.
Color composite can be prepared as true color composite in which blue wavelength blue color, green wavelength green color and red wavelength is assigned red color.
Whereas, false color composite (FCC) is composite of green, red and infra-red wavelength, which are assigned blue, green and red respectively.
Satellite images are acquired in bands and to make color composites, bands need to stack into a single layer or file.
Image Processing is defined as the “examination, processing and analysis of satellite images to identifying objects and extracting information”
In this process visual and computer-based processing techniques, tools and equipment is used to detect, identify, classify, measure and evaluate physical objects, their spatial patterns and relationships through remotely sensed data.
Image processing basically includes the following below steps.
- Acquiring/Procure satellite image through freely available sources or purchase satellite images.
- Satellite image may available in layer or bands, so it requires several operations before digital image processing.
- Importing the satellite image from one format to other format as per need.
- Image enhancement to create a better image.
- Image classification to assign or create categories through interpretation.
- Image classification can be performed by two basic methods supervised or unsupervised.
- Post processing assessment to improve classification.
- Accuracy assessment to find out the quality of image classification.
- Output is the last stage in which result can be displayed in the image form, report and data for monitoring, decision making and planning.
This process is provided in image processing software. In below matter, the layer stacking process is explained using ERDAS image processing software.
Layer Stacking
Following steps followed in Erdas image processing softwareto perform layer stacking
1. Steps to access layer stack tool
- Erdas Main Tool Bar
- Select Interpreter
- Utilities
- Layer Stack icon
2. In the Layer stack dialog, enter the name of the first file in the Input File.
Be sure the “Files of Type” is set to the same input type as your data downloaded or acquired. GEO-TIFF is normally used in Raster images.
The output image will contain the bands in the order that they have been input, so first add band 1, second will be band 2, etc.
3. Once a name is entered click Add. You will see the name of the file appear in the gray dialog list box. Continue entering and adding each of your files.
4. With this process all the bands are added one by one by selecting downloaded geotiff files/bands.
5. Define output file name, and select output file type. Generally it should be same as format supported by your image processing software. In case of Erdas it will be img format.
6. Click the Union output option, and choose to ignore Zero in stats (this will produce brighter image). Click OK. The process will run. Your layers are now stacked in one image or file.
It's very good and bigger platform to learn about image processing of satellite.
Thank you for appreciation, please visit regularly for next article.