Remotely sensed images are fast becoming an important source of spatial information. Digital remotely sensed images are widely recognized as one most practical means for spatial information updating of modern Geographical Information Systems (GIS), especially in real-time applications.

However, for most applications, the digital data can only be used if it can be correctly interpreted, classified and presented in the same way as data from other sources, such as maps and charts.

Remotely sensed images, as seen in a hard copy or on a computer monitor, look much like normal photographs, except that they are rendered by numbers that represent brightness/color values.  To extract thematic information from the array of digital numbers constituting an image, three basic factors must be considered, namely spectral, spatial and temporal characteristics. The spectral characteristics of an image refer to the nature of electromagnetic radiation (emitted or reflected from the earth’s surface) that is recorded by the sensor. The spatial characteristics describe the size of the earth’s surface features that may be identified, and this in turn depends upon the spatial resolution of the sensor.  The temporal characteristics of images must be considered while attempting to monitor the environment, not only the nature of natural phenomena, but also human induced disasters.  The temporal resolution of remote sensing data depends upon the characteristics of the platform. 

The goal of image processing for earth science applications is to enhance geographic data in so as to make it more meaningful to the user, extract quantitative information and solve problems.  The term digital image processing refers to the use of a computer to manipulate image data stored in a digital format.  A digital image is stored as a two-dimensional array (or grid) of small areas called pixels (picture elements), and each pixel corresponds spatially to an area on the earth’s surface. This array or grid structure is also called a raster, so image data is often referred to as raster data. The raster data is arranged in horizontal rows called lines, and vertical columns called samples. Each pixel in the image raster is represented by a digital number (or DN).

These image DNs can represent many different types of data depending on the data source. For satellite data such as Landsat and SPOT, the DNs represent the intensity of reflected light in the visible, infrared or other wavelengths. For imaging radar (Synthetic Aperture Radar - SAR) data, the DNs represent the strength of a radar pulse returned to the antenna. For digital terrain models (DTMs), the DNs represent terrain elevation. No matter what the source, all these types of data can be stored in a raster format. By applying mathematical transformations to the digital numbers, ER Mapper can enhance image data to highlight and extract very subtle information that would be impossible using traditional manual interpretation techniques. This is why image processing has become such a powerful tool for all types of earth science applications. The exercises in this course provide many examples that illustrate how image processing is typically used to enhance image data and extract information. Many image datasets have multiple bands (or layers) of data covering the same geographic area, each containing a different type of information. For example, a SPOT HRV-XS satellite image has three bands of data, each recording reflectance from the earth’s surface in a different wavelength of light. Since each band records reflectance in a different part of the spectrum, this type of data is often called multispectral data. Many powerful image processing techniques have been developed to combine various bands from multispectral images to highlight specific types of earth science information such as vegetation abundance, water quality parameters, or the types of minerals present at the earth’s surface.

Image Processing - Traditional & Modern:

Digital image processing was developed on large mainframe computers in the 1960s to process images from planetary satellites. To process an image, the operator had to specify the name of the file to process, the type of operation required to be performed, and then wait for the system to process the data and write the results to a new image file on disk.  He then used a separate display program to view the output file and evaluate the results.

With the introduction of powerful workstations in the 1980s, processing of large images could be performed on desktop computers. Processing for digital images involves the manipulation and interpretation of digital images with the aid of a computer. The approach to digital image processing is quite simple. The digital image is fed into a computer one pixel at a time with its brightness value, or its digital number (DN). The computer is programmed to insert these data into an equation, or series of equations, and then store the results of the computation for each pixel.  These results form a new digital image that may be displayed or recorded in pictorial format, or may be further manipulated by additional programs.

The possible forms of digital image processing are literally infinite, the ultimate goal of image processing is to extract as much information about the terrain as possible.  Some important operations related to thematic information identification and extraction, are geometric correction, image enhancement, image arithmetic and image classification.

Image Interpretation:

All image processing and classifying activities are undertaken to lead to some sort of end results. The purpose is to obtain new information, and make informed decisions.  For instance, a Geographic Information System program will require a variety of data that may be gathered and processed simply to answer a question like: "Where is the best place in a region of interest to locate (site) a new school?" Computers and human beings work together to find appropriate answers.

The person or persons involved in image processing and information extraction must have a suitable knowledge base and adequate experience in evaluating data, solving problems, and making decisions. In instances where pattern recognition is a tool to be used in interpretation, the interpreter must also be familiar with the principles and procedures underlying these techniques and have expertise in selecting the right data inputs, the most appropriate method of processing, and be capable of obtaining intelligible results in order to reach satisfactory interpretations and therefore informed decisions.  With the computer age it has also become possible to have software and display programs that make some interpretations. Needless to say that these automated results must ultimately be evaluated by qualified people. As the field of Artificial Intelligence develops and decision rules become more sophisticated, a greater proportion of the interpretation and evaluation can be carried out by the computers. But at some stage the human mind must interact directly.

Image processing for interpretation of landcover features has become an important tool for a wide range of earth science mapping, analysis, and modeling applications. Following are just a few of the many applications for which image processing is commonly used:

• Land use/land cover mapping and change detection

• Agricultural assessment and monitoring

• Coastal and marine resource management

• Mineral exploration

• Oil & gas exploration

• Forest resource management

• Urban planning and change detection

• Telecommunications siting and planning

• Physical oceanography

• Geology and topographic mapping

• Sea ice detection and mapping

Notes & Handouts

The Himalayas

Kumaon Himalayas

Askot Basemetals



This website is hosted by

S. Farooq

Department of Geology

Aligarh Muslim University, Aligarh - 202 002 (India)

Phone: 91-571-2721150