Page 114 - Digital Analysis of Remotely Sensed Imagery

CHAPTER 3

Storage of Remotely

Sensed Data

o matter whether they are directly downloaded from the
Internet site of a data vendor or scanned from analog materials,
Nremotely sensed data must be stored digitally in a certain
format before they can be imported to an image analysis system for
processing. Besides, the analyzed results must be saved in a certain
format appropriate for them to be integrated with data from another
source for further analysis. This chapter on storage of remotely sensed
data consists of four sections. The first section is devoted to space
needed and formats for storing multispectral remote sensing data in a
computer. This discussion is followed by a survey of various data storage
media, and generic graphic formats in which images are commonly
stored. With the emergence of hyperspectral remote sensing data, a
huge storage space is required to store both raw and intermediate results
during data processing. Thus, the image analyst faces the problem of
how to compress these data to a manageable level prior to, during, and
after image processing. Included in the last section is a brief review of
common data-compression methods and their major characteristics.

3.1 Storage of Multispectral Images

3.1.1 Storage Space Needed
Storage space is measured in bytes (8 bits), kilobytes (kb) or 1000 bytes,
megabytes (Mb) or 1 million bytes, gigabytes (Gb) or 1 billion bytes, and
terabytes (Tb) or 1 trillion bytes. The space needed to store a multispectral
image is affected by several variables, including the physical size of the
image (e.g., number of rows r and columns c), the number of spectral
bands it has (b), and its radiometric resolution or quantization level q. So
the total space D required is calculated as below:

D r c b q (3.1)
85