1
Digital Data and Binary Numbers
1.Digital Data
2. Binary numbers
3. How digital data is displayed on the monitor
4. Color
5. Data formats
2
Digital Remote Sensing Data
-Matrix of rows and columns
-Each pixel has a value
-corresponds to brightness range of the color in each array
3
Binary Numbers
101001
4
193= 100+90+3
193 is really {(1*10*10)+(9*10)+(3*1)} =100+90+3
Binary numbers are base two digits
3 9 1
Ones Tens Hundreds
3 9 1
10
0
10
1
10
2
Binary numbers are base two digits
2
0
2
1
2
2
5
7= 4+2+1
Binary numbers are base two digits
1 1 1
2
0
2
1
2
2
3 = 0 + 1 + 1
1 1 0
2
0
2
1
2
2
Binary number= 111
Binary number= 11
- Each column represents one bit
- 8-bit image stores 8 bits for each pixel (from 00000000-11111111)
2
8
= 256 (grayscale)
1 1 1
2
0
2
1
2
2
6
Additive color
-yellow is not yellow, it is a combination of red and green
-Red, Green, and Blue are primary colors, all others can be made from them
True Color
-Adds Red, Green, Blue arrays (each are 8 bits) together
- 24-bit image
7
But data collected from sensors not in 0-255 values, so must be converted.
Signed vs. Unsigned
Single, R4,
Real*4
Long, I4,
Integer*4
Short, I2,
Integer*2
Byte
Other Names,
Abbreviations
& Symbols in
Use
3.403*10
38
-3.403*10
38
32
Single
Precision
Floating Point
(SAR
image)
2,147,483,647 -2,147,483,648 32
Signed Long
Integer
32,767 -32,768 16
Signed Short
Integer
255 0 8
Unsigned
Byte
Maximum
Possible
Value
Minimum
Possible Value
No. of
Bits
Number Type
Name
8
2 ways to convert:
Take numerical range of data for each of the three colors
and create a linear map (equal class intervals)
Problem: if you have outliers, image will be really dark or really light.
1. Linear Mapping
0 255
2 ways to convert:
2. Equalization (equal class frequencies)
Produces image with more contrast
9
3 types of images can be stored and seen on the computer:
1. Color images
a) Natural color images:
Red, green, blue bands represent these
actual colors as we see them
b) False color images:
assign three bands that do not represent actual
green blue and red
c) Pseudocolor image:
one band, values range from 0-265 and levels
of values are assigned a color on a look up table.
3 types of images can be stored and seen on the computer:
2. Grayscale images
-One single band
-has all colors, but colors have same values (127, 127, 127)
10
3 types of images can be stored and seen on the computer:
3. Labeled/classified images
-pixels value represents a tag that indicates a property
-used for landcover, e.g.: 1=ag., 2=water, etc.
Data Formats
11
Data format
Describes the way that data is written to storage
Usually contain:
1. Metadata (description about the data: projection, scan lines,
pixels per line)
2. Image Data
-pixel values of each band, arranged band by band
a) Band Sequential (BSQ)
b) Band Interleaved by Line (BIL)
12
File compression
Lossy: loses some information in compression (J PEG)
-smaller file sizes, easy sharing
-cannot obtain original data
Lossless: preserve all data
-can always obtain original data
- file size may not be compressed that much
Lossy vs. Lossless
8888883333
Lossy: 83
Lossless: 8[6]3[4]
File compression
quadtree: as a 2-D compression scheme
-image must be square and length must be a power of 2.
If many homogeneous areas, file will be compressed
If many non-homogeneous areas are there, file may be much larger
13
System Processing
*System Processing:
need to take info collected from sensors and turn into usable format
-image geometry corrected
determined by orbital height,
direction of motion
-remove noise
-correct for pixel misalignments from sensors
(called geometric correction in Ch.4)
-pixels relocated (projected) (Chapter 4)
14
