Comprehensive Remote Sensing: Spectral, Data, and Image Analysis Techniques

Spectral reflectance curves and objects identification

A spectral reflectance curve shows how a material reflects EMR across wavelengths.

Each object has a unique spectral signature

Example:

Vegetation: low in red, high in NIR

Water: low reflectance in all bands

Soil: gradual increase with wavelength.

Used to identify and distinguish objects in imagery.

Basic classification of remote sensing on the basis of data collected

Based on:

1. Source of energy: Passive (sunlight), Active (radar, LiDAR)

2. Platform: Satellite, Aircraft

3. Data type: Imaging (photos, raster images), Non-imaging (spectral graphs)

EMR, atmospheric windows, and remote sensing

EMR (Electromagnetic Radiation) carries information

Atmosphere absorbs some wavelengths

Atmospheric windows = regions where EMR passes through

Radiation laws and their applications in remote sensing

1. Wien's Displacement Law

Shorter wavelength = hotter object

Used to determine temperature (thermal sensing)

2. Stefan-Boltzmann Law

Total emitted energy ∝ temperature⁴

3. Planck's Law

Describes spectral distribution of radiation

Used in thermal remote sensing & sensor design.

The birth and advances in satellite remote sensing

Began with Landsat 1

Progress: Higher resolution, More spectral bands, Hyperspectral sensors, Real-time data

Image resolutions

Spatial (pixel size)

Spectral (number of bands)

Temporal (revisit time)

Radiometric (bit depth)

Image resolutions and trade-offs

High spatial → lower temporal/spectral

High spectral → lower spatial

More detail = more data cost

Remote sensing systems for locating cellular phone transceiver

Cell phone locating systems rely on a combination of triangulation, GPS, and spatial data modeling

Justification for the high cost of satellite remote sensing

precision, coverage, reliability, and long-term value

Satellite versus aerial photography

Satellite - large coverage, moderate resolution, high upfront cost, low flexibility

Aerial - small coverage, high resolution, lower cost, high flexibility

Calculate brightness levels and range of brightness levels from bits

Formula:

Levels = 2ⁿ Example:

8-bit → 256 levels

10-bit → 1024 levels

Convert DN to binary digits

DN = 13 → binary = 1101

Landsat band for obtaining bathymetric information and masking water from land areas

Bathymetry → Blue band (Band 1)

Landsat band for masking water from land areas

Water masking → NIR (Band 4, water absorbs)

Pushbroom and whiskbroom imaging technology

Pushbroom: Line sensors, no moving parts

Whiskbroom: Scanning mirror

Lidar technology and data collection

Active sensor (laser)

Measures distance using time-of-flight

Produces: Elevation, 3D structure

Off-nadir and nadir viewing of satellite in remote sensing practice

Nadir: straight down

Off-nadir: angled view, increases coverage but causes distortion

Mask and use in remote sensing

Removes unwanted features (e.g., water, clouds)

Image registration and relevance in remote sensing

Image Registration

Aligns images to same coordinate system

Essential for comparison

Fusion technique and usefulness

Combines multiple images (e.g., high spatial + high spectral)

Image preprocessing techniques

Radiometric correction

Atmospheric correction

Geometric correction

Converting DN to radiance

Radiance = gain × DN + offset

Atmospheric effects and remote sensing image interpretation

Scattering → haze

Absorption → signal loss

Reduces contrast

Layer stacking

Combine multiple bands into one dataset

Supervised and unsupervised digital image processing techniques

Supervised: User-defined training data

Unsupervised: Algorithm clusters data

Density slicing

Converts continuous values into categories

Interpretation of imagery visual on the basis of energy interaction with matter

Objects reflect, absorb, transmit EMR differently

Why objects appear in different colors

Based on reflected wavelengths

Example: Leaves reflect green → appear green

FCIR image interpretation

NIR → red

Red → green

Green → blue

Vegetation appears red

FCIR, true color, and pseudo-color

True: natural colors

False: reassigned bands

Pseudo: artificial color mapping

Why we perceive vegetation as green or yellow

Green: reflects green light

Yellow: stressed vegetation

Lidar DTM, DSM, and Intensity Image

DTM: bare earth

DSM: includes buildings/trees

Intensity: reflectance strength

Distinguishing coniferous and deciduous trees in aerial photograph

Coniferous: darker, uniform

Deciduous: lighter, varied texture

Interpretation of NDVI density sliced image

High NDVI → healthy vegetation

Low NDVI → water/soil

Spectral reflectance curve and its usefulness in two features identification

Distinguishes objects like vegetation vs soil

Explain Wein's displacement law

Hotter objects emit shorter wavelengths

Interpret the color of Vegetation in Color Infrared Image

Appears red (high NIR reflectance)

Interpret the color of Urban in Color Infrared Image

Appears cyan/gray

Understand primary and secondary colors to explain why objects may appear white, black, or steel gray in a composite image

White = all bands high

Black = no reflection

Gray = moderate reflection

Ground truth data versus insitu ground reference data

Ground truth: validation data

In-situ: measured on site

spectral reflectance curve of water, street and tree

Water: Low

Street: Moderate

Tree: High NIR

Fusion and Layer stacking definition and uses

Fusion: combines images of the same area but different resolutions to create a new, improved image.

Layer stacking: combines multiple spectral bands into a single multiband image.

Fusion = Improvement

Stacking = Combination

Explain the fusion technique and the benefits for image interpretation

Process of combining two or more images of the same area that have different resolutions to create a single, more informative image.

Example: High spatial resolution panchromatic image (sharp detail, black & white) with Lower spatial resolution multispectral image (color/spectral information)

Instrument/sensor design and function (f/stop, film, aperture, and shutter system)

Aperture: light entry

f/stop: controls brightness

Shutter: exposure time

Biophysical data

Vegetation health

Biomass

Soil moisture

Characterize remote sensing on the basis of radiation detected, source of radiation, information carried by EMR about the target

Remote sensing can be classified based on:

(1) the type of radiation detected (visible, infrared, microwave),

(2) the source of radiation (passive or active), and

(3) the information carried by EMR, including spectral, spatial, radiometric, and temporal characteristics of the target.

Imaging and Non-imaging

Imaging: pictures

Non-imaging: graphs/spectra

Satellite and human vision comparision

Humans: visible only

Sensors: broader spectrum

Response of vegetation, wet soil and water in NIR image (single band)

Vegetation: High

Wet soil: Medium

Water: Low

The particle theory equation and significance in remote sensing

E = hν

Energy proportional to frequency

Spectral reflectance curve diagnoses of low contrast in image

Small reflectance difference

Atmospheric effects and low contrast

Caused by haze, atmosphere

Geoferencing technique and image overlay

Align image to coordinates

Enables overlay

What image resolution would you require for tree census, and monitoring Traffic flow, urban expansion, and flood inundated areas

Tree census: High

Traffic: High temporal

Urban: Moderate

Flood: Moderate

Trade-offs in image resolutions

Higher detail = less coverage

Wavelength regions for B, G, R, and NIR

Blue: 0.45-0.52 µm

Green: 0.52-0.60 µm

Red: 0.63-0.69 µm

NIR: 0.76-0.90 µm

Color assignment in CIR image

NIR → red

Red → green

Green → blue

What is remote sensing?

Acquiring information about an object without physical contact, using electromagnetic radiation (EMR).

What are the two types of energy sources in remote sensing?

Passive (Sun) and Active (sensor emits energy like RADAR, LiDAR).

What happens during atmospheric interaction in remote sensing?

Scattering and absorption modify the incoming signal.

What are the outcomes of surface interaction in remote sensing?

Reflection, absorption, and transmission.

What does sensor detection measure in remote sensing?

It measures radiance (Lλ).

What are the steps in the remote sensing process?

Energy source, atmospheric interaction (incoming and outgoing), surface interaction, sensor detection, processing, and interpretation.

What are the advantages of remote sensing?

Large spatial coverage, repeat observations, multi/hyperspectral capability, and access to remote areas.

What are the limitations of remote sensing?

Atmospheric distortion, resolution trade-offs, calibration required, high costs, and not a complete data source.

What is in-situ data used for in remote sensing?

Calibration and validation.

What are collateral data in remote sensing?

Data such as DEMs, soil maps, land use, and census information.

What does the wave model of electromagnetic radiation describe?

The relationship between speed of light (c), wavelength (λ), and frequency (ν).

What does a longer wavelength indicate in electromagnetic radiation?

A lower frequency.

According to the Stefan-Boltzmann Law, how does temperature affect energy emission?

Total energy is proportional to temperature raised to the fourth power (T⁴).

What are the three outcomes of energy-matter interaction?

Reflection (ρ), absorption (α), and transmission (τ).

What are spectral reflectance curves?

Identity fingerprints of materials showing how they reflect different wavelengths.

What causes Rayleigh scattering?

Small particles scatter light, affecting blue light and causing a blue sky.

What is the effect of haze on image contrast?

Haze reduces image contrast.

What is the significance of atmospheric windows in remote sensing?

Regions where energy passes through the atmosphere, allowing remote sensors to operate effectively.

What is the role of preprocessing in digital image processing?

It includes geometric, radiometric, and atmospheric corrections.

What does radiometric resolution refer to?

The sensitivity of a sensor, determined by bit depth.

What is spatial resolution in remote sensing?

The size of a pixel on the ground; smaller pixels yield higher resolution.

What is temporal resolution?

The frequency at which a sensor revisits the same area.

What is the importance of change detection in remote sensing?

It allows for multi-date comparisons to identify changes over time.

What are the elements of image interpretation?

Tone/color, shape, size, texture, pattern, shadow, site, and association.

What is NDVI used for in remote sensing?

To assess vegetation health.

What is a common exam trap regarding digital numbers (DN)?

DN values are not comparable across images without conversion to radiance.

What does supervised classification involve?

Using user-defined training data to classify images.