Remote Sensing for Forest Resources Assessment

2/20/2024
Biratu, 2024 OBU 1
Oda Bultum University
Institute of Land Administration
Department of Geographic Information Sciences
Course Title: GIS and RS for Forest Resource Assessment
Course code: GISc 4090
Academic Year: 2024
Semester: II
Course inst.: Biratu B.
Oda Bultum University
Institute of Land Administration
Department of Geographic Information Sciences
Course Title: GIS and RS for Forest Resource Assessment
Course code: GISc 4090
Academic Year: 2024
Semester: II
Course inst.: Biratu B.

2/20/2024
Biratu, 2024 OBU 2
UNIT ONE
An Overview of Optical, Microwave,
RADAR, LiDAR, and thermal remote
sensing for forest resources
assessment

Classification of satellite remote sensing system
2/20/2024
Biratu, 2024 OBU
3
 Remote sensing systems can be classified on
two bases:
1) The Source of Radiation
I. Passive Remote Sensing
II.Active Remote Sensing

Classification of…
2/20/2024
Biratu, 2024 OBU
4
2) The spectral regions used for data acquisition
1. Optical remote sensing systems (including visible, near IR and SWIR
systems)
2. Thermal infrared remote sensing systems
3. Microwave remote sensing systems
A. Passive remote sensing systems:
 A passive system generally consists of an array of
sensors or detectors that record the amount of EM
radiation reflected and/or emitted from the Earth’s
surface.

2/20/2024
Biratu, 2024 OBU
5
B). Active remote sensing systems
Make use of active artificial sources of
radiation generally mounted on the
remote sensing platform.
An active system, on the other hand,
emits EM radiation and measures the
intensity of the return signal.

2/20/2024
Biratu, 2024 OBU
6
 Both active and passive sensors can be further
classified as:
1. Scanning sensors: the field of interest
scanned sequentially.
2. Non-scanning sensors: the entire field of
interest is explored in one take.
 Active non-scanning sensor systems include microwave
altimeters, laser distance meters and water depth meters.

2/20/2024
Biratu, 2024 OBU
7
 Active scanning sensor systems include
synthetic aperture radar (SAR), in which
microwave pulses are transmitted by an antenna
towards the earth’s surface and the energy
scattered back to the sensor is measured.

Classification of satellite remote sensing system based
on spectral regions
2/20/2024
Biratu, 2024 OBU
8
Optical Remote Sensing
 The images are formed by detecting the solar radiation
reflected by objects on the ground.
 Optical remote sensing systems mostly make use of the visible
(300 -700 nm), near IR (720 -1300 nm) and shortwave IR
(1300 - 3000nm) wavelength bands to form images of the earth
surface.
 Optical remote sensing systems are classified into the following
types, depending on the number of spectral bands used in the
imaging process.

A. Panchromatic imaging system
2/20/2024
Biratu, 2024 OBU
9
 The sensor is a single channel detector sensitive to
radiation within a broad wavelength range.
 If the wavelength range coincides with the visible
range, then the resulting image resembles a "black-
and-white" photograph taken from space.
 The physical quantity being measured is the apparent
brightness of the targets.
 The spectral information or "colour" of the targets is
lost.
Examples of panchromatic imaging systems are:
 IKONOS PAN
 SPOT HRV-PAN

B) Multispectral imaging system:
2/20/2024
Biratu, 2024 OBU
10
 The sensor is a multichannel detector with a
few spectral bands.
 Each channel is sensitive to radiation within a
narrow wavelength band.
 The resulting image is a multilayer image
which contains both the brightness and spectral
(colour) information of the targets being
observed.

Cont.’…
2/20/2024
Biratu, 2024 OBU
11
 Examples of multispectral systems are:
 LANDSAT- MSS
 LANDSAT -TM
 SPOT HRV-XS
 IKONOS - MS
 ETRSS-1
 ET-SMART-RSS

C) Superspectral Imaging Systems:
2/20/2024
Biratu, 2024 OBU
12
 A Superspectral imaging sensor has many more
spectral channels (typically >10) than a multispectral
sensor.
 The bands have narrower bandwidths, enabling the
finer spectral characteristics of the targets to be
captured by the sensor.
 Examples of superspectral systems are:
 MODIS
 MERIS
 SENTINEL 2

D) Hyperspectral Imaging Systems
2/20/2024
Biratu, 2024 OBU
13
 A hyperspectral imaging system is also
known as an "imaging spectrometer".
 It acquires images in about a hundred or
more contiguous spectral bands.
 The precise spectral information contained
in a hyperspectral image enables better
characterization and identification of
targets.

2/20/2024
Biratu, 2024 OBU
14
 Hyperspectral images have potential applications in
such field as:
 precision agriculture (e.g. monitoring the types,
health, moisture status and maturity of crops),
 coastal management (e.g. monitoring of
phytoplanktons, pollution, bathymetry changes).
An example of a hyperspectral system is:
 Hyperion on EO1 satellite
 AVERIS (Airborne Visible and Infrared
Spectrometer
Cont.’…

Thermal Infrared Remote Sensing Systems
2/20/2024
Biratu, 2024 OBU
15
 Thermal infrared remote sensing systems
employ the mid wave IR (3000-5000 nm) and
the long wave IR (8000-140000 nm)
wavelength bands.
 The imagery here is derived from the
thermal radiation emitted by the earth’s
surface and objects.

2/20/2024
Biratu, 2024 OBU
16
 Thermal images provide information on
temperature of the ground and water
surfaces and objects on them.
Cont.’…
This image cannot currently be displayed.

2/20/2024
Biratu, 2024 OBU
17
 Microwave sensing encompasses both active and
passive forms of remote sensing.
 The microwave portion of the spectrum covers the
range from approximately 1cm to 1m in wavelength.
 Because of their long wavelengths, compared to the
visible and infrared, microwaves have special
properties that are important for remote sensing.
Microwave Remote Sensing Systems

2/20/2024
Biratu, 2024 OBU
18
 Longer wavelength microwave radiation can
penetrate through cloud cover, haze, dust, and all
but the heaviest rainfall as the longer wavelengths
are not susceptible to atmospheric scattering which
affects shorter optical wavelengths.
 This property allows detection of microwave energy
under almost all weather and environmental
conditions so that data can be collected at any time.
Cont.’…

Microwave Remote Sensing Systems
2/20/2024
Biratu, 2024 OBU
19
 Generally operate in the 1 cm to 1 m wavelength band
 Used by Radars b/c Microwave radiation can
penetrate through clouds, haze and dust, making
microwave remote sensing a weather independent
technique.
 Active microwave remote sensing systems provide
their own source of microwave radiation to illuminate
the target object.

2/20/2024
Biratu, 2024 OBU
21
 Because of their long wavelengths, compared to the
visible and infrared, microwaves have special properties
that are important for remote sensing.
 Longer wavelength microwave radiation can penetrate
through cloud cover, haze, dust, and all but the
heaviest rainfall as the longer wavelengths are not
susceptible to atmospheric scattering which affects
shorter optical wavelengths.
Cont.’…

2/20/2024
Biratu, 2024 OBU
22
 This property allows detection of microwave energy
under almost all weather and environmental conditions
so that data can be collected at any time.
 Passive microwave sensing is similar in concept to
thermal remote sensing.
 All objects emit microwave energy of some
magnitude, but the amounts are generally very
small.
Cont.’…

2/20/2024
Biratu, 2024 OBU
23
 A passive microwave sensor detects the naturally
emitted microwave energy within its field of view.
 This emitted energy is related to the temperature and
moisture properties of the emitting object or surface.
 Passive microwave sensors are typically radiometers or
scanners and operate in much the same manner as
systems discussed previously except that an antenna is
used to detect and record the microwave energy.
Cont.’…

2/20/2024
Biratu, 2024 OBU
24
 The microwave energy recorded by a passive sensor
can be:
 emitted by the atmosphere,
 reflected from the surface
 emitted from the surface, or
 transmitted from the subsurface
Cont.’…

2/20/2024
Biratu, 2024 OBU
25
 Because the wavelengths are so long, the energy
available is quite small compared to optical
wavelengths.
 Thus, the fields of view must be large to detect
enough energy to record a signal.
 Most passive microwave sensors are therefore
characterized by low spatial resolution.
Cont.’…

2/20/2024
Biratu, 2024 OBU
26
 Active microwave sensors provide their own source of
microwave radiation to illuminate the target.
 Active microwave sensors are generally divided into two
distinct categories: imaging and non-imaging.
 The most common form of imaging active microwave
sensors is RADAR.
 RADAR is an acronym for RAdio Detection and Ranging,
which essentially characterizes the function and operation of
a radar sensor.
Cont.’…

2/20/2024
Biratu, 2024 OBU
27
 The sensor transmits a microwave (radio) signal
towards the target and detects the backscattered
portion of the signal.
 The strength of the backscattered signal is
measured to discriminate between different targets
and the time delay between the transmitted and
reflected signals determines the distance (or range)
to the target.
Cont.’…

2/20/2024
Biratu, 2024 OBU
28
 Non-imaging microwave sensors
include altimeters and scatterometers.
 In most cases these are profiling devices which
take measurements in one linear dimension, as
opposed to the two-dimensional representation of
imaging sensors.
Cont.’…

Remote Sensing in Forest Resources Assessment
2/20/2024
Biratu, 2024 OBU
29
 Remote sensing is the investigation, acquisition,
and processing of information about an area of
interest without contacting it.
 It has been used in environmental sciences to
capture images of the Earth’s surface acquired
from sensors mounted in the air or space
platforms.

2/20/2024
Biratu, 2024 OBU
30
 These images have been used for mapping the
distribution of forest ecosystems, the 3D
structure of forests, and measure the global
fluctuations in plant productivity across different
seasons.
 Remote sensing has made it possible to
consistently and repeatedly monitor forest
characteristics in qualitative and quantitative
ways.
Cont.’…

2/20/2024
Biratu, 2024 OBU
31
 Such data collection and reporting are a
significant factor that assists in research and
development processes.
 It also makes it easier to integrate forestry
with other agencies.
 Nowadays, remote sensing is applied in
different areas of forest management.
Cont.’…

Forest conversion studies
2/20/2024
Biratu, 2024 OBU
32
 Using remote sensing satellites has helped provide
spatial change maps for the forests.
 It has made it possible to determine, monitor, and
curb forest destructions and encroachment.
 Furthermore, image differencing methods and
logical operations have been adopted to determine
the extent of forest degradation globally.

Mapping
2/20/2024
Biratu, 2024 OBU
33
 Remote satellites have made it possible for geographical
sizes of forests to be determined and maintained.
 The satellites gather data like; forest cover types,
forest density and height maps, tree composition, thinning
volume estimation, etc., which have been crucial in
determining forest land distribution.
 Mapping data is extensive. Hence, it has provided
opportunities for researchers to study such data and
delve deeper into botanical studies.

Quantitative Estimation and monitoring of forest cover
2/20/2024
Biratu, 2024 OBU
34
 Remote sensing measurements have made it
possible to understand the forest ecosystem
functions and processes through various variables.
 Variables such as vegetation chemistry and
moisture, biodiversity, soil characteristics, and
vegetation structure make it possible to classify
forests into specific biomass and communities.

Forest fire damage
2/20/2024
Biratu, 2024 OBU
35
 The use of remote satellite sensors has been
beneficial in identifying and assessing forest ground
fire-damaged areas.
 Apart from this, using satellite sensors enhances the
capability of identifying hot spot areas in the forests
that are annually prone to fires.
 Collecting such data is of importance in forest
management.

Forest surveillance
2/20/2024
Biratu, 2024 OBU
36
Using remote sensing helps in keeping track of
forest resources and how to manage them;
 For example, we can track how the
environmental factors react to specific
forces within the environment, like thermal
energy.
 Day-to-day surveillance of the forests is
critical for determining potential risks and
overall forest health.
 Climatic changes and their impact on the forests’ are
easier to monitor when conducting remote sensing
forest surveillance (investigation).

2/20/2024
Biratu, 2024 OBU 37
Unit Two
Information Requirements for Forest
Resources Management

2/20/2024
Biratu, 2024 OBU 38
 Aerial photographs, primarily digital
orthophotos, are the main source of information
for forest data.
 Orthophotos can be used for manual
classification of compartments and as a
background for a variety of digital map products.
 To some extent, manual measurement and
interpretation of stereo aerial photographs in
digital photogrammetric workstation
Role of remote sensing in providing forest
information

2/20/2024
Biratu, 2024 OBU 39
 Forest data estimation using this method, however,
requires great skill of the photo-interpreter if it is
to replace field work.
 An alternative is to use aerial photos or orthophotos
for a preliminary evaluation of the area before field
work.
 Digital stereo aerial photograph is used for
forestry estimates in the same way as 3D point
cloud from laser data.
 However, the data obtained from aerial photography
does not give as good information about forest
density as laser data does,

2/20/2024
Biratu, 2024 OBU 40
 A drone (Unmanned Aerial Vehicles or UAVs) is being
put into use in the preparation of forest management
plans for individual properties.
 It is so far mainly orthophotos created from
mosaicking drone flyovers that are being used.
 The availability of frequent and free satellite
images has increased.
 In 2014 Landsat 8 was launched and in 2015 the
European Sentinel 2A satellite with 10 m pixels was
launched.

2/20/2024
Biratu, 2024 OBU 41
 Currently there is much ongoing research and testing
on a range of technologies with great potential to
streamline the future measurement of forest.
 An example of this is that laser scanners can also
be placed:
i. on the ground (the name of this technology is
often shortened to TLS for Terrestrial Laser
Scanning), or mounted in a backpack or;
ii. on a vehicle (often abbreviated to MLS for
Mobile Laser Scanning), or even carried in the
hand (also MLS).

2/20/2024
Biratu, 2024 OBU 42
 With TLS and MLS detailed information can be
obtained about tree stem position and shape,
which, inter alia, can be used as reference data
to create estimates from the airborne data

2/20/2024
Biratu, 2024 OBU 43
 Remote sensing provides raw data which can be used
simply as a visual background image, such as it is
done now with Google Earth.
 However, remote sensing data are digital data, and
can be analyzed and manipulated and therefore
converted into map data that give users the
information they need.
 We can categorize map data as being thematic or
continuous.
Products from remote sensing

2/20/2024
Biratu, 2024 OBU 44
 Thematic maps have discrete classes, such as
Land Cover map with the thematic classes of
“forest”, “mire”, “water”, ...
 Continuous variable maps consist of a range of
values for a single phenomenon.
 An example of this is a map of timber
volume, with values ranging from 0 up to
the maximum value.

2/20/2024
Biratu, 2024 OBU 45
 Another common product from remote sensing
data is a change map, which can show the
differences between two or more dates of
remote sensing data.
 Another product from remote sensing data is as
input to visualization;
 the remote sensing data input may be as raw data,
such as from Terrestrial Laser Scanning, or
 it may be the map products that form the baseline
for visualization.

2/20/2024
Biratu, 2024 OBU 46
The relationship between remote sensing and
GIS, forest inventory and forest planning
 A GIS consists of four components, namely;
i. data acquisition,
ii. data storage,
iii. data analysis, and
iv. map production.
 Remote sensing fills the function of “data
acquisition” in a GIS; the remote sensing data input
may be raw data (e.g., images) or processed data
(e.g., map data derived from remote sensing).

2/20/2024
Biratu, 2024 OBU 47
 Forest inventory can be done with manual methods,
but remote sensing plays an increasing role for
providing both wall-to-wall data and as ancillary data
in statistical estimations of forest resources.
 Remote sensing and forest inventory have an
intertwined relationship
 The field of forest inventory is concerned with
techniques and methods for measuring and
estimation of forest resources.

2/20/2024
Biratu, 2024 OBU 48
 Forest inventory data use as reference data to
help interpret remote sensing data (i.e., training
data) or to use in validation of the map or product
from remote sensing (i.e., validation data).
 Remote sensing data are used within the process
of forest inventory, and they are analyzed or
displayed within a GIS.

2/20/2024
Biratu, 2024 OBU 49
 However, the subject of remote sensing includes
not just measuring (i.e., inventory), but also
knowledge of how to acquire and process the raw
remotely sensed in a correct way.
 This may require background knowledge in physics,
statistics, photogrammetry, programming, and
certainly geography.

2/20/2024
Biratu, 2024 OBU 50
 The subject of remote sensing also involves
knowing which remote sensing data source is
best suited for the purpose (i.e., strengths and
limitations), and how to perform and present an
accuracy assessment of the map products from
remote sensing data.

2/20/2024
Biratu, 2024 OBU 51
Remote sensing for global to
individual tree applications

2/20/2024
Biratu, 2024 OBU 52
 The remote sensing data source chosen is dependent upon
the aim and goal of the project, the availability of remote
sensing data, and the cost, among other factors.
 For example, if we consider remote sensing data
acquisition for forest inventory purposes, we need
to consider at what scale we wish to produce
information.
 This can be at the landscape scale and may range
downwards to the individual tree scale.
 The large area, landscape scale coverage that can be
provided by remote sensing (synoptic views) makes it a
useful tool.

2/20/2024
Biratu, 2024 OBU 53
 Aerial photographs have been used between 1930’s
to present for delineating forest stands and
measurement of tree height.
 Satellite data have also played a role in providing
forest information over estates, or whole countries.
 With the current innovation of airborne LiDAR,
which measures tree heights and forest density
with high accuracy, the use of remote sensing
for forest inventory is increasing rapidly.

2/20/2024
Biratu, 2024 OBU 54
 At the individual tree scale, remote sensing technologies
are providing data from both the air and from the
ground.
 The ground-based remote sensing includes terrestrial
laser scanning, and ground-based photogrammetry.
 These sensors may be placed on platform which is a stative,
or may be mobile (e.g., placed on a car or hand-held).
 The level of spatial detail in the remote sensing data, the
accuracy of the map products, and cost effectiveness have
had an influence on whether the data will be useful for
forestry applications

2/20/2024
Biratu, 2024 OBU 55
The growth of remote sensing for forestry purposes
 Developments in the subject and use of remote sensing are growing
exponentially due to:
 the ability to acquire highly accurate and useful 3D data
(from laser, radar, and digital photogrammetry) ;
 access to free open-source remote sensing and geographic
data (with the ability to use free, open-source software)
 economic cost-effectiveness of remote sensing data in
businesses (e.g., forestry);
 increased computing power; and
 Increased knowledge about remote sensing and GIS by the
public with access to sites such as Google Earth, and
personal use of GPS.

2/20/2024
Biratu, 2024 OBU 56
Forest information from remote sensing data
 Different remote sensing data sources are more
or less appropriate given the application.
 To determine certain forest characteristics, you
will need to choose one data source over another,
and therefore should know the possibilities and
limitations of different data sources.
 What can and cannot be determined,
given the available data?

2/20/2024
Biratu, 2024 OBU 57
Optical Data
 When using data in the visible and near-infrared
spectral region, which is what is most commonly
available from optical sensors, the main spectral
components in a forest stand are: -
o sunlit canopy,
o shadowed canopy,
o sunlit ground and,
o Shadowed ground.

2/20/2024
Biratu, 2024 OBU 58
 For example, if using satellite data with visible
and near infrared as well as shortwave infrared
wavelengths, the data from the blue, green, red,
and the two middle infrared bands are highly
correlated with each other, and
 the darkness seen in the pixels in these bands
over forest areas will be mostly a function of the
size of the trees and the stem number, since
shadows are a dominant factor behind the
reflectance.

2/20/2024
Biratu, 2024 OBU 59
 A difference, or ratio, measure between the
Near Infrared band and the red band, (as well as
other combinations of near infrared and the sum
of the other bands), is well correlated with the
photosynthetic activity.
 The thermal band is influenced by factors that
are difficult to control, like moisture, and is
seldom used for forestry studies.

2/20/2024
Biratu, 2024 OBU 60
 The shortwave infrared (SWIR) bands
have been shown to be especially useful
for forestry, since that wavelength region
(around 2 µm) is sensitive to the amount of
shadows, and thus to the size of the trees.
 The SWIR bands are the ones that are
best correlated with forest biomass and
thus also with forest stem volume.

2/20/2024
Biratu, 2024 OBU 61
 The SWIR bands have been shown to be of
significance in forest parameter estimation,
most likely due to their sensitivity to shadow
patterns.
 The driving factors behind the reflectance
from forest canopies are:
─ canopy closure
─ leaf area index
─ tree species composition and;
─ The understory vegetation.

2/20/2024
Biratu, 2024 OBU 62
Radar Remote Sensing of Forest
 Radar is a technique that is barely influenced of
cloud coverage or rain.
 Radar (acronym for RAdio Detection And Ranging)
is an active sensor for detecting, locating, tracking,
and identifying objects even at a considerable
distance.
 The radar system is a sensor transmitting and
receiving electromagnetic energy at micro wave
frequencies

2/20/2024
Biratu, 2024 OBU 63
 Generally, three basic types of radiation are
commonly used in remote sensing.
1. Radiation emitted by the object itself, due to
its material properties and physical conditions,
for example thermal radiation.
2. Diffuse scattering of natural illumination from
an incoherent radiation source, for example
the sun.
3. Backscattered radiation from artificial
coherent sources, for example, radar and laser
systems.

2/20/2024
Biratu, 2024 OBU 64
Microwaves
 Remote sensing can be performed with
electromagnetic waves at a wide spectrum of
frequencies.
 Electromagnetic radiation at frequencies 300 MHz
to 300 GHz are generally considered microwaves,
where “micro” represents “small waves compared to
radio waves”.
 The radar bands include frequencies from the entire
range 3 MHz – 300 GHz,

2/20/2024
Biratu, 2024 OBU 66
 The penetration behavior of microwaves is also
considered as an advantage when imagery of
forests is concerned.
 Optical waves are mainly reflected at the top of
the canopy, while microwaves can often
penetrate the canopy, with the attenuation
(weakening of the signal) being dependent on
wavelength, moisture content and the vegetation
density.

2/20/2024
Biratu, 2024 OBU 67
Illustration of the scattering (penetration) in a forest canopy
with common remote sensing radar bands

2/20/2024
Biratu, 2024 OBU 68
Forest Stratification
Diversity in vertical structure
 The vertical arrangement of vegetation in a forest is as
important to many species as the size of the forest
itself.
 Introduced wildlife species are dependent upon
different vegetative layers in the forest subterranean,
understory, midstory, and canopy layers.
 Each layer offers a unique set of habitat features. Fallen
logs, snags, and cavity trees also add to vertical
structure and enhance biodiversity.

2/20/2024
Biratu, 2024 OBU 70
Tree Classifications
 The position of a tree crown affects how well a tree
grows relative to its closest competitors.
 Trees that get the most sunlight generally grow
fastest.
 Tree crowns are classified as dominant, codominant,
intermediate or suppressed.
 Dominant trees have crowns that rise above
the general canopy level.
 They get full sunlight form above and on all sides.

2/20/2024
Biratu, 2024 OBU 71
 Codominant trees make up the average canopy
level.
 Their crowns receive overhead light but
surrounding trees restrict some sunlight
from the sides.
 Intermediate trees occupy a position underneath
the dominant and codominants below the general
crown canopy.
 They receive sunlight only from directly
above.

2/20/2024
Biratu, 2024 OBU 72
 Suppressed trees:
 receive no overhead sunlight.
 They usually are slow-growing and
weakened.
 Shade tolerant tree species can grow in the
suppressed level of the canopy or in the
understory for many years and then, upon
the death of a tree overhead, they respond
with a spurt of growth to take their place in
the general canopy.

2/20/2024
Biratu, 2024 OBU 73
GIS and RS Application
in
Forestry
Chapter Three

2/20/2024
Biratu, 2024 OBU 74
 Remote sensing satellites are capable of collecting
and transmitting data on different measures and at
varying degrees of detail.
 Satellite sensors with potential applications to
forestry can be grouped into two types:
i. electro-optical and
ii. synthetic aperture radar.

2/20/2024
Biratu, 2024 OBU 75
Electro-optical systems
 Electro-optical (EO) systems are passive sensors
which use an electronic sensor to detect optical
signals but operate only when there is sufficient
light reflected from the target for them to form
an image.
 EO systems typically monitor blue, green, and red
portions (or bands) of the visible spectrum, as well
as near-infrared, and long-wave, or thermal,
infrared bands.

2/20/2024
Biratu, 2024 OBU 76
 The near-IR bands have a wide variety of
applications, including the ability to detect certain
indicators of the health of vegetation.
 Most EO systems operate in two different regimes:
a. panchromatic (Pan), in which imagery from all
portions of the spectrum are recorded as shades
of gray (i.e. producing a “black-and-white” image);
and
b. multispectral (MS), in which the discrete bands
of light are recorded individually to be analyzed
either individually or in composite, producing a
“color” image.

2/20/2024
Biratu, 2024 OBU 77
Synthetic aperture radar
 Synthetic aperture radar (SAR) is an active sensor,
that is, it provides its own illumination and can
therefore operate at night.
 SAR collects information by measuring reflected
microwaves which generally have very good
atmospheric penetration capability, so the satellite
can collect data through clouds and light rain—an
important capability in many forest regions.

2/20/2024
Biratu, 2024 OBU 78
 A notable disadvantage of SAR is that
the imagery shows only those objects
that reflect the microwave signal energy
back to the imaging platform; organic
targets, such as vegetation, are much less
visible than physical structures.

2/20/2024
Biratu, 2024 OBU 79
Lidar Technology in Forestry
 Lidar is an active remote sensing technology used
to measure distances with high accuracy. This
technology provides horizontal and vertical
information at high spatial resolution and vertical
 accuracies, offering opportunities for enhanced
forest monitoring, management and planning.
 Lidar systems for forestry are classified according
to the platforms they are mounted on (airborne,
space-borne or ground based)

2/20/2024
Biratu, 2024 OBU 80
 The most common Lidar systems used today are
small-footprint, discrete return laser scanners
mounted on aircraft (although this is rapidly
changing).
 The laser scanner is used to measure distances to a
target by emitting pulses at rapid frequency, up to
150 kilohertz (kHz), and recording the time it takes
for each pulse to complete the distance from the
scanner to the object and back to the scanner

2/20/2024
Biratu, 2024 OBU 81
 Airborne laser scanning systems have four
major components:
1) a laser scanning unit,
2) a global positioning system (GPS),
3) an inertial measurement unit
(IMU) and
4) a computer to store data.

2/20/2024
Biratu, 2024 OBU 82
 Lidar has been used successfully to:
 capture forest structure,
 map individual trees in forests and
critical wildlife habitat
characteristics,
 predict forest volume and biomass, to
develop inputs for forest fire behavior
modeling and to map forest topography
and infra-structure.

2/20/2024
Biratu, 2024 OBU 83
Forest structure
 the tree density, volume and height characteristics
is critical for management, fire prediction and
wildlife assessment.
 Optical remote sensors such as Landsat do not
provide detailed depictions of forest structure.
There are typically two methods to perform large-
scale forest inventory with small-footprint Lidar
data:
a) at the scale of individual trees and
b) at the stand or plot scale.

2/20/2024
Biratu, 2024 OBU 84
 The ability to delineate individual trees from a Lidar
point cloud has been proven for heterogeneous and
complex forests such as oak savanna and mixed-
conifer stands.
 Delineating the individual trees is done by
segmenting the Lidar-derived canopy height model —
the raster image interpolated from Lidar points
depicting the top of the vegetation canopy.

2/20/2024
Biratu, 2024 OBU 85
 After accurate segmentation, relationships can
be derived between Lidar and field-measured
structural attributes such as tree height,
crown diameter and canopy base height, which
are directly measured, and basal area, diameter
at breast height, wood volume, biomass and
species type, which are derived by correlations.

2/20/2024
Biratu, 2024 OBU 86
Canopy height from LiDAR data
 A requirement for calculating canopy heights using
both discrete return and full waveform LiDAR
data is the ability to identify some ground
reference level below the canopy.
 In the case of discrete return LiDAR data, canopy
height estimates are calculated by taking the
difference between those LiDAR returns not
classified as ground

2/20/2024
Biratu, 2024 OBU 87
 Terrain mapping surveys are typically carried out
using a discrete return scanning LiDAR system.
 Full waveform LiDAR systems are primarily used by
researchers for science applications and have yet to
be truly commercialized.
 For full waveform LiDAR data, canopy heights can be
calculated by converting the elapsed time difference
between the peaks of the two most prominent modes
in the amplitude waveform into range.

2/20/2024
Biratu, 2024 OBU 88
Chapter Four
Remote Sensing and GIS for Forest/tree
Biophysical Parameter Mapping

2/20/2024
Biratu, 2024 OBU 89
Thematic mapping in forestry
Different types of thematic maps could exist in
forestry. Examples of thematic maps in forestry:
 Forest cover map
 Plantation forest cover map
 Natural forest cover map
 Forest tree species distribution map

2/20/2024
Biratu, 2024 OBU 90
 Thematic map can be produced using different
data.
 If the different data have different coordinate
systems and/or datum, the different data should
be projected/transformed to a common
coordinate system.
 There are two types of datum that are used in
Ethiopia including Adindan and WGS84.

2/20/2024
Biratu, 2024 OBU 91
 The thematic maps should have:
 Appropriate contents
 Primary contents (layer that delivers the
massage of the map)
 Secondary / supporting contents
 Appropriate legend, north arrow, scale
(graphical),
 Title
 Label
 Data source
 Date of data acquisition
 datum
 Participated institution / personnel for the
map production
 Contact address

2/20/2024
Biratu, 2024 OBU 92
Assignment 1 (15%)
 Be in a group of five and do the following
tasks!
A. Analyze the land cover types (forest and non-
forest) around (in) west Hararghe zone (Use
landsat 8)
B. Analyze the forest cover change and produce
forest cover change thematic map (use landsat
of the last two decades)

2/20/2024
Biratu, 2024 OBU 93
 Once a digital terrain model (DTM) is generated,
useful information about vegetation “height”
over forested areas can be immediately derived
by subtracting the underlying DTM elevation
from the elevation of each point, a process
called point cloud normalization or DTM
detrending.
Canopy Surface Height Modeling (CHM) and
Mapping

2/20/2024
Biratu, 2024 OBU 94
Type of vegetation height product:
• a raster grid that stores the upper surface
which is called the canopy height model (CHM).
• CHM is essentially 2.5-dimensional (2.5D)
instead of 3-dimensional (3D)
• if a user is interested in constructing the
complete 3D envelope of tree crowns that
includes the upper surface and lower boundary,
more sophisticated techniques such as
“wrapped surface reconstruction”.

2/20/2024
Biratu, 2024 OBU 95
• CHM, as a surface model, by itself carries
important information about the amount and
spatial distribution of vegetation materials over
a geographical area.
• It is also the basis for mapping individual trees
and deriving tree level information (such as
height and crown size)

2/20/2024
Biratu, 2024 OBU 96
 A CHM can be generated via two different
processes:
1. first generate a digital surface model (DSM)
from the original LiDAR point cloud and then
subtract DTM from DSM to derive CHM
2. first create a DTM-detrended point cloud
(with x, y, height values) by subtracting each
point’s Z elevation from its DTM elevation,
and then generate a CHM from the
detrended point cloud.

2/20/2024
Biratu, 2024 OBU 97
 Whatever method is used, interpolation
is needed to generate continuous DSM
or CHM models from discrete points.

2/20/2024
Biratu, 2024 OBU 98
 The main challenge of CHM generation lies in the
fact that the LiDAR sensor does not continuously
measure but just samples the earth surface (i.e.,
LiDAR data acquisition is essentially a sampling
process).
 Therefore, if CHM is generated by simply
searching the laser point of maximum elevation or
height within each cell, it would underestimate the
true maximum height.

2/20/2024
Biratu, 2024 OBU 99
 Another problem with sampling is that
some cells, especially when they are small,
might not have laser points within them,
which appear as “pits” over crowns on a
CHM
 An inaccurate CHM causes problems in the
retrieval of individual trees and crown
attributes especially when crown shapes are
complex and irregular.

2/20/2024
Biratu, 2024 OBU 100
Individual Tree Isolation and Mapping
 In forestry and ecology studies, tree-level
information (e.g., height, crown size, diameter at
breast height (DBH), stem volume, and biomass) is
often needed.
 Airborne LiDAR data of high point density (~5
points/m2 or higher).
 It can produce better results than high spatial
resolution optical imagery because:

2/20/2024
a) all points in a LiDAR point cloud are accurately geo-
referenced in 3D whereas pixels in an image or
photo are projected to 2D with distortion (such as
relief displacement),
b) the basis of detecting trees from LiDAR data is
the 3D shape of the trees whereas photo- or
image-based tree detection replies on the pixel
brightness variations within tree crowns (the latter
is often affected by sun illumination conditions),
and

2/20/2024
• C) LiDAR can provide direct estimates of tree
height whereas it is much more difficult to do so
using imagery.
• Because of these advantages, airborne LiDAR
has gained popularity for tree mapping in the
21st century.

2/20/2024
1. Grid-Based Tree Mapping
 The classical approaches to map individual trees are
based on CHM, a 2D grid or raster of canopy height.
 A common strategy is to identify treetops by searching
local maxima from CHM.
2. Point-Based Tree Mapping
 Another strategy of tree mapping is to group laser points
into clusters that correspond to individual tree crowns.
Such clustering algorithms usually assign a 3D point to its
tree crown based on its proximity to the center of the
tree crown.

2/20/2024
Modeling, Mapping, and Estimating Biomass
 The elevated concentration of carbon dioxide (CO2), as
a greenhouse gas, in the atmosphere is of major
concern to our earth.
 Forests can absorb CO2 in the atmosphere via
photosynthesis and release O2 to the atmosphere via
respiration, the balance of which results in changes of
forest biomass.
 Like other forest attributes, biomass can be estimated
using individual-tree or area-based approaches from
LiDAR.

2/20/2024

Remote Sensing for Forest Resources Assessment

More Related Content

What's hot (10)

Similar to Remote Sensing for Forest Resources Assessment (20)

Recently uploaded (20)

Remote Sensing for Forest Resources Assessment