Lidar Accuracy
-
Upload
oscar-riveros -
Category
Documents
-
view
241 -
download
1
description
Transcript of Lidar Accuracy
AIRBORNE 1’ S BIMONTHLY LIDAR PUBLICATION
THE LIDAR NETWORK DECEMBER 2002 / VOL.. 1 NO. 4
LiDARing the World Everything you need to know before surveying abroad Baseline Length in LiDAR Surveys Achieving Profitability and Keeping Accuracy LiDAR & Forestry Changes to the forestry market are right around the corner
CONTENTS
LIDAR ACCURACYAN AIRBORNE 1 PERSPECTIVE
Airborne 1 Corporation
300 N. Sepulveda Blvd.
Suite 1060El Segundo, CA 90245
USA Phone: 310.414.7400Fax: 310.414.7409 E-mail: [email protected]
T H E L I D A R N E T W O R K
The Link [‘(th)& ‘li(ng)k] nl.: your connection to the LiDAR industry
link [‘li(ng)k] n.: a connecting structure, element or factor
Send your e-mail to [email protected]
and ask for your FREE SUBSCRIPTION to our
newsletter, THE LINK!
This publication is the second in a series of Briefing Notes we will be releasing in the coming months. If you’d like to subscribe to this series,
send an e-mail to [email protected] with the subject line “subscribe”.
Get linked to
Flexible. Accessible. Affordable.
We’re on the web! www.airborne1.com
ABOUT AIRBORNE 1 CORPORATION Airborne 1 Corporation provides flexible access to advanced LiDAR technology for partners in the photogrammetry, surveying and mapping fields. Airborne 1's digital mapping services and solutions include a dedicated airborne LiDAR survey group operating Optech ALTM sensors; flexible fractional ownership plans for partners without dedicated access to LiDAR technology; LiDAR data processing, analysis and application development; LiDAR field survey coordination and project management. Airborne 1 was incorporated in 1998, the same year its management team took home the coveted USC MBA "Best Business Plan" award.
Airborne 1 Corporation
PHN. 310.414.7400 / FAX 310.414.7409
300 N. SEPULVEDA BLVD. #1060, EL SEGUNDO, CA 90245
BN#01 Briefing Note
Airborne 1 Corporation
B A C K G R O U N D
Special points of interest:
• An overview of the issues affecting lidar accuracy.
• Discussion of range, position, and orienta-tion errors.
• Results from our last 5 rojects
dar and mapping andards.
I N S I D E T H I S B R I E F I N G :
Background 3
Technical 3
Specifications 5
Project #1 6
Laser Error 8
Project #2 8
GPS Error 11
IMU Error 12
Project #3 12
Project #4 15
Standards 16
Project #5 19
Summary 21
References 21
T E C H N I C A L C O N S I D E R A T I O N S : E R R O R B U D G E T F O R L I D A R
Lidar DSM of San Francisco by Airborne 1.
Lid ar Accur acy Page 1
Flexible …
Accessible …
Affordable
A i rbo rn e 1 Co r po r ati o n Briefing Note
BN#01
solution and the IMU orientation solution. Contributions to the error budget include such diverse factors as the inherent pointing error of the laser, inaccuracies
(Continued on page 4)
The error budget for a given airborne lidar mapping system is primarily driven by the contributing error budgets from the core subsystems; the laser rangefinder, the GPS position
(Continued on page 5)
conducted using an Optech ALTM 1225. It is important when discussing lidar accuracy to keep in mind that the theoretical system error based on a rigorous engineering analysis of the system is generally not achievable in the field.
There is growing concern and confusion in the lidar end user community about the achievable accuracy of airborne lidar data and how this accuracy impacts products certified to various established mapping standards. The purpose of this briefing note is to provide background information and discussion points for lidar users in order to better educate them aboutlidar accuracy specifications. A general discussion of system design error budgets is complemented with examples from recent commercial surveys we have
p
• List
Page 1
T E C H N I C A L C O N S I D E R A T I O N S : E R R O R B U D G E T F O R L I D A R
0.5 accuracy”
Br ie f ing Note Page 1
B A C K G R O U
S Y S T E M A C C U R
service providers are currently a source of significant debate
within the community.”
Page 4 Page 5
Operational consideration
such as variations in GPSquality, will significantly affect the final accuracy. addition an end user or contracting agency will befaced with various interpretations – and misinterpretations – of wis meant by the accuracythe lidar data. Service providers are often not clabout how stated accurawill vary under different conditions across the project, for example in arof steep slope, or how thestated accuracy will be quality checked and
in the response time (Continued from page 3)
(latency) of the receiver electronics, view angle mismatch between the transmitter and receiver, sensor mounting biases - which are the small angular misalignments between the laser reference frame and the IMU reference frame - inaccuracies in measuring the lever arm (antenna offset)
budgets for a given lidar sensor it is important to understand that the final accuracy will be limited by the convolution of all contributors to the error budget. System engineers need to balance each subsystems contribution against desired system performance to avoid
rangefinder with sub-centimeter accuracy would not necessarily improve overall system performance if GPS positioning accuracy were not improved; an extremely accurate IMU will not necessarily improve pointing accuracy if scanner angle measurements are still made to only 0.5° accuracy. For an excellent discussion and detailed examples of how each
While the error budget foparticular lidar system cabe reasonably well defineby proper system engineering analysis thatconsiders the inherent inaccuracies of its components, the final operational accuracy thatcan be achieved is generworse than the theoreticalimit and hence more diffto specify and open to interpretation. The accurspecifications quoted by system manufacturers anservice providers are currently a source of significant debate within community, especially
ongst end users of lidata. This is in part due tk of clear definitions oat is meant when
N D
A C Y S P E C I F I C A T I O N S
Br ie f ing Note
s,
In
hat of
ear cies
eas
controlled. Lidar system manufacturers are often vague about the conditions under which system specifications apply and are generally known to quote specifications that are best case or averaged across expected results. At Airborne 1 we feel an educated end user will be able to make better decisions about lidar as a suitable tool for their project. We hope these briefing notes will help.
stating accuracy for lidardata. Different interpretations of common terms, competing claims among stakeholders, along with confusion
r a n d
“extremely accurate IMUs would not necessarily improve pointing accuracy if
scanner angle measurements are still made to only
“The accuracy specifications quoted by LiDAR system manufacturers and
between the GPS antenna phase center and the reference point of the laser output - usually taken to be the center of the output mirror - and the error inherent in recording thescanner angle at the moment of each pulse. In considering error
system parameter contributes to overall lidar system accuracy, see Baltsavias (1999a). For an independently published analysis of lidar accuracy using commercial sensors, see Kraus & Pfeifer (1998). Shrestha et. al. (2000) and Gutierrez et. al. (1998).
between theoretical or operational accuracy specifications also contribute to the lack of acceptance and skepticism regarding lidar accuracy claims. The ASPRS LiDAR Subcommittee is working hard to establish guidelines for calibrating lidar systems and providing commonly accepted definitions of lidar accuracy in an effort to standardize these specifications.
ally l icult
acy lidar d
the
r o a f
(Continued on page 6)
amdalacwh
price/performancediscrepancies. Forexample, a laser
Page 1 Br ie f ing Note
System Accuracy Specifications
A recent volumetric survey conducted using our ALTM 1225 was flown for 0.5 ft RMS accuracy. The survey was flown at 2500 feet and 20,000,000 points were collected covering 5 sq. miles. Ground control was based on 400 control points and kinematic GPS profiles. The results are shown in Table 1 which lists the min, max, mean and standard deviation of the lidar data from
ground control (all units US Survey Feet).
P R O J E C T S T U DY #1
Min -0.58 Max 0.40 Mean -0.07 Stdev 0.16 RMS 0.17
Table 1
Pro
ject
#1
- Err
or H
isto
gram
Nor
mal
ized
to 4
02 C
nt P
ts
0.15
0
0.20
0
0.25
0
0.30
0
nt (Normalized)Page 6
this limit. 2 sigma (95%) or 90% (1.6 sigma) specifications are generally not mentioned.
2. Accuracy specifications are generally taken across the entire scan width of a system despite the fact accuracywill decrease with increasing scan angle; it is common to see the quoted accuracy being the average of the error at min and max scan angles.
3. Accuracy is generally taken in the GPS reference frame so effects of geoid modelingare ignored.
4. Accuracy analysis is generally taken by comparing to known ground control points butdetails of how this was done are generally not included.
(Continued on page 15)
A brief survey of accuracy specifications published current service providers and system manufacturers would reveal specifications of 15 cm vertical and 50 – 100 cm horizontal are common. The published specifications generally provide little information on how this accuracy is measured including such critical information as operating altitude, full or limited scan angle, target type and slope, GPS quality required to reproduce etc. Some notes to keep in mind when discussing lidar accuracy:
1. Manufacturer’s accuracy specifications are derived from statistical sampling of the lidar data and are generally quoted as a 1 sigma spec, meaning ~68% of the data will fall within
0.00
0
0.05
0
0.10
0
Less
-0.5
0-0
.40
-0.3
0-0
.20
-0.1
00.
000.
100.
200.
300.
400.
500.
60M
ore
Dev
iatio
n Fr
om C
ontr
ol (U
S S
urve
y Fe
et)
Cou
P
L A S E R R A N G E F I N D E R E R R O R
Br ie f ing Note
.30
.50
.70
.91.
11.
31.
5
S S
urve
y Fe
et)
Page 8
ra
m
position on transmitted and received pulses or range walk due to spreading of the pulse from a sloped target. In addition, atmospheric affects can impact the accuracy of the laser rangefinder and become significantly more critical at higher altitudes. Proper modeling of these affects is necessary to prevent
Laser rangefinders are mature technology that is reasonably well understood. The engineering requirements, system design parameters and performance envelopes achievable with current technology are well known. The majority of today’s airborne lidar systems are based on diode-pumped solid-state lasers with pulse widths on the order of 10 ns and rise
g s
-0.5
-0.3
-0.1
0.1
0
iatio
n Fr
om C
ontr
ol (Uintroducing additional error times on the order of 1 ns.
ro
ject
#2
- Err
or H
isto
Norm
aliz
ed
to 4
428
Cnt
Pt
in to the error budget. These atmospheric affects are wavelength-dependent so they can vary in magnitude depending on the particular wavelength used in the system. The correction for the refraction and velocity change of light in the atmosphere is given as a spherical range correction
Under normal operating conditions, the range error from a properly calibrated laserrangefinder of this caliber can be expected to be on the order of 5 – 7 cm, independent of altitude. However, proper calibration of the rangefinder requires the system engineer to take in to consideration such parameters as the timing
A recent preliminavey was flown for vey was flown at 3points were collectGround control wapoints and kinematare shown in Table
mean and standard deviatiground control (all units U
P R O J E C T S T
Min -1.39 Max 1.19 Mean 0.04 StDev 0.40 RMS 0.40
Table 2
jitter in the on-board clock(time interval meter), the ato select the same relative
ry engineering and design sur-0.5 RMS accuracy. The sur-000 feet and 150,000,000 ed covering 40 sq. miles. s based on 4500 control ic GPS profiles. The results 2 which lists the min, max, on of the lidar data from S Survey Feet).
U DY # 2
P
0.00
0
0.02
0
0.04
0
0.06
0
0.08
0
0.10
0
0.12
0
Less
-1.3
-1.1
-0.9
-0.7 Dev
Count (Normalized)
(Marini and Murray, 1973) and needs to be calculated at both the platform location (Continued on page 10)
bility
Use of the Global Positioning System (GPS) is an important part of lidar mapping. Air-borne GPS systems are used in lidar to provide positioning information regarding the trajectory of the sensor. When performing lidar map-ping, it is important to have a good understanding of GPS-related errors. GPS errors fall into two broad categories. The first cate-gory is carrier phase GPS positioning errors. Differen-tial carrier phase positioning gives centimeter-level accu-racy and differential code positioning gives meter-level accuracy. Sources of error include satellite geometry (PDOP), the number of satel-lites, orbital biases, multi-path, antenna phase center modeling, integer resolution, and atmospheric errors. At-mospheric errors consist of either tropospheric or iono-spheric errors. Compound-ing some of these errors is the distance from the ground GPS stations to the lidar sen-sor located in the aircraft. Another area of GPS-related errors falls in the category of network or ground control surveys. These error sources all have an impact on the accuracy of lidar products. One example of GPS-related errors is the accuracy of ge-oid height models. GPS heights are relative to an
ellipsoid, or mathefigure of the earthengineering and mprojects are refereorthometric heighe. elevations), so Grequiring orthomeneed to perform geling. Empirical ermates using curreheight models proNGS show large dusing single-tie geing techniques. Nthe current geoid model (GEOID99) cision of ±5.2 cm km distance, ±5.5a 10-km distance,cm (1s) at a 100-ktance (Smith 2000vertical GPS error,geoid height modedirectly influence racy of any lidar p It is important to rthat for every GPSerror source, a mebe employed to denate, or minimize This generally invocluding the right kchecks into the prdetect these dataOne of the checksby Airborne 1 inclutistical comparisoderived DEMs agamatic GPS profilesformed by field su
G P S P O S I T I O N I N G ER
P
L A S E R R A N G E F I N D E R E R R O R
“In general, non-uniform targets with
differences in reflectivity and significant slope
across the footprint introduce uncertainty in to the actual XY
position being referenced by the
return pulse with a corresponding
uncertainty in Z position.”
Br ief ing Note
For small footprint, time-of-flight lidar sensors, the footprint size is an important consideration when considering accuracy. Lidar sensors are not infinitepoint sampling instruments and the complex interaction of the transmitted pulse energy with the target needsto be considered carefully. The return signal from a target surface will be a function of the integrated energy distribution across the footprint weighted by the reflectivity profile of the terrain within the footprint. Increasing slope across the footprint can further complicate this situation compared to a flat target, ascan returns where only part of the footprint intersects the target. In general, non-uniform targets with differences in reflectivity and significant slope across the footprint introduce uncertainty in to the actual XY position being referencedby the return pulse with a corresponding uncertainty inZ position. This situation can be minimized by using time-of-flight systems with the smallest divergence possible, resulting in the smallest footprint on the ground.
Page 10
(Continued from page 8) and the laser footprint. The magnitude of the atmospheric correction is dependent on temperature, pressure, and humidity and to some degree on the altitude above sea level and the latitude. These atmospheric affects are usually minimized - but not eliminated - by incorporating an appropriate atmospheric model in the post-processing of the lidar data. If proper system design considerations are implemented, the laser rangefinder accuracy is the best-defined and smallest contributor to the overall error budget except in cases of low altitude, low scan widths. An excellent discussion on geolocation of laser altimetry data can be found in Hoften et.al. (2000).
A Note on Divergence and Footprint Size The divergence of a laser represent the physical spread of the beam as it propagates. For example an output beam with a spot size of 0.1 cm and a divergence of 0.25 mrad as it exits the sensor will illuminate a footprint on the ground of ~25 cm from an altitude of 1000 m and ~50 cm from 2000 m. Divergence is a physical property of the laser source that can be modified incorporating appropriate optical elements in the transmitter. Typical divergence values for commercial sensors range from 0.25 mrad to 5 mrad.
matical . Most apping nced to an
t surface (i.PS users
tric heights eoid mod-ror esti-nt geoid duced by ifferences oid model-ationally, height has a pre-(1s) at a 5- cm (1s) at and ±6 m dis-). Any
such as ling, will
the accu-roduct.
emember -related thod can tect, elimi-
that error. lves in-ind of ocess to outliers. performed des a sta-
n of lidar-inst kine- per-rveyors.
RO R
“Empirical error estimates using
current geoid height models produced by NGS show large differences using single-tie geoid
modeling techniques.”
Page 11
P
I MU O R I E N TA T I O N E R R O R
Br ie f ing Note
0.1
00
.15
0.2
00
.25
0.3
00
.35
rvey
Fee
t)
(POS/AV™ 510 from Applanix – post-
Knowing the correct orientation of the sensor in
Page 12
processed solution) although some systems perform to a 0.0025º pitch/roll accuracy. A 0.005º angular error corresponds to a 0.17 m positioning error on the ground from 2,000 m, a 0.35 m error from 4,000 m and a 0.52 m error
space is a necessary but notsufficient condition for calculating an accurate transformation from the local sensor reference frame to the Earth-centered reference frame (WGS84). Accurate measurements of the roll, pitch and heading of the platform are required to correctly determine the
A recent high resolution DTM survstudio was flown for highest densit
racy. The survey was flowover 150,000,000 points ering 10 sq. miles. Grouon 90 points and kinemaThe results are shown in the min, max, mean and sthe lidar data from groun
US Survey Feet).
P R O J E C T S T U DY
Min -0.40 Max 0.24 Mean 0.01 StDev 0.10 RMS 0.10
Table 3
#3
- Err
or H
isto
gram
mal
ized
to 8
8 C
nt P
ts
.10
-0.0
50
.00
0.0
5
Fro
m C
ontr
ol (U
S S
ufrom 6,000 m. However there are additional contributions to the angular pointing error tha nerally not dis by sensor ma rers or service pro These include co ns from the sc ubsystem, wh dd to the errorbu to the finite acthean
pointing direction for each laser pulse. In practice, the orientation of the platform isrecorded by an on-board inertial measurement unit (IMU) that is hard mounted to the lidar sensor. While a variety of IMUs are available commercially, a typical specification for the price/performance levels common in most commercial lidar sensors would be 0.005º pitch/roll, 0.008º heading
t are gecussed nufactuviders.
ntributioanning sich will adget due
ey for a Hollywood y, 0.5 RMS accu-n at 2000 feet and
were collected cov-nd control was based tic GPS profiles. Table 3 which lists tandard deviation of d control (all units in
# 3
Pro
ject No
r
0.00
0
0.05
0
0.10
0
0.15
0
0.20
0
0.25
0
0.30
0
-0.3
5-0
.30
-0.2
5-0
.20
-0.1
5-0
Dev
iatio
n
Count (Normalized)
curacy in measuring scanner mirror angle
d non-linear dynamics
(Continued on page 15)
Lid ar Accur acy
IMU Orientation Error
Another recent preliminary engineering and vey was flown for 0.5 RMS accuracy. The suflown at 3000 feet and over 50,000,000 points were collected covering 5 sq. miles. Ground control was based on 90 control points and kinematic GPS profiles. The re-sults are shown in Table 4 which lists the min, max, mean and standard deviation of the lidar data from ground control (all units in US Survey Feet).
P R O J E C T S T U DY # 4
ccuracy Specifications
Proj
ect #
4 - E
rror
His
togr
amNo
rmal
ized
to 8
9 C
nt P
ts
0.00
00
0.02
00
0.04
00
0.06
00
0.08
00
0.10
00
0.12
00
0.14
00
0.16
00
0.18
00
-0.9
0-0
.80
-0.7
0-0
.60
-0.5
0-0
.40
-0.3
0-0
.20
-0.1
00.
000.
100.
200.
300.
400.
500.
600.
70M
ore
Dev
iatio
n Fr
om C
ontr
ol (U
S Su
rvey
Fee
t)
Count (Normalized)
(Continued from page12) from the IMU but ratthe sum of its compoA sensor with a perfeproviding absolute accuracy measuremthe platform orientawould be wasted if tscanner subsystem had an accuracy of 0general, when consifinal achievable accufor the lidar sensor, engineers must consthe entire error budgnot simply focus on IMU.
A
namics in the scanner mirror motion, especially for single-axis as opposed to rotating scanner systems, and laser pointing errors. Many of these effects can be minimized - but not eliminated - by proper system calibration prior to data collection and proper system modeling during post-processing. However, it is important to keep in mind that the angular error budget – the orientation error – is not just derived
(Continued from page 6) It is always prudto request a certsystem calibratioand accuracy analysis from thelidar system manufacturer orservice provider.
5. Accuracy analysis tends to focus on vertical accuracy (Z) and details on how planimetric accuracy (XY) is verified are vague.
Page 1
design sur-rvey was
Min -0.85 Max 0.54 Mean -0.02 StDev 0.29 RMS 0.29
Table 4
Page 15
her is nents. ct IMU
ents of tion he only .5º. In
dering racy
system ider et and
the
ent ifiedn
the
P
M E A S U R I N G U P T O T H E S TA N DA R D S
“Can you do 1’ contour-
interval mapping with your laser?”
Br ief ing Note Lid ar Accur acy
Feet Min -0.577 Max 0.403 Mean -0.067 Stdev 0.159 RMS 0.172
Table 5
of tested elevatiwithin ½ the coninterval, or in thifeet. For this proRMS value is 0.1This equates to 0at the 90% levelconfidence. Thiscalculated by simmultiplying the R1.6449. Clearly,meets and exceevertical requiremNMAS standard,the distribution i Horizontally, ourlaser unit reliablrepeatedly demobetter than 1/50flight altitude, incalibration and tprovided by the manufacturer, OTheir systems arcertified to bette1/1000th, 1/20some higher (“benumber if applicproven to a specImportantly, all tstatements are f(68% +/-) only. Tunit we’re flying thoroughly testeproven at severathese ratios, butor exceeds the 91/30th and/or 1/inch at the publifor 1”:40’ map p1.33’ and 1.0 ferespectively). Thconservatively ca1/3,500th the flias follows: 2000altitude) dividedhorizontal accurayields 0.5714 fesigma. Multiply (Continu
Applying Mapping Standards to LiDAR NMAS: Vertical requirements are that at least 90%
(Continued from page 16)
(Continued on page 17)
chart for each of our five most recent projects. It is our hope that sharing some of that analysis will be at least informative, and may open doors to exchanges of information across providers and users alike. Here’s one example of how our recent project statistics measure up against the applicable NMAS, ASPRS, and NSSDA standards. Project: Fullerton, CA Deliverable: DEM in support of 1’ contours to NMAS standards Control Points/Total Control Points: 403/407 Lidar DEM: TIN values against control points: Vertical in Feet
guidelines and standards toaddress these issues. However, with enough effort and thought, there’s usually a much more helpful answer than “it depends.” If this booklet achieves its design, it will illustrate and substantiate our response to this question, which says: A: “For the vast majority of terrain and vegetation types, our lidar DEM data will support 1’ c.i. mapping which meets or exceeds the1’ NMAS accuracy requirements, and of course those of any larger contour interval. We will not certify data to the more rigorous 1’ c.i. requirements defined by ASPRS Class 1 map accuracy standards. Given the capability of today’s lidar technology, we will also reliably and repeatedlycertify to ASPRS Class 1 standards for a 2’ c.i. map product. At the margin, we find most datasets will in fact support certification at up to ½ meter contours according to the more rigorous standards.” The support for that position comes directly from the statistical analysis on the millions of acres andthousands of flight-lines of lidar data we’ve collected. In this booklet, we’ve posted an analysis
Page 16
Q: “Can you do 1’ contour-interval mapping with your laser?” The question comes across our phone lines with ever-increasing frequency. As with most every lidar service provider, our answer remains unchanged. The problem is, the answers varysignificantly across companies. As a relatively new tool, it is important to remember that lidar represents a different measurement system when compared to photogrammetry. Current mapping standards were developed with classical photogrammetric systems inmind. Certain artifacts can occur with lidar-derived products such as grade-break definition, planimetric accuracy issues, and vegetation removal. It is important for lidar usersto understand when these artifacts occur so that mapping products meet the desired tolerance for the project at hand. To date, there is a lack of specifications or standards for lidar products. Because of the lack of lidar standards and specifications, one must useexisting guidelines to estimate the accuracy of lidar products. At Airborne 1we are committed to supporting ASPRS’s effort to establish and publish
Page 1
“At Airborne 1 we are committed
to supporting ASPRS’s effort to establish and
publish guidelines and standards to
address these issues.”
Page 17
ons be tour s case 0.5 ject, the 72 feet. .283 feet
of was ply
MS error x the data ds the ents of the
assuming s normal.
ALTM 1225y and nstrates 00th the
cluding all est data
ptech. e typically r than 00th , or tter”)
able and ific system. hese or 1-sigma he 1225
has been d and l times still meets 0% test at 40th of an
cation scale roducts (or et, is is lculated at
ght altitude ’ (flight
by 1:3500 cy factor
et at 1-this
ed on page 18)
P
M E A S U R I N G U P T O T H E S TA N DA R D S
Br ief ing Note Lid ar Accur acy Page 1
M E A S U R I N G U P T O T H E S TA N DA R D S
Lidar DSM of San Francisco
Page 18 Page 19
(Continued from page 17) feet. For a system flying at challenging specification to
xam172
.9racevertic
NSt thr than ur level of mple
which yields a horizontal error of 0.421 feet on thisdata example.
2000’ AGL, and carrying a “1/2000th the altitude” horizontal certification, the expected planimetric accuracy is going to be 1.0’, clearly falling outside the specification. Because our horizontal is tested to something better than this, we can plan and fly to generally achieve a 0.4’ horizontal requirement ondeliver on. In our ethe RMS error of 0.times the required 1factor yield an accu0.337’ at the 95% lconfidence. The verequirement of the standard states thadata must be bette0.5958 x the contointerval at the 95% confidence. The sa
times the factor of 1.6449 (to ascertain the 90% confidence level), and he resultant accuracy is 0.940 feet. This falls inside the frequently cited requirement of 1 contour interval, and exceeds the 1/30th of an inch requirement at 40-scale. ASPRS: Class 1 map vertical accuracy requirements are for
rd
The accuracy specifications for planimetric (XY) accuracy as opposed to vertical (Z) accuracy are different for lidar data. The planimetric accuracy is strongly corre-lated to the pointing angle accuracy and due to the an-gular error component in-creases with altitude. Unlike
imagery, planimetric accu-racy for lidar data is generally 2 –5 times worse than verti-cal accuracy.
N O T E ? P L A N I M E T R I C V S . V E R T I C A L A C C U R A C Y A recent DEM surve
was flown for 0.5 Rvey was flown at 30100,000,000 points 10 sq. miles. Groun150 control points afiles. The results arewhich lists the min, deviation of the lidatrol (all units in US
P R O J E C T S
certain projects. We cannot yet do so reliably, repeatedly, and certifiably, and therefore will not certify datasets to this standard at
me. For 100-scale roducts, this ntal requirement is to 1.0’, and therefore dily certified to.
A: This standard on accuracies
reported at the 95% level of confidence, a geometrically more (Continued on page 19)
project meets this vertstandard. For horizontal accuracthis standard again sethe results at a 95% leconfidence. As with althese tests, a normal distribution is assumedAssuming independenin the x and y axis, a faof 2.4477 is used to determine the horizontNSSDA accuracy,
a limiting RMS error of 1/3 the contour interval. The sample project meets this vertical standard, with an RMS error of 0.172 feet for the mass points. Notably, thedata fails to meet the spot elevation accuracy requirement of 1/6th the contour interval (0.167’), so these would have to be compiled from imagery or ground collection methods. Horizontally, though, the limiting RMS error for a 40- scale map (1”:40’) is .40
ple, ’
6 y of l of al
SDA e
ical
ies, eks vel of l
. t errorctor
al
this timap phorizoeasedis rea NSSDrelies
y in support of photogrammetry MS accuracy. The sur-00 feet and over were collected covering d control was based on nd kinematic GPS pro- shown in Table 5 max, mean and standard r data from ground con-Survey Feet).
T U DY # 5
Min -0.40 Max 0.35 Mean -0.02 St Dev 0.14 RMS 0.15
Table 6
Lid ar Accur acy Page 1
Baltsavias, E.P. (1999). Airborne laser scanning: basic rela-tions and formulas. ISPRS J. Photogramm. Remote Sensing 54 (2/3), 199-214. Gutierrez, R. et. al. (1998). Airborne Laser Swath Mapping Of Galveston Island And Bolivar Peninsula, Texas, in Proceed-ings, Fifth International Conference: Remote Sensing for Ma-rine and Coastal Environments, San Diego: p. I-236–i-243. Hoften, M.A. et. al. (2000). An airborne scanning laser altim-etry survey of Long Valley, California. Int. J. Remote Sensing, Vol 21, #12, 2413-2437 Kraus, K., Pfeifer, N., (1998). Determination of terrain mod-els in wooded areas with airborne laser scanning data. ISPRS J. Photogramm. Remote sensing 53 (4), 193-203. Marini, J.W. and Murray, C.W. (1973). Correction of laser range tracking data for atmospheric refraction at elevation angles above 10º. X59173351 NASA Technical Report Shrestha, R.L. et. al. (2000). Airborne Laser Swath Map-ping: Accuracy Assessment For Surveying and Mapping Applications. University of Florida (see http://www.alsm.ufl.edu/pubs/accuracy/accuracy.htm) Smith, D.A. (2000). Gravity and the Geoid at NGS. Pre-sented at the 2000 Geodetic Advisor Convocation. National Geodetic Survey, Silver Spring, MD
In summary, the vertical accuracy we are seeing from our Optech ALTM 1225 based on our last five field projects is as follows:
A C C U R A C Y S U M M A R Y F O R L A S T 5 A I R B O R N E 1 P R O J E C T S
Project Goal (RMS, feet)
RMS (feet)
#1 0.50 0.17
#2 0.50 0.40
#3 0.50 0.10
#4 0.50 0.29
#5 0.50 0.15
Avg 0.22
Proj
ect #
5 - E
rror
His
togr
amN
orm
aliz
ed to
144
Cnt
Pts
0.00
0
0.05
0
0.10
0
0.15
0
0.20
0
0.25
0
0.30
0
-0.5
0-0
.40
-0.3
0-0
.20
-0.1
00.
000.
100.
200.
300.
400.
50M
ore
Dev
iatio
n Fr
om C
ontr
ol (U
S S
urve
y Fe
et)
Count (Normalized)
R E F E R E N C E S
Table 7
Page 21
N O T E S :
Lid ar Accur acy Page 1
FEMA PUBLISHES APPENDIX FOR LIDAR
As part of its Map Modernization Plan, FEMA has developed a specification for the production of elevation data for flood studies using LiDAR systems. This has been incorporated into Appendix A of the new Guidelines and Specifications for Flood Hazard Mapping Partners.
The Appendix includes FEMA's requirements for LIDAR sys-tems to gather the necessary data to create digital eleva-tion models, digital terrain maps and other National Flood Insurance Program products. FEMA is working with the Na-tional Digital Elevation Program and coordinating with the American Society of Photogrammetry and Remote Sensing to support the develop of broad government and industry standards for LIDAR and other advanced remote sensing technologies. As these standards are adopted, FEMA in-tends to replace the current FEMA standard with these broader standards. This guidelines are available at http://www.fema.gov/mit/tsd/dl_cgs.htm.
Para
met
er
Data
Siz
e (ty
pica
l)
Raw
Dat
a –
Dual
Ret
urn
wt I
nten
sity
L
aser
Ran
ge
4
byt
e
Inte
nsity
2 b
yte
L
aser
Ran
ge
4
byt
e
Inte
nsity
2 b
yte
S
can
Angl
e
2 b
yte
T
ime
Stam
p
8 b
yte
D
ata
Size
- Du
al R
etur
n wt
Inte
nsity
2
2 by
te
Proc
esse
d Da
ta –
Dua
l Ret
urn
wt I
nten
sity
3
XYZ
Co-
ordi
nate
3
x 8
byte
= 2
4 by
te
Inte
nsity
2 b
yte
XYZ
Co-
ordi
nate
3
x 8
byte
= 2
4 by
te
Inte
nsity
2 b
yte
Dat
a Si
ze –
Dua
l Ret
urn
wt I
nten
sity
5
2 by
te
Dat
a St
orag
e Re
quire
men
ts (2
5,00
0 Hz
, dua
l ret
urn
with
inte
nsity
) P
er H
our o
f Dat
a C
olle
ctio
n (“
Butto
n O
n Ti
me”
)
25,0
00 H
z x
3,60
0 s
x 52
byt
e =
4
.57
GBy
te
Per
Pro
ject
Are
a (a
ssum
ing
1 pu
lse
per 1
0 fe
et2 (~
1 m
2 ))
100
sq.
mile
13.
16 G
Byte
1
00 s
q. k
m
5.0
8 G
Byte
1
00 a
cre
2
0.50
MBy
te
Per
Acr
e
205.
00 K
Byte
LID
AR
DA
TA
ST
OR
AG
E R
EQ
UIR
EM
EN
TS
Page 23