SlideShare ist ein Scribd-Unternehmen logo
1 von 75
An Image-based Disdrometer
Verification and Raindrop Analysis
影像式雨滴譜儀系統驗證與雨滴分析
指導教授: 鐘太郎 教授
學生: 102061539 黃政翰
Outline
 Introduction
 Raindrop Analysis Theories
 Structure of the Proposed Disdrometer and Software Algorithm
 Result of Experiments
 Conclusion
Introduction
Precipitation Observation
 Weather forecasting must tell the information of precipitation
 Behavior of rainfall phenomena are due to local and sudden precipitation
 Predicting rainfall intensity
 Advance preparation can prevent potential disasters from happening
Observing Systems
 Primitive method
 Collect drops by a box with dye paper
 Wasting time and low efficiency
 Radar and Satellite sensors
 Collect data in large region
 Not accurate enough in analyzing rather small region
 Disdrometers
 Analyze raindrop particles
Why Disdrometer?
 Increase the accuracy of the raining condition in a small region
 Increase the accuracy of the radar measurement
 Raindrop feature analysis can be used in
 Air traffic control
 Scientific examination
 Weather observation system
Types of Disdrometers
 Joss-Waldvogel Disdrometer
 Acoustic Disdrometer
 Optical Particle Size Velocity Disdrometer
 2-D Video Disdrometer
 Image-based Disdrometer
Joss-Waldvogel Disdrometer
 Discriminate drop size by receiving the impact kinetic energy
 It cannot determine the drop shape
 Low sensitivity in drizzling and heavy rain
Acoustic Disdrometer
 Drops hitting on the sensor induces sonic wave
 The piezoelectric sensor can measure the rainfall intensity
 It does not provide raindrop distribution data
 Wind influences the measurement easily
Optical Particle Size Velocity Disdrometer
 A row of laser beam points to a sensor
 Measure drop size by calculating the duration of light extinction
 2 dimension is possible; speed measurement is available
 Drop mismatch makes errors
2-D Video Disdrometer
 Using 2 line-scan cameras to measure size and shape
 Velocity can be calculated by traveling time and distance between two frames
 Slanted particle falling path makes image distortion and lead to errors
Image-based Disdrometer
 Use CCD camera to capture raindrops
 Double exposures in each frame
 Too many drops in one image influences matching accuracy a lot
Disdrometer Comparison
Disdrometer Type Measuring Mechanics Advantages Disadvantages
Joss-Waldvogel Falling impact
Good performace in
small size variation
Poor at measuring
drops that are too
small or too large
Acoustic Inducing sonic wave
Good at monitoring
large size drops
Wind influence easily
OTT Parsivel
Make drops falling through
a laser beam
With good accuracy in
measuring drop size
Drop mismatch and
near drops makes
error
2D Image
Use 2 line-scan camera to
measure size and shape
Size, shape and
velocity measurement
are available
Slanted falling drops
are distorted
Image-based
use CCD camera to
capture images and find
parameters by image
processing
Low cost of recording
and flexible system
setting
Camera frame rate and
resolution influence the
result
Raindrop Analysis Theories
Drop Size Distribution (DSD)
 Marshall and Palmer (1948) [11]
 Announced that DSD can be described by an exponential distribution
D
D eNN 
 0
4
0 08.0 
 cmN
121.0
41 
 cmR
Drop Size Distribution (DSD)
 The relationship comes up with errors in small drops
 Ulbrich (1983) [12]-> Gamma function
 are the parameters
D
D eDNN 
 
0
,, 0N
Drop Size Distribution (DSD)
 The relationship comes up with errors in small drops
 Feingold and Levin (1986) [13]-> Lognormal function
 are the parametersTg ND ,,
DSD is related to
 DSD has some relations with rain rate and rain types
 Kozu and Nakamura (1991) [14]
 DSD is related to reflectivity factor measured by radar
 Doviak and Zrnic (1984) [15]
 DSD can calibrate and increase the accuracy of the radar
 Joss and Waldvogel (1969) [16]
Drop Velocity
 Gunn and Kinzer (1949) [17]
 Experiments of raindrop terminal velocities through stagnant air
 Battan (1964) [18]
 Experiments in thunderstorm
 Foote and DuToit (1969) [19], Beard (1976) [20]
 Experiments in different air density
Drop Velocity
Different velocity curve under different conditions Different velocity curve under different air density
Atlas et al. [5]
Mitchell [21]
Beard [20]
Structure of the Proposed Disdrometer and
Software Algorithm
System Structure
 Optical Unit
 Light source (Part A)
 Image Acquisition Unit
 Lens (Part B)
 CCD camera (Part C)
 Data Processing Unit
 Processing Algorithm (Part D)
System Structure Diagram
System Structure
System Structure Photograph
Optical Unit
 Viswell HBL-100
 Uniform blue LED light source
 Bring up light intensity
 Enhance image contrast
Optical Unit
 Relative position between light source and camera
 It depends on the lens used in the system
 Using a telecentric lens
 We can only put the light source facing to the camera
Optical Unit
Light source on the side Light source facing to the camera
Optical Unit
 Proper adjustment of light intensity
Light Intensity 50 52.5 55 57.5
Contrast under
0 degree light
0.241 0.232 0.053 0
Image Acquisition Unit: Camera
 CCD Camera: Pylon Basler Aca640-90gm
 Monochrome, adjustable gain and exposure time
 High frame rate (90 fps in stable)
 659 pixels * 494 pixels
 SDK is provided
Image Acquisition Unit: Lens
 Lens: OPTO Engineering TC13064
 Telecentric lens
 Minimizes blur effect
 Long depth of field
Image Acquisition Unit: Lens
 Compare with the old one: Computar M1214-MP2
TC13064 M1214-MP2
FOV 6.5cm*4.8cm 5cm*3.7cm
DoF about 15cm about 3~5cm
Focus Fixed Manual adjustable
Iris Fixed Manual adjustable
Out of focus
Blurriness
Slight Severe
Image Acquisition Unit: Lens
Image taken by M1214-MP2 Image taken by TC13064
Data Processing Unit
 Software platform: Visual Studio 2012, using Visual C++
 Combined with: OpenCV 2.4.9, Pylon 4 SDK, Matlab2010
Data Processing Unit:
Camera Parameter Setting
 Done before taking every set of images
 Critical parameters
 Image size
 Exposure time
 Gain
 Recording duration
Data Processing Unit:
Camera Parameter Setting
Set the duration time of recording
or the number of images taken
Set the exposure time
Set the gain
Set the image size
Start taking images
Start the program
Data Processing Unit:
Drop Extraction and Analysis
Drop Extraction and Analysis:
Make a Background
 Background calculation: average all the frames
 is the ith frame, is the number of taken images
k
if
Bg
k
i


][
][if k
k
average
Drop Extraction and Analysis:
Make a Background
Original image Image without background
Drop Extraction and Analysis:
Image Binarization
 Median Filter 5*5 -> reduce noise
 Choose proper threshold method
 Depend on the image we get
 Max Entropy, Iterative, Otsu, Region Growing, Level Set had been tried
 100us exposure time image: Max Entropy Thresholding
 2000us exposure time image: Iterative Thresholding
Drop Extraction and Analysis:
Max Entropy Thresholding
 The threshold determines by maximizing the entropy of foreground and
background
 is the gray-level probability density function for the image
)
)(
)(
log(
)(
)(255
1 Tq
ip
Tq
ip
H
fTi f
f  
)
)(
)(
log(
)(
)(
0 Tq
ip
Tq
ip
H
b
T
i b
b  
)(ip
Drop Extraction and Analysis:
Max Entropy Thresholding
 and are the probabilities that a given pixel belongs to foreground or
background when the threshold is


255
1
)()(
Ti
f ipTq


T
i
b ipTq
0
)()(
)(Tqf )(Tqb
T
)max( bf HHT 
Drop Extraction and Analysis:
Iterative Thresholding
 An initial threshold is chosen, typically the average intensity of the image
 Mean gray value of foreground and background are calculated
 is the gray-level probability density function for the image
T


255
1
)(
Ti
f iip


T
i
b iip
0
)(
2
bf
T
 

)(ip
Drop Extraction and Analysis:
Binary images after thresholding
100us binary image 2000us binary image
Data Processing Unit:
Drop Extraction and Analysis
Drop Extraction and Analysis:
Contour Finding
 Binary images have high contrast
 Easy to determine edges
Drop Extraction and Analysis:
Ellipse Fitting
Drop Extraction and Analysis:
Minimal Bounding Rectangle
Drop Extraction and Analysis:
Unsuitable Objects Elimination
 Out of bound elimination
 Any contour touches the border are eliminated
 100us images
 Axis ratio 0.4~1.2 -> treated as raindrops
 2000us images
 Eliminate if the width is larger than height
Drop Extraction and Analysis:
Size Calculation
1.01.0  PA
2
1
)(2

A
Dm 
mm1.0
mm1.0
P
Drop Extraction and Analysis:
Velocity Calculation
2
2
w
hd 
tdv / h
w
2
w
2
w
Drop Extraction and Analysis:
DSD calculation
 According to Liu et al. (2013)
 : DSD
 : number of drops of each bin
 : bin interval (1mm)
 : Sampling volume of the drop-falling space
)(DN )( 13 
mmm
)(DNum
dD
)(15.0 3
mWHV 
m
m
m
dDV
DNum
DN


Pr
)(
)(
framemymx
osuremyosuremx
tvv
tvhHtvwW



)()(
Pr expexp
Drop Extraction and Analysis:
Rain Rate Calculation
 : velocity of the measured diameter



0
3
)(
6
mmmm dDvDDNR

mv
Result of Experiments
Marble Experiments
 Throwing marbles from 1mm to 5mm separately
 100us exposure time, 300 gain, maximum light
 3 seconds duration, 270 frames, as one set of images
Marble Experiments
Measured diameter vs Calculated Diameter Measured diameter vs Measured Area
Marble Experiments
Theoretical Value Average Std Min Max Error
diameter(100us) (mm) 0.25 0.3110 0.0902 0.0707 0.4998 24.3815
area(100us) (mm
2
) 0.0491 0.1317 0.0511 0.055 0.215 168.1602
Theoretical Value Average Std Min Max Error
diameter(100us) (mm) 1 1.0219 0.2217 0.5 1.4948 2.1872
area(100us) (mm
2
) 0.785 0.7369 0.4036 0.09 1.9 6.1687
Theoretical Value Average Std Min Max Error
diameter(100us) (mm) 2 2.0560 0.2808 1.5 2.4749 2.8018
area(100us) (mm
2
) 3.142 3.2074 1.195 0.535 5.305 2.0950
Theoretical Value Average Std Min Max Error
diameter(100us) (mm) 3 3.0887 0.2514 2.5 3.4883 2.9583
area(100us) (mm
2
) 7.069 7.4855 1.5203 1.62 10.735 5.8988
Theoretical Value Average Std Min Max Error
diameter(100us) (mm) 4 4.0374 0.2346 3.5 4.4721 0.9339
area(100us) (mm
2
) 12.566 12.4727 2.2965 4.225 17.085 0.7456
Theoretical Value Average Std Min Max Error
diameter(100us) (mm) 5 4.9206 0.3012 4.5 5.4447 1.5886
area(100us) (mm
2
) 19.635 18.1703 3.5536 5.985 24.27 7.4594
Marble Experiments
 Overlapping leads to the presence of outliers
 Larger marbles have higher error
 Axis ratio recognition gives larger range to be distinguished in large size objects
 Small marbles error
 Some noises are remained after thresholding
Water Sprinkling Experiments
 Spread water by sprinkler
 100us exposure time, 300 gain, maximum light
 2000us exposure time, 300 gain, half light
 5 seconds duration, 450 frames, as one set of images
Water Sprinkling Experiments
Measured diameter vs Calculated Diameter Measured diameter vs Measured Area
Water Sprinkling Experiments
Axis ratio distribution
Water Sprinkling Experiments
Diameter vs Speed Canting angle histogram
Water Sprinkling Experiments
Theoretical Value Average Std Min Max Error
diameter(2000us) 0.25 0.324 0.096 0.045 0.500 29.7393
speed(2000us) 0.7847 0.643 0.530 0.069 2.705 18.0382
diameter(100us) 0.25 0.456 0.045 0.300 0.499 82.2501
area(100us) 0.0491 0.136 0.045 0.005 0.235 177.8743
Theoretical Value Average Std Min Max Error
diameter(2000us) (mm) 1 0.9795 0.2773 0.5 1.4997 2.0474
speed(2000us) (m/s) 3.9972 2.1910 0.8306 0.05 8.6185 45.1867
diameter(100us) (mm) 1 0.7777 0.1927 0.5 1.4977 22.2321
area(100us) (mm2
) 0.7854 0.4596 0.2566 0.02 1.795 41.4878
Theoretical Value Average Std Min Max Error
diameter(2000us) (mm) 2 1.8524 0.2785 1.5 2.4989 7.3819
speed(2000us) (m/s) 6.5477 3.4311 1.1124 0.1 10.454 47.5979
diameter(100us) (mm) 2 1.7033 0.2171 1.5 2.4660 14.8353
area(100us) (mm2
) 3.1416 1.6046 0.5596 0.475 3.685 48.9244
Theoretical Value Average Std Min Max Error
diameter(2000us) (mm) 3 2.8973 0.2750 2.5 3.4985 3.4220
speed(2000us) (m/s) 7.9474 4.3873 1.4764 0.75 10.8093 44.7961
diameter(100us) (mm) 3 2.8418 0.0189 2.82843 2.8552 5.2725
area(100us) (mm
2
) 7.0686 2.97 0.9334 2.31 3.63 57.9831
Theoretical Value Average Std Min Max Error
diameter(2000us) (mm) 4 3.8909 0.2694 3.5 4.4933 2.7287
speed(2000us) (m/s) 8.7156 5.1360 1.9515 0.75 12.8701 41.0717
diameter(100us) (mm) 4 N/A N/A N/A N/A N/A
area(100us) (mm2
) 12.5664 N/A N/A N/A N/A N/A
Theoretical Value Average Std Min Max Error
diameter(2000us) (mm) 5 4.9799 0.2777 4.5 5.4811 0.4011
speed(2000us) (m/s) 9.1372 5.9424 2.1716 1.4 11.4378 34.9649
diameter(100us) (mm) 5 N/A N/A N/A N/A N/A
area(100us) (mm2
) 19.6350 N/A N/A N/A N/A N/A
Average Std Min Max
canting angle -35.6758 26.6744 -180 0
Water Sprinkling Experiments
 Larger error of speed difference in larger drop size
 Overlapping issue
 Sprinkled water are not in terminate velocity
 Few drops are grabbed in this size interval
Raining Experiments
 Real raining condition at 17:00, 27 Aug 2015 at Hsinchu, Taiwan
 5 seconds duration, 450 frames, as one set of images, 30 seconds in total
 2700 images taken in 100us and 2000us respectively
Raining Experiments
Histogram of size distribution Axis ratio distribution
Raining Experiments
Measured diameter vs Calculated Diameter Measured diameter vs Measured Area
Raining Experiments
Histogram of size distribution Histogram of canting angle distribution
Raining Experiments
Diameter vs Speed
Raining Experiments
Theoretical Value Average Std Min Max Error
diameter(2000us) (mm) 0.25 0.374864 0.108827 0.14 0.497947 49.9457
speed(2000us) (m/s) 0.7847 1.15072 0.598629 0.28 3.2 46.6446
diameter(100us) (mm) 0.25 0.441055 0.065593 0.31305 0.494975 76.4218
area(100us) (mm2
) 0.0491 0.137188 0.060991 0.025 0.215 179.4043
Theoretical Value Average Std Min Max Error
diameter(2000us) (mm) 1 0.9394 0.2447 0.5 1.4999 6.0578
speed(2000us) (m/s) 3.9972 2.6945 0.8629 1.05 5.3062 32.5902
diameter(100us) (mm) 1 0.8658 0.2196 0.5 1.4863 13.4154
area(100us) (mm2
) 0.7854 0.6174 0.3419 0.055 1.87 21.3847
Theoretical Value Average Std Min Max Error
diameter(2000us) (mm) 2 1.7479 0.1486 1.5402 2.0803 12.6055
speed(2000us) (m/s) 6.5477 5.3268 0.5195 4.0784 6.0691 18.6465
diameter(100us) (mm) 2 1.9637 0.1418 1.5 2.2030 1.8129
area(100us) (mm
2
) 3.1416 3.0464 0.3600 1.695 3.695 3.0294
Average Std Min Max
canting angle -10.276 29.3070 -180 -0.273
Raining Experiments
 Small raindrops dominant
 Small canting angle -> almost no wind
 Speed are lower than theoretical value
 Image processing leads to the error
 Raining condition difference
Raining Experiments:
Image Processing Error
 If there is one-pixel error in width
 An 1mm raindrop is in 10% error
 Theoretical velocity is in 9% error
Raining Experiments:
Image Processing Error
Raining Experiments
 According to the statistical data of Central Weather Bureau
 Rain rate = 0.5 mm/h
 Wind speed = 0.3 m/s
 The calculated data
 Rain rate = 0.5721 mm/h, Error = 14.42%
 Wind speed = 0.2095 m/s, Error = 30.01%
Conclusion
Conclusion
 We have built an image-based disdrometer:
 Low cost and Easy-assembling
 Results are in the tendency of the empirical formula
 Keep good performance in windy situation
 Three kinds of experiments were done to verify the system
 The structure and processing procedures are feasible
 Thresholding calibration is needed
 Calculated rain rate is in the error around 15%
Future Work
 Still need further calibration in every set of images to increase measurement
accuracy
 Increasing FOV or frame rate to increase capture probability
 Improve contrast in field experiment
 Overlapping issue -> Set 2 CCD camera to make 3D images

Weitere ähnliche Inhalte

Was ist angesagt?

Was ist angesagt? (20)

meteorological instruments
meteorological instrumentsmeteorological instruments
meteorological instruments
 
Weather instruments
Weather instrumentsWeather instruments
Weather instruments
 
Meteorology
MeteorologyMeteorology
Meteorology
 
El nino and La nina
El nino and La ninaEl nino and La nina
El nino and La nina
 
Weather and Weather Elements
Weather and Weather ElementsWeather and Weather Elements
Weather and Weather Elements
 
El Nino
El NinoEl Nino
El Nino
 
El nino
El ninoEl nino
El nino
 
Sediment characteristics.ppt
Sediment characteristics.pptSediment characteristics.ppt
Sediment characteristics.ppt
 
Ppt on monsoon
Ppt on monsoonPpt on monsoon
Ppt on monsoon
 
Weather instruments ppt for students
Weather instruments ppt for studentsWeather instruments ppt for students
Weather instruments ppt for students
 
Measuring strike and dip
Measuring strike and dipMeasuring strike and dip
Measuring strike and dip
 
Air Masses
Air MassesAir Masses
Air Masses
 
Weather
WeatherWeather
Weather
 
Geological controls on ground water movement
Geological controls on    ground water movementGeological controls on    ground water movement
Geological controls on ground water movement
 
Presentation Meteorology
Presentation MeteorologyPresentation Meteorology
Presentation Meteorology
 
Meteorology
MeteorologyMeteorology
Meteorology
 
Thunderstorms, Tornadoes, and Hurricanes
Thunderstorms, Tornadoes, and HurricanesThunderstorms, Tornadoes, and Hurricanes
Thunderstorms, Tornadoes, and Hurricanes
 
Air sea interaction
Air sea interactionAir sea interaction
Air sea interaction
 
Earth’s Atmosphere
Earth’s AtmosphereEarth’s Atmosphere
Earth’s Atmosphere
 
WEATHER AND CLIMATE
WEATHER AND CLIMATEWEATHER AND CLIMATE
WEATHER AND CLIMATE
 

Ähnlich wie An image based disdrometer verification and raindrop analysis

LiDAR Technology and Geospatial Services
LiDAR Technology and Geospatial Services LiDAR Technology and Geospatial Services
LiDAR Technology and Geospatial Services MattBethel1
 
Coastal erosion management using image processing and Node Oriented Programming
Coastal erosion management using image processing and Node Oriented Programming Coastal erosion management using image processing and Node Oriented Programming
Coastal erosion management using image processing and Node Oriented Programming AbdAllah Aly
 
A Review on Haze Removal Techniques
A Review on Haze Removal TechniquesA Review on Haze Removal Techniques
A Review on Haze Removal TechniquesIRJET Journal
 
Inkjet quality measurement
Inkjet quality measurementInkjet quality measurement
Inkjet quality measurementYair Kipman
 
Rapid Laser Scanning the process
Rapid Laser Scanning the processRapid Laser Scanning the process
Rapid Laser Scanning the processSeeview Solutions
 
Digital Image Correlation Presentation
Digital Image Correlation PresentationDigital Image Correlation Presentation
Digital Image Correlation Presentationtrilionqualitysystems
 
TH1.L10.5: VALIDATION OF TIE-POINT CONCEPTS BY THE DEM ADJUSTMENT APPROACH OF...
TH1.L10.5: VALIDATION OF TIE-POINT CONCEPTS BY THE DEM ADJUSTMENT APPROACH OF...TH1.L10.5: VALIDATION OF TIE-POINT CONCEPTS BY THE DEM ADJUSTMENT APPROACH OF...
TH1.L10.5: VALIDATION OF TIE-POINT CONCEPTS BY THE DEM ADJUSTMENT APPROACH OF...grssieee
 
Shade Measurement Overview
Shade Measurement OverviewShade Measurement Overview
Shade Measurement Overviewphoberg
 
RIM Poster Optics r2.1 - 2-OP-05 Glatzel_Tinsley Poster
RIM Poster Optics r2.1 - 2-OP-05 Glatzel_Tinsley PosterRIM Poster Optics r2.1 - 2-OP-05 Glatzel_Tinsley Poster
RIM Poster Optics r2.1 - 2-OP-05 Glatzel_Tinsley PosterKevin Nouri
 
IRJET - Change Detection in Satellite Images using Convolutional Neural N...
IRJET -  	  Change Detection in Satellite Images using Convolutional Neural N...IRJET -  	  Change Detection in Satellite Images using Convolutional Neural N...
IRJET - Change Detection in Satellite Images using Convolutional Neural N...IRJET Journal
 
Shape Mediçao
Shape MediçaoShape Mediçao
Shape Mediçaonilson
 
Shape Mediçao
Shape MediçaoShape Mediçao
Shape Mediçaonilson
 
Deblurring Image and Removing Noise from Medical Images for Cancerous Disease...
Deblurring Image and Removing Noise from Medical Images for Cancerous Disease...Deblurring Image and Removing Noise from Medical Images for Cancerous Disease...
Deblurring Image and Removing Noise from Medical Images for Cancerous Disease...IRJET Journal
 
Unmanned Aerial Systems (UAS) Data Quality and Accuracy Realities
Unmanned Aerial Systems (UAS) Data Quality and Accuracy RealitiesUnmanned Aerial Systems (UAS) Data Quality and Accuracy Realities
Unmanned Aerial Systems (UAS) Data Quality and Accuracy RealitiesUAS Colorado
 
20051031 Biomass Smoke Emissions and Transport: Community-based Satellite and...
20051031 Biomass Smoke Emissions and Transport: Community-based Satellite and...20051031 Biomass Smoke Emissions and Transport: Community-based Satellite and...
20051031 Biomass Smoke Emissions and Transport: Community-based Satellite and...Rudolf Husar
 
Radiology lecture 2 CR and DR .pptx
Radiology lecture 2 CR and DR  .pptxRadiology lecture 2 CR and DR  .pptx
Radiology lecture 2 CR and DR .pptxMahrukhMunawar1
 

Ähnlich wie An image based disdrometer verification and raindrop analysis (20)

Photogrammetry 1.
Photogrammetry 1.Photogrammetry 1.
Photogrammetry 1.
 
LiDAR Technology and Geospatial Services
LiDAR Technology and Geospatial Services LiDAR Technology and Geospatial Services
LiDAR Technology and Geospatial Services
 
D04432528
D04432528D04432528
D04432528
 
Coastal erosion management using image processing and Node Oriented Programming
Coastal erosion management using image processing and Node Oriented Programming Coastal erosion management using image processing and Node Oriented Programming
Coastal erosion management using image processing and Node Oriented Programming
 
A Review on Haze Removal Techniques
A Review on Haze Removal TechniquesA Review on Haze Removal Techniques
A Review on Haze Removal Techniques
 
Inkjet quality measurement
Inkjet quality measurementInkjet quality measurement
Inkjet quality measurement
 
Rapid Laser Scanning the process
Rapid Laser Scanning the processRapid Laser Scanning the process
Rapid Laser Scanning the process
 
Raskar Ilp Oct08 Web
Raskar Ilp Oct08 WebRaskar Ilp Oct08 Web
Raskar Ilp Oct08 Web
 
Digital Image Correlation Presentation
Digital Image Correlation PresentationDigital Image Correlation Presentation
Digital Image Correlation Presentation
 
TH1.L10.5: VALIDATION OF TIE-POINT CONCEPTS BY THE DEM ADJUSTMENT APPROACH OF...
TH1.L10.5: VALIDATION OF TIE-POINT CONCEPTS BY THE DEM ADJUSTMENT APPROACH OF...TH1.L10.5: VALIDATION OF TIE-POINT CONCEPTS BY THE DEM ADJUSTMENT APPROACH OF...
TH1.L10.5: VALIDATION OF TIE-POINT CONCEPTS BY THE DEM ADJUSTMENT APPROACH OF...
 
Raskar Paris Nov08
Raskar Paris Nov08Raskar Paris Nov08
Raskar Paris Nov08
 
Shade Measurement Overview
Shade Measurement OverviewShade Measurement Overview
Shade Measurement Overview
 
RIM Poster Optics r2.1 - 2-OP-05 Glatzel_Tinsley Poster
RIM Poster Optics r2.1 - 2-OP-05 Glatzel_Tinsley PosterRIM Poster Optics r2.1 - 2-OP-05 Glatzel_Tinsley Poster
RIM Poster Optics r2.1 - 2-OP-05 Glatzel_Tinsley Poster
 
IRJET - Change Detection in Satellite Images using Convolutional Neural N...
IRJET -  	  Change Detection in Satellite Images using Convolutional Neural N...IRJET -  	  Change Detection in Satellite Images using Convolutional Neural N...
IRJET - Change Detection in Satellite Images using Convolutional Neural N...
 
Shape Mediçao
Shape MediçaoShape Mediçao
Shape Mediçao
 
Shape Mediçao
Shape MediçaoShape Mediçao
Shape Mediçao
 
Deblurring Image and Removing Noise from Medical Images for Cancerous Disease...
Deblurring Image and Removing Noise from Medical Images for Cancerous Disease...Deblurring Image and Removing Noise from Medical Images for Cancerous Disease...
Deblurring Image and Removing Noise from Medical Images for Cancerous Disease...
 
Unmanned Aerial Systems (UAS) Data Quality and Accuracy Realities
Unmanned Aerial Systems (UAS) Data Quality and Accuracy RealitiesUnmanned Aerial Systems (UAS) Data Quality and Accuracy Realities
Unmanned Aerial Systems (UAS) Data Quality and Accuracy Realities
 
20051031 Biomass Smoke Emissions and Transport: Community-based Satellite and...
20051031 Biomass Smoke Emissions and Transport: Community-based Satellite and...20051031 Biomass Smoke Emissions and Transport: Community-based Satellite and...
20051031 Biomass Smoke Emissions and Transport: Community-based Satellite and...
 
Radiology lecture 2 CR and DR .pptx
Radiology lecture 2 CR and DR  .pptxRadiology lecture 2 CR and DR  .pptx
Radiology lecture 2 CR and DR .pptx
 

Kürzlich hochgeladen

Unit 1 - Soil Classification and Compaction.pdf
Unit 1 - Soil Classification and Compaction.pdfUnit 1 - Soil Classification and Compaction.pdf
Unit 1 - Soil Classification and Compaction.pdfRagavanV2
 
KubeKraft presentation @CloudNativeHooghly
KubeKraft presentation @CloudNativeHooghlyKubeKraft presentation @CloudNativeHooghly
KubeKraft presentation @CloudNativeHooghlysanyuktamishra911
 
Thermal Engineering Unit - I & II . ppt
Thermal Engineering  Unit - I & II . pptThermal Engineering  Unit - I & II . ppt
Thermal Engineering Unit - I & II . pptDineshKumar4165
 
Generative AI or GenAI technology based PPT
Generative AI or GenAI technology based PPTGenerative AI or GenAI technology based PPT
Generative AI or GenAI technology based PPTbhaskargani46
 
Design For Accessibility: Getting it right from the start
Design For Accessibility: Getting it right from the startDesign For Accessibility: Getting it right from the start
Design For Accessibility: Getting it right from the startQuintin Balsdon
 
Intro To Electric Vehicles PDF Notes.pdf
Intro To Electric Vehicles PDF Notes.pdfIntro To Electric Vehicles PDF Notes.pdf
Intro To Electric Vehicles PDF Notes.pdfrs7054576148
 
Double rodded leveling 1 pdf activity 01
Double rodded leveling 1 pdf activity 01Double rodded leveling 1 pdf activity 01
Double rodded leveling 1 pdf activity 01KreezheaRecto
 
notes on Evolution Of Analytic Scalability.ppt
notes on Evolution Of Analytic Scalability.pptnotes on Evolution Of Analytic Scalability.ppt
notes on Evolution Of Analytic Scalability.pptMsecMca
 
Call Girls Wakad Call Me 7737669865 Budget Friendly No Advance Booking
Call Girls Wakad Call Me 7737669865 Budget Friendly No Advance BookingCall Girls Wakad Call Me 7737669865 Budget Friendly No Advance Booking
Call Girls Wakad Call Me 7737669865 Budget Friendly No Advance Bookingroncy bisnoi
 
Call Girls Pimpri Chinchwad Call Me 7737669865 Budget Friendly No Advance Boo...
Call Girls Pimpri Chinchwad Call Me 7737669865 Budget Friendly No Advance Boo...Call Girls Pimpri Chinchwad Call Me 7737669865 Budget Friendly No Advance Boo...
Call Girls Pimpri Chinchwad Call Me 7737669865 Budget Friendly No Advance Boo...roncy bisnoi
 
Call Girls In Bangalore ☎ 7737669865 🥵 Book Your One night Stand
Call Girls In Bangalore ☎ 7737669865 🥵 Book Your One night StandCall Girls In Bangalore ☎ 7737669865 🥵 Book Your One night Stand
Call Girls In Bangalore ☎ 7737669865 🥵 Book Your One night Standamitlee9823
 
ONLINE FOOD ORDER SYSTEM PROJECT REPORT.pdf
ONLINE FOOD ORDER SYSTEM PROJECT REPORT.pdfONLINE FOOD ORDER SYSTEM PROJECT REPORT.pdf
ONLINE FOOD ORDER SYSTEM PROJECT REPORT.pdfKamal Acharya
 
Unit 2- Effective stress & Permeability.pdf
Unit 2- Effective stress & Permeability.pdfUnit 2- Effective stress & Permeability.pdf
Unit 2- Effective stress & Permeability.pdfRagavanV2
 
VIP Call Girls Ankleshwar 7001035870 Whatsapp Number, 24/07 Booking
VIP Call Girls Ankleshwar 7001035870 Whatsapp Number, 24/07 BookingVIP Call Girls Ankleshwar 7001035870 Whatsapp Number, 24/07 Booking
VIP Call Girls Ankleshwar 7001035870 Whatsapp Number, 24/07 Bookingdharasingh5698
 
UNIT - IV - Air Compressors and its Performance
UNIT - IV - Air Compressors and its PerformanceUNIT - IV - Air Compressors and its Performance
UNIT - IV - Air Compressors and its Performancesivaprakash250
 
Top Rated Call Girls In chittoor 📱 {7001035870} VIP Escorts chittoor
Top Rated Call Girls In chittoor 📱 {7001035870} VIP Escorts chittoorTop Rated Call Girls In chittoor 📱 {7001035870} VIP Escorts chittoor
Top Rated Call Girls In chittoor 📱 {7001035870} VIP Escorts chittoordharasingh5698
 

Kürzlich hochgeladen (20)

(INDIRA) Call Girl Bhosari Call Now 8617697112 Bhosari Escorts 24x7
(INDIRA) Call Girl Bhosari Call Now 8617697112 Bhosari Escorts 24x7(INDIRA) Call Girl Bhosari Call Now 8617697112 Bhosari Escorts 24x7
(INDIRA) Call Girl Bhosari Call Now 8617697112 Bhosari Escorts 24x7
 
Unit 1 - Soil Classification and Compaction.pdf
Unit 1 - Soil Classification and Compaction.pdfUnit 1 - Soil Classification and Compaction.pdf
Unit 1 - Soil Classification and Compaction.pdf
 
KubeKraft presentation @CloudNativeHooghly
KubeKraft presentation @CloudNativeHooghlyKubeKraft presentation @CloudNativeHooghly
KubeKraft presentation @CloudNativeHooghly
 
Thermal Engineering Unit - I & II . ppt
Thermal Engineering  Unit - I & II . pptThermal Engineering  Unit - I & II . ppt
Thermal Engineering Unit - I & II . ppt
 
Generative AI or GenAI technology based PPT
Generative AI or GenAI technology based PPTGenerative AI or GenAI technology based PPT
Generative AI or GenAI technology based PPT
 
Call Girls in Ramesh Nagar Delhi 💯 Call Us 🔝9953056974 🔝 Escort Service
Call Girls in Ramesh Nagar Delhi 💯 Call Us 🔝9953056974 🔝 Escort ServiceCall Girls in Ramesh Nagar Delhi 💯 Call Us 🔝9953056974 🔝 Escort Service
Call Girls in Ramesh Nagar Delhi 💯 Call Us 🔝9953056974 🔝 Escort Service
 
Design For Accessibility: Getting it right from the start
Design For Accessibility: Getting it right from the startDesign For Accessibility: Getting it right from the start
Design For Accessibility: Getting it right from the start
 
Intro To Electric Vehicles PDF Notes.pdf
Intro To Electric Vehicles PDF Notes.pdfIntro To Electric Vehicles PDF Notes.pdf
Intro To Electric Vehicles PDF Notes.pdf
 
FEA Based Level 3 Assessment of Deformed Tanks with Fluid Induced Loads
FEA Based Level 3 Assessment of Deformed Tanks with Fluid Induced LoadsFEA Based Level 3 Assessment of Deformed Tanks with Fluid Induced Loads
FEA Based Level 3 Assessment of Deformed Tanks with Fluid Induced Loads
 
Double rodded leveling 1 pdf activity 01
Double rodded leveling 1 pdf activity 01Double rodded leveling 1 pdf activity 01
Double rodded leveling 1 pdf activity 01
 
notes on Evolution Of Analytic Scalability.ppt
notes on Evolution Of Analytic Scalability.pptnotes on Evolution Of Analytic Scalability.ppt
notes on Evolution Of Analytic Scalability.ppt
 
Call Girls Wakad Call Me 7737669865 Budget Friendly No Advance Booking
Call Girls Wakad Call Me 7737669865 Budget Friendly No Advance BookingCall Girls Wakad Call Me 7737669865 Budget Friendly No Advance Booking
Call Girls Wakad Call Me 7737669865 Budget Friendly No Advance Booking
 
NFPA 5000 2024 standard .
NFPA 5000 2024 standard                                  .NFPA 5000 2024 standard                                  .
NFPA 5000 2024 standard .
 
Call Girls Pimpri Chinchwad Call Me 7737669865 Budget Friendly No Advance Boo...
Call Girls Pimpri Chinchwad Call Me 7737669865 Budget Friendly No Advance Boo...Call Girls Pimpri Chinchwad Call Me 7737669865 Budget Friendly No Advance Boo...
Call Girls Pimpri Chinchwad Call Me 7737669865 Budget Friendly No Advance Boo...
 
Call Girls In Bangalore ☎ 7737669865 🥵 Book Your One night Stand
Call Girls In Bangalore ☎ 7737669865 🥵 Book Your One night StandCall Girls In Bangalore ☎ 7737669865 🥵 Book Your One night Stand
Call Girls In Bangalore ☎ 7737669865 🥵 Book Your One night Stand
 
ONLINE FOOD ORDER SYSTEM PROJECT REPORT.pdf
ONLINE FOOD ORDER SYSTEM PROJECT REPORT.pdfONLINE FOOD ORDER SYSTEM PROJECT REPORT.pdf
ONLINE FOOD ORDER SYSTEM PROJECT REPORT.pdf
 
Unit 2- Effective stress & Permeability.pdf
Unit 2- Effective stress & Permeability.pdfUnit 2- Effective stress & Permeability.pdf
Unit 2- Effective stress & Permeability.pdf
 
VIP Call Girls Ankleshwar 7001035870 Whatsapp Number, 24/07 Booking
VIP Call Girls Ankleshwar 7001035870 Whatsapp Number, 24/07 BookingVIP Call Girls Ankleshwar 7001035870 Whatsapp Number, 24/07 Booking
VIP Call Girls Ankleshwar 7001035870 Whatsapp Number, 24/07 Booking
 
UNIT - IV - Air Compressors and its Performance
UNIT - IV - Air Compressors and its PerformanceUNIT - IV - Air Compressors and its Performance
UNIT - IV - Air Compressors and its Performance
 
Top Rated Call Girls In chittoor 📱 {7001035870} VIP Escorts chittoor
Top Rated Call Girls In chittoor 📱 {7001035870} VIP Escorts chittoorTop Rated Call Girls In chittoor 📱 {7001035870} VIP Escorts chittoor
Top Rated Call Girls In chittoor 📱 {7001035870} VIP Escorts chittoor
 

An image based disdrometer verification and raindrop analysis

  • 1. An Image-based Disdrometer Verification and Raindrop Analysis 影像式雨滴譜儀系統驗證與雨滴分析 指導教授: 鐘太郎 教授 學生: 102061539 黃政翰
  • 2. Outline  Introduction  Raindrop Analysis Theories  Structure of the Proposed Disdrometer and Software Algorithm  Result of Experiments  Conclusion
  • 4. Precipitation Observation  Weather forecasting must tell the information of precipitation  Behavior of rainfall phenomena are due to local and sudden precipitation  Predicting rainfall intensity  Advance preparation can prevent potential disasters from happening
  • 5. Observing Systems  Primitive method  Collect drops by a box with dye paper  Wasting time and low efficiency  Radar and Satellite sensors  Collect data in large region  Not accurate enough in analyzing rather small region  Disdrometers  Analyze raindrop particles
  • 6. Why Disdrometer?  Increase the accuracy of the raining condition in a small region  Increase the accuracy of the radar measurement  Raindrop feature analysis can be used in  Air traffic control  Scientific examination  Weather observation system
  • 7. Types of Disdrometers  Joss-Waldvogel Disdrometer  Acoustic Disdrometer  Optical Particle Size Velocity Disdrometer  2-D Video Disdrometer  Image-based Disdrometer
  • 8. Joss-Waldvogel Disdrometer  Discriminate drop size by receiving the impact kinetic energy  It cannot determine the drop shape  Low sensitivity in drizzling and heavy rain
  • 9. Acoustic Disdrometer  Drops hitting on the sensor induces sonic wave  The piezoelectric sensor can measure the rainfall intensity  It does not provide raindrop distribution data  Wind influences the measurement easily
  • 10. Optical Particle Size Velocity Disdrometer  A row of laser beam points to a sensor  Measure drop size by calculating the duration of light extinction  2 dimension is possible; speed measurement is available  Drop mismatch makes errors
  • 11. 2-D Video Disdrometer  Using 2 line-scan cameras to measure size and shape  Velocity can be calculated by traveling time and distance between two frames  Slanted particle falling path makes image distortion and lead to errors
  • 12. Image-based Disdrometer  Use CCD camera to capture raindrops  Double exposures in each frame  Too many drops in one image influences matching accuracy a lot
  • 13. Disdrometer Comparison Disdrometer Type Measuring Mechanics Advantages Disadvantages Joss-Waldvogel Falling impact Good performace in small size variation Poor at measuring drops that are too small or too large Acoustic Inducing sonic wave Good at monitoring large size drops Wind influence easily OTT Parsivel Make drops falling through a laser beam With good accuracy in measuring drop size Drop mismatch and near drops makes error 2D Image Use 2 line-scan camera to measure size and shape Size, shape and velocity measurement are available Slanted falling drops are distorted Image-based use CCD camera to capture images and find parameters by image processing Low cost of recording and flexible system setting Camera frame rate and resolution influence the result
  • 15. Drop Size Distribution (DSD)  Marshall and Palmer (1948) [11]  Announced that DSD can be described by an exponential distribution D D eNN   0 4 0 08.0   cmN 121.0 41   cmR
  • 16. Drop Size Distribution (DSD)  The relationship comes up with errors in small drops  Ulbrich (1983) [12]-> Gamma function  are the parameters D D eDNN    0 ,, 0N
  • 17. Drop Size Distribution (DSD)  The relationship comes up with errors in small drops  Feingold and Levin (1986) [13]-> Lognormal function  are the parametersTg ND ,,
  • 18. DSD is related to  DSD has some relations with rain rate and rain types  Kozu and Nakamura (1991) [14]  DSD is related to reflectivity factor measured by radar  Doviak and Zrnic (1984) [15]  DSD can calibrate and increase the accuracy of the radar  Joss and Waldvogel (1969) [16]
  • 19. Drop Velocity  Gunn and Kinzer (1949) [17]  Experiments of raindrop terminal velocities through stagnant air  Battan (1964) [18]  Experiments in thunderstorm  Foote and DuToit (1969) [19], Beard (1976) [20]  Experiments in different air density
  • 20. Drop Velocity Different velocity curve under different conditions Different velocity curve under different air density Atlas et al. [5] Mitchell [21] Beard [20]
  • 21. Structure of the Proposed Disdrometer and Software Algorithm
  • 22. System Structure  Optical Unit  Light source (Part A)  Image Acquisition Unit  Lens (Part B)  CCD camera (Part C)  Data Processing Unit  Processing Algorithm (Part D) System Structure Diagram
  • 24. Optical Unit  Viswell HBL-100  Uniform blue LED light source  Bring up light intensity  Enhance image contrast
  • 25. Optical Unit  Relative position between light source and camera  It depends on the lens used in the system  Using a telecentric lens  We can only put the light source facing to the camera
  • 26. Optical Unit Light source on the side Light source facing to the camera
  • 27. Optical Unit  Proper adjustment of light intensity Light Intensity 50 52.5 55 57.5 Contrast under 0 degree light 0.241 0.232 0.053 0
  • 28. Image Acquisition Unit: Camera  CCD Camera: Pylon Basler Aca640-90gm  Monochrome, adjustable gain and exposure time  High frame rate (90 fps in stable)  659 pixels * 494 pixels  SDK is provided
  • 29. Image Acquisition Unit: Lens  Lens: OPTO Engineering TC13064  Telecentric lens  Minimizes blur effect  Long depth of field
  • 30. Image Acquisition Unit: Lens  Compare with the old one: Computar M1214-MP2 TC13064 M1214-MP2 FOV 6.5cm*4.8cm 5cm*3.7cm DoF about 15cm about 3~5cm Focus Fixed Manual adjustable Iris Fixed Manual adjustable Out of focus Blurriness Slight Severe
  • 31. Image Acquisition Unit: Lens Image taken by M1214-MP2 Image taken by TC13064
  • 32. Data Processing Unit  Software platform: Visual Studio 2012, using Visual C++  Combined with: OpenCV 2.4.9, Pylon 4 SDK, Matlab2010
  • 33. Data Processing Unit: Camera Parameter Setting  Done before taking every set of images  Critical parameters  Image size  Exposure time  Gain  Recording duration
  • 34. Data Processing Unit: Camera Parameter Setting Set the duration time of recording or the number of images taken Set the exposure time Set the gain Set the image size Start taking images Start the program
  • 35. Data Processing Unit: Drop Extraction and Analysis
  • 36. Drop Extraction and Analysis: Make a Background  Background calculation: average all the frames  is the ith frame, is the number of taken images k if Bg k i   ][ ][if k k average
  • 37. Drop Extraction and Analysis: Make a Background Original image Image without background
  • 38. Drop Extraction and Analysis: Image Binarization  Median Filter 5*5 -> reduce noise  Choose proper threshold method  Depend on the image we get  Max Entropy, Iterative, Otsu, Region Growing, Level Set had been tried  100us exposure time image: Max Entropy Thresholding  2000us exposure time image: Iterative Thresholding
  • 39. Drop Extraction and Analysis: Max Entropy Thresholding  The threshold determines by maximizing the entropy of foreground and background  is the gray-level probability density function for the image ) )( )( log( )( )(255 1 Tq ip Tq ip H fTi f f   ) )( )( log( )( )( 0 Tq ip Tq ip H b T i b b   )(ip
  • 40. Drop Extraction and Analysis: Max Entropy Thresholding  and are the probabilities that a given pixel belongs to foreground or background when the threshold is   255 1 )()( Ti f ipTq   T i b ipTq 0 )()( )(Tqf )(Tqb T )max( bf HHT 
  • 41. Drop Extraction and Analysis: Iterative Thresholding  An initial threshold is chosen, typically the average intensity of the image  Mean gray value of foreground and background are calculated  is the gray-level probability density function for the image T   255 1 )( Ti f iip   T i b iip 0 )( 2 bf T    )(ip
  • 42. Drop Extraction and Analysis: Binary images after thresholding 100us binary image 2000us binary image
  • 43. Data Processing Unit: Drop Extraction and Analysis
  • 44. Drop Extraction and Analysis: Contour Finding  Binary images have high contrast  Easy to determine edges
  • 45. Drop Extraction and Analysis: Ellipse Fitting
  • 46. Drop Extraction and Analysis: Minimal Bounding Rectangle
  • 47. Drop Extraction and Analysis: Unsuitable Objects Elimination  Out of bound elimination  Any contour touches the border are eliminated  100us images  Axis ratio 0.4~1.2 -> treated as raindrops  2000us images  Eliminate if the width is larger than height
  • 48. Drop Extraction and Analysis: Size Calculation 1.01.0  PA 2 1 )(2  A Dm  mm1.0 mm1.0 P
  • 49. Drop Extraction and Analysis: Velocity Calculation 2 2 w hd  tdv / h w 2 w 2 w
  • 50. Drop Extraction and Analysis: DSD calculation  According to Liu et al. (2013)  : DSD  : number of drops of each bin  : bin interval (1mm)  : Sampling volume of the drop-falling space )(DN )( 13  mmm )(DNum dD )(15.0 3 mWHV  m m m dDV DNum DN   Pr )( )( framemymx osuremyosuremx tvv tvhHtvwW    )()( Pr expexp
  • 51. Drop Extraction and Analysis: Rain Rate Calculation  : velocity of the measured diameter    0 3 )( 6 mmmm dDvDDNR  mv
  • 53. Marble Experiments  Throwing marbles from 1mm to 5mm separately  100us exposure time, 300 gain, maximum light  3 seconds duration, 270 frames, as one set of images
  • 54. Marble Experiments Measured diameter vs Calculated Diameter Measured diameter vs Measured Area
  • 55. Marble Experiments Theoretical Value Average Std Min Max Error diameter(100us) (mm) 0.25 0.3110 0.0902 0.0707 0.4998 24.3815 area(100us) (mm 2 ) 0.0491 0.1317 0.0511 0.055 0.215 168.1602 Theoretical Value Average Std Min Max Error diameter(100us) (mm) 1 1.0219 0.2217 0.5 1.4948 2.1872 area(100us) (mm 2 ) 0.785 0.7369 0.4036 0.09 1.9 6.1687 Theoretical Value Average Std Min Max Error diameter(100us) (mm) 2 2.0560 0.2808 1.5 2.4749 2.8018 area(100us) (mm 2 ) 3.142 3.2074 1.195 0.535 5.305 2.0950 Theoretical Value Average Std Min Max Error diameter(100us) (mm) 3 3.0887 0.2514 2.5 3.4883 2.9583 area(100us) (mm 2 ) 7.069 7.4855 1.5203 1.62 10.735 5.8988 Theoretical Value Average Std Min Max Error diameter(100us) (mm) 4 4.0374 0.2346 3.5 4.4721 0.9339 area(100us) (mm 2 ) 12.566 12.4727 2.2965 4.225 17.085 0.7456 Theoretical Value Average Std Min Max Error diameter(100us) (mm) 5 4.9206 0.3012 4.5 5.4447 1.5886 area(100us) (mm 2 ) 19.635 18.1703 3.5536 5.985 24.27 7.4594
  • 56. Marble Experiments  Overlapping leads to the presence of outliers  Larger marbles have higher error  Axis ratio recognition gives larger range to be distinguished in large size objects  Small marbles error  Some noises are remained after thresholding
  • 57. Water Sprinkling Experiments  Spread water by sprinkler  100us exposure time, 300 gain, maximum light  2000us exposure time, 300 gain, half light  5 seconds duration, 450 frames, as one set of images
  • 58. Water Sprinkling Experiments Measured diameter vs Calculated Diameter Measured diameter vs Measured Area
  • 59. Water Sprinkling Experiments Axis ratio distribution
  • 60. Water Sprinkling Experiments Diameter vs Speed Canting angle histogram
  • 61. Water Sprinkling Experiments Theoretical Value Average Std Min Max Error diameter(2000us) 0.25 0.324 0.096 0.045 0.500 29.7393 speed(2000us) 0.7847 0.643 0.530 0.069 2.705 18.0382 diameter(100us) 0.25 0.456 0.045 0.300 0.499 82.2501 area(100us) 0.0491 0.136 0.045 0.005 0.235 177.8743 Theoretical Value Average Std Min Max Error diameter(2000us) (mm) 1 0.9795 0.2773 0.5 1.4997 2.0474 speed(2000us) (m/s) 3.9972 2.1910 0.8306 0.05 8.6185 45.1867 diameter(100us) (mm) 1 0.7777 0.1927 0.5 1.4977 22.2321 area(100us) (mm2 ) 0.7854 0.4596 0.2566 0.02 1.795 41.4878 Theoretical Value Average Std Min Max Error diameter(2000us) (mm) 2 1.8524 0.2785 1.5 2.4989 7.3819 speed(2000us) (m/s) 6.5477 3.4311 1.1124 0.1 10.454 47.5979 diameter(100us) (mm) 2 1.7033 0.2171 1.5 2.4660 14.8353 area(100us) (mm2 ) 3.1416 1.6046 0.5596 0.475 3.685 48.9244 Theoretical Value Average Std Min Max Error diameter(2000us) (mm) 3 2.8973 0.2750 2.5 3.4985 3.4220 speed(2000us) (m/s) 7.9474 4.3873 1.4764 0.75 10.8093 44.7961 diameter(100us) (mm) 3 2.8418 0.0189 2.82843 2.8552 5.2725 area(100us) (mm 2 ) 7.0686 2.97 0.9334 2.31 3.63 57.9831 Theoretical Value Average Std Min Max Error diameter(2000us) (mm) 4 3.8909 0.2694 3.5 4.4933 2.7287 speed(2000us) (m/s) 8.7156 5.1360 1.9515 0.75 12.8701 41.0717 diameter(100us) (mm) 4 N/A N/A N/A N/A N/A area(100us) (mm2 ) 12.5664 N/A N/A N/A N/A N/A Theoretical Value Average Std Min Max Error diameter(2000us) (mm) 5 4.9799 0.2777 4.5 5.4811 0.4011 speed(2000us) (m/s) 9.1372 5.9424 2.1716 1.4 11.4378 34.9649 diameter(100us) (mm) 5 N/A N/A N/A N/A N/A area(100us) (mm2 ) 19.6350 N/A N/A N/A N/A N/A Average Std Min Max canting angle -35.6758 26.6744 -180 0
  • 62. Water Sprinkling Experiments  Larger error of speed difference in larger drop size  Overlapping issue  Sprinkled water are not in terminate velocity  Few drops are grabbed in this size interval
  • 63. Raining Experiments  Real raining condition at 17:00, 27 Aug 2015 at Hsinchu, Taiwan  5 seconds duration, 450 frames, as one set of images, 30 seconds in total  2700 images taken in 100us and 2000us respectively
  • 64. Raining Experiments Histogram of size distribution Axis ratio distribution
  • 65. Raining Experiments Measured diameter vs Calculated Diameter Measured diameter vs Measured Area
  • 66. Raining Experiments Histogram of size distribution Histogram of canting angle distribution
  • 68. Raining Experiments Theoretical Value Average Std Min Max Error diameter(2000us) (mm) 0.25 0.374864 0.108827 0.14 0.497947 49.9457 speed(2000us) (m/s) 0.7847 1.15072 0.598629 0.28 3.2 46.6446 diameter(100us) (mm) 0.25 0.441055 0.065593 0.31305 0.494975 76.4218 area(100us) (mm2 ) 0.0491 0.137188 0.060991 0.025 0.215 179.4043 Theoretical Value Average Std Min Max Error diameter(2000us) (mm) 1 0.9394 0.2447 0.5 1.4999 6.0578 speed(2000us) (m/s) 3.9972 2.6945 0.8629 1.05 5.3062 32.5902 diameter(100us) (mm) 1 0.8658 0.2196 0.5 1.4863 13.4154 area(100us) (mm2 ) 0.7854 0.6174 0.3419 0.055 1.87 21.3847 Theoretical Value Average Std Min Max Error diameter(2000us) (mm) 2 1.7479 0.1486 1.5402 2.0803 12.6055 speed(2000us) (m/s) 6.5477 5.3268 0.5195 4.0784 6.0691 18.6465 diameter(100us) (mm) 2 1.9637 0.1418 1.5 2.2030 1.8129 area(100us) (mm 2 ) 3.1416 3.0464 0.3600 1.695 3.695 3.0294 Average Std Min Max canting angle -10.276 29.3070 -180 -0.273
  • 69. Raining Experiments  Small raindrops dominant  Small canting angle -> almost no wind  Speed are lower than theoretical value  Image processing leads to the error  Raining condition difference
  • 70. Raining Experiments: Image Processing Error  If there is one-pixel error in width  An 1mm raindrop is in 10% error  Theoretical velocity is in 9% error
  • 72. Raining Experiments  According to the statistical data of Central Weather Bureau  Rain rate = 0.5 mm/h  Wind speed = 0.3 m/s  The calculated data  Rain rate = 0.5721 mm/h, Error = 14.42%  Wind speed = 0.2095 m/s, Error = 30.01%
  • 74. Conclusion  We have built an image-based disdrometer:  Low cost and Easy-assembling  Results are in the tendency of the empirical formula  Keep good performance in windy situation  Three kinds of experiments were done to verify the system  The structure and processing procedures are feasible  Thresholding calibration is needed  Calculated rain rate is in the error around 15%
  • 75. Future Work  Still need further calibration in every set of images to increase measurement accuracy  Increasing FOV or frame rate to increase capture probability  Improve contrast in field experiment  Overlapping issue -> Set 2 CCD camera to make 3D images

Hinweis der Redaktion

  1. 為何要觀測降雨? 氣象預報降雨資訊重要 許多降雨引起的現象是很LOCAL的 預測降雨可以事先準備預防災害發生
  2. 增加了幾個參數來更好的描述DSD
  3. 不同降雨型態反映出的DSD也會不同 參數的表現也會不同 降雨型態分為層狀對流 回波值也是如此 回波值>40dBZ 幾乎為對流性
  4. Stagnant air 滯流