2009 spie dirfilt

Directional Analysis and Filtering for Dust Storm detection

in NOAA-AVHRR Imagery

S. Janugani a, V. Jayaram a, S. D. Cabrera a, J. G. Rosiles a, T. E. Gill b,c, N. Rivera Rivera b

aDept. of Electrical & Computer EngineeringbEnvironmental Science & Engineering Program

cDept. of Geological Sciences

The University of Texas at El Paso

500 W. University Ave., El Paso, TX 79968-0523

ABSTRACT

In this paper, we propose spatio-spectral processing techniques for the detection of dust storms and automaticallyfinding its transport direction in 5-band NOAA-AVHRR imagery. Previous methods that use simple band mathanalysis have produced promising results but have drawbacks in producing consistent results when low signalto noise ratio (SNR) images are used. Moreover, in seeking to automate the dust storm detection, the presenceof clouds in the vicinity of the dust storm creates a challenge in being able to distinguish these two types ofimage texture. This paper not only addresses the detection of the dust storm in the imagery, it also attemptsto find the transport direction and the location of the sources of the dust storm. We propose a spatio-spectralprocessing approach with two components: visualization and automation. Both approaches are based on digitalimage processing techniques including directional analysis and filtering. The visualization technique is intendedto enhance the image in order to locate the dust sources. The automation technique is proposed to detect thetransport direction of the dust storm. These techniques can be used in a system to provide timely warnings ofdust storms or hazard assessments for transportation, aviation, environmental safety, and public health.

Keywords: Multispectral imagery, NOAA-AVHRR imagery, Dust storm, Band math analysis, Directionalfiltering.

1. INTRODUCTION

Dust Storms are serious causes of physical, environmental and economic hazards. Due to this growing concern,detection of dust storms using multispectral imagery has become a topic of several recent studies.1–6 Theneed to develop a real-time capability to detect, identify, quantify and track various dust storms is a wellknown problem to scientists and climate monitoring agencies. Detection of particulate effluents, the studyof volcanology and environmental risk management are just a few of the many applications that demand anautomated/in-situ processing system. The multispectral images from AVHRR, GOES, MODIS and other sensorshave opened up new possibilities for detecting and monitoring atmospheric dust events and tracking the potentialenvironmental risks. Although several traditional methods for detecting dust storms exist, they have difficultiesin consistently detecting and distinguishing the dust storm features. Traditional methods include: band mathanalysis,2 differencing of radiation temperature or the infrared split-window technique, single band thresholdingand multi-band combinations. They have achieved good results in detecting dust storms,5however, they havelimitations in detecting transport direction and in identifying the location of the dust storm sources.

In this paper, we propose a spatio-spectral processing scheme to detect the presence of dust storms visuallyand with an automated technique for finding dust storm direction. The goal is to present seamless integration ofadvanced data analysis and modeling tools to scientists and climate monitoring personnel, advancing the state-of-the-practice in the utilization of satellite image data to various types of dust storm studies. The proposedmethodology is a twofold approach. First, we present a scheme to visualize the directionality and point sources

Further author information: (Send correspondence to S. D. Cabrera)Prof. S. D. Cabrera: E-mail: [email protected], Telephone: 1 915 747 5470

Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XV,edited by Sylvia S. Shen, Paul E. Lewis, Proc. of SPIE Vol. 7334, 73341G · © 2009 SPIE

CCC code: 0277-786X/09/$18 · doi: 10.1117/12.820235

Proc. of SPIE Vol. 7334 73341G-1

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 09/04/2013 Terms of Use: http://spiedl.org/terms

of these dust storm events. The second approach provides an automated technique to compute the transportdirection of the dust storm. In the visualization technique, image processing algorithms like the spectral-domainprincipal component analysis (PCA), minimum noise fraction (MNF) transforms7 and the spatial-domain k-means unsupervised classification method are used as tools to help locate dust storms visually. Pointing outthe dust sources is done based on image information, directional filtering in combination with edge detectorsand spectral-domain classification techniques. Next, edge detectors like Sobel and Frei-Chen are applied tothe selected filtered images for further enhancement of the streaks produced by the directional texture. Falsecolor composite images are created to visually enhance the directional streaks to be able to easily locate thedust sources. The second approach comprising the automation technique for finding direction of the dust storminvolves performing the power spectrum analysis on bands 4 and 5. This step is performed because thesewavelengths highlight the absorption and subsequent emission of thermal radiation by the silicate particles inthe dust storms. The scheme involves block processing consisting of power spectrum analysis followed by binarythresholding and morphological enhancement. This local spectral analysis is first used to confirm the presence ofhigh directionality information in certain regions of an image. Binary thresholding is performed on these blocksto enhance the directional texture. Further morphological enhancement is performed on these binary imagesto compute the dust storm’s area and orientation. The similarity in orientation among neighboring blocks alsogives confirming evidence of the presence of a dust storm in the area formed by putting together these blocks.

2. EARLIER WORK

The motivation for this paper is derived from the experiments documented in Rivera and Gill’s work.4, 5 Themethodology developed in that previous work, primarily uses band math analysis as the technique to detectdust storm areas. Dust sources are located by visual analysis and manual labeling. Other previous methods fordust storm detection include: differencing of radiation temperature, single band thresholding and multi-bandcombining. Some have achieved good results in detecting dust storms, however, these techniques have limitationsin detecting direction of transport of the dust storm, and the visibility index. Above all, there is an absence ofan automated detection scheme for in-situ processing systems.

In a related work, hierarchial principal component analysis (HPCA) techniques8 produced significant seg-mentation results. However, the technique is computationally challenging and expensive as they involve datafusion both spectrally and spatially from cameras placed at different angles. In most practical scenarios, it is notpossible to get the raw data sets of the same scene (same geographical area at a particular time) from camerasplaced at different angles.

Koren et al.9 proposed a dust source detection method using geo-referencing of the dust-storm image withthe non-dust-storm image of the same location. The limitation of this method is that the technique is dependenton the availability of the image at the same location, with identical geometry and resolution. To avoid any falsealarms, resulting from clouds or pollution, the image should also be captured at close to the same time.

3. SPATIO-SPECTRAL PROCESSING FOR DUST STORM DETECTION

In this paper, we focus on developing an automated approach for locating dust sources and finding dust stormdirection using NOAA-AVHRR imagery. Images collected with the AVHRR, as well as the MODIS sensor, areof particular interest in dust storm detection because of their relatively higher temporal resolutions (once perday) compared with LANDSAT-type systems (once every 16 days). The nighttime thermal imaging capabilityof AVHRR and the higher spectral resolution of MODIS can also assist in detecting dust emission and mappingthe vulnerability of the landscape to wind erosion.2 As mentioned earlier, our approach involves both spatialand spectral processing. A step-by-step view of our approach, involving visualization and automation, is shownin the form of a flowchart in Figure 1.

The band math analysis, also used in this paper, produces results as shown in Figure 2. This differencebetween bands 4 and 5 is used as an initial step to detect the presence of the dust storm area. This is the firststep in detecting the dust storm location as shown in the flowchart in Figure 1. The dark regions in the imageindicate the presence of dust storm. The presence of clouds in the vicinity of the dust storm would make theproblem more difficult since they have a very similar texture on the resulting imagery. In order to overcome



(FHt..lng

I

(F.. Coo, CompotoI

1

Enh.no.d dk.tion..t..b.. to, lonottog

d.t ,o.on., I

NOOA.AV000Fbo bond ,.w doto

(B,dn.th PCA/ MNF

Corn Doto,t,00)

IK.00nnn CImthcbon

)Vor,fic.bo., Stop)

Ic

Crop doot Oton,,regIon45l2x5I2)

VsuaIization Automa)ion

ttonlAutornaUonj

BIob Pr0000ing125,128 blonko I

(bond4B.nd5) I-II

IEdgo Ornroflon Powo, Spootront

)Sobol nd Fro-Chwr) onolyoO

IIBtoroy Threnholng

IMorphologbotEnbonoentOnt

Dtrwotiorrollty

L_CornPntottJ

Figure 1. Block diagram for proposed spatio-spectral processing scheme of dust storm detection.

the false alarm of misinterpreting the cloud pixels as part of the dust storm area, we can employ the classicalunsupervised pattern recognition technique such as the k-means classification.10 This step is performed to verifythe dust storm presence and to match the results with those of band math analysis.

The next step in the processing involves selecting a subimage (in this case a 512x512 pixel region) thatincludes the dust storm area. This is a cropping step from all bands of the original raw NOAA-AVHRR imagedata set. The k-means classification is performed on these bands to verify if it is able to segment the same duststorm region again. An illustration of this process can be seen in Figure 3. Here, the k-means classificationis able to segment the same dust storm region. If band math analysis and k-means classification are able tosegment a similar region, it indicates with high confidence the presence of the dust storm. Next, we proceedto locate the sources of the dust storms using the visualization technique and to find the direction of the duststorm with the automation technique.

4. LOCATING DUST SOURCES

In the visualization approach, directional filtering is applied on the cropped individual image bands at differentangles. Directional filters are significant in many image processing applications such as edge sharpening, featureenhancement, texture analysis and object recognition. In this paper we have employed steerable Gaussian filtersand 2-D convolution filter masks at various angles. Steerable filters are a class of filters in which a filter ofarbitrary orientation is synthesized as a linear combination of a set of basis filters.11 In using this approach, thesame filter is rotated to multiple angles to obtain a response for each angle.



Figure 2. Original band 4 (left), band 5 (center) and the bandmath (band 4 - band 5) result (right) for April 15, 2003event over southwestern North America.

Figure 3. k-means classification of cropped 512x512 region of Band 1 (April 15, 2003 event).



Consider the two-dimensional circularly symmetric Gaussian function G(x, y) = e−(x2+y2). Let the nthderivative of a Gaussian in the x direction be Gn. Let fθ(x, y) represent f(x, y) rotated through an angle θ

about the origin. The first x derivative of a Gaussian is

G0◦

1 =∂

∂x[e−(x2+y2)] = −2xe−(x2+y2),

and if rotated by 90◦

G90◦

1 =∂

∂y[e−(x2+y2)] = −2ye−(x2+y2).

Thus, the filter G1 for any arbitrary orientation θ can be synthesized by a linear combination

Gθ1 = cos(θ)G0◦

1 + sin(θ)G90◦

1 .

Here G0◦

1 and G90◦

1 are called the basis filters for Gθ1 and cos(θ), sin(θ) components are the interpolation

functions for the basis filters. We can synthesize the image filtered at an arbitrary orientation using convolutionby taking linear combinations of images filtered with G0◦

1 and G90◦

1 .

Therefore,

R0◦

1 = G0◦

1 ∗ I

R90◦

1 = G90◦

1 ∗ I

Rθ1 = cos(θ)R0◦

1 + sin(θ)R90◦

1 .

Figure 4 illustrates the results of directional filtering at different angles. The directional filtering was per-formed on the cropped version of band 4 for different angles: 20◦, 30◦, 40◦, 50◦, 70◦, and 90◦. Notice thatin Figure 4, the texture of the dust storm is more prominent around 70◦. To concretize our findings, we havealso performed directional filtering using direct convolution with another type of filters (those provided in theENVI software) wherein a certain resulting output pixel value is the function of some weighted average of thebrightness of the neighboring input pixels. The sum of the directional filter kernel/mask elements is zero forthis family. Convolution of this filter mask with the input image results in a spatially filtered output image.These first derivative edge enhancement filters selectively bring out image features having a specific prominentdirection. The outcome is a spatially high pass filtered image where the convolution with the directional filtermask contains areas with uniform pixel values represented by ’0’ and those with variable values that form abright edge enhanced texture. It is found that computing directionality or orientation angle of the dust stormwith both directional and convolution filters provides nearly identical results. It is observed that the 8 dustsources marked in “yellow”in Figure 5 match with the dust sources shown by Rivera.5

Energy and entropy measurements are computed on the outputs of the directional filters to find the prominentdirection of the dust storm. To enhance the streaks representing the directional components of the storm, weuse edge detection operators like Sobel and Frei-Chen. To improve the visualization of the resulting images weuse false coloring over composites from three out of the five bands. The display color assignment for any band isdone in an arbitrary manner. In general, many environmental features are more readily discernible when satelliteimages are represented as false color composites. In our case, pseudo-coloring enhances the localization of theorigin points of a dust storm. Figure 5 shows the located origin points of a dust storm on false color compositeimages of different band combinations are show in the Figure. It is observed that band combination 1, 4, 5 forcolors Red, Green, and Blue is possibly the best one to locate the dust sources easily by visual interpretation.



50 100 150 200 250

50

100

150

200

25050 100 150 200 250

50

100

150

200

25050 100 150 200 250

50

100

150

200

250

50 100 150 200 250

50

100

150

200

25050 100 150 200 250

50

100

150

200

25050 100 150 200 250

50

100

150

200

250

Figure 4. Directional filtering of cropped band 4 for different angles 20◦, 30◦, 40◦, 50◦, 70◦, 90◦ (from left to right).

Figure 5. (Left) False Color Composite image - Band 1, 4, 5 (R, G, B) with marked dust sources for April 15, 2003 event.(Center) False Color Composite image - Band 1, 4, 3 (R, G, B) with marked dust sources for April 15, 2003 event. TheFigure on the right depicts the False Color Composite image - Band 1, 3, 4 (R, G, B) with marked dust sources for April

15, 2003 event.(for color image see electronic version of manuscript)



5. TRANSPORT DIRECTION OF DUST STORM

In the automation process, we perform block processing on subimages that are cropped regions (512x512 pixels)of bands 4 and 5. Block processing12 means dividing an image into blocks and processing them individually in anindependent fashion to provide spatial localization. The smallest block size used is 128x128 based on keeping aminimum amount of frequency domain resolution in the Discrete Fourier Transform (DFT) representation of eachblock. A local power spectral density analysis in each of these blocks is used to find the prominent direction ofthe texture in that image block. This analysis can also be used to define a candidate dust storm region consistingof multiple blocks. The presence of a prominent direction in the texture of the candidate dust storm region canalso be used to verify its presence as an automated detection scheme. Binary thresholding based on Otsu’s13

method is performed on the power spectrum representations for all blocks in order to define the orientation ofthe spectral information. Otsu’s method chooses a threshold to minimize the intra-class variance of the twoclasses of pixels that are converted to 0 or 1 in forming the output binary image. Morphological enhancementsare then applied to these binary images to better define the bright regions corresponding to large spectral energy.The size and prominent orientation of each largest bright region for each block are then calculated. These twonumbers are the features that can be used to define dust storm blocks (large areas) with their correspondingorientations. The results of these experiments are presented in the following section.

6. EXPERIMENTS

The current on-orbit operational satellites maintained by NOAA are NOAA-18, 17, 16, 15, 14 and 12.14 Othersensors which are of particular interest in dust storm detection and monitoring are the Moderate ResolutionImaging Spectrometer (MODIS) and a sensor onboard the Geostationary Operating Environmental Satellite(GOES). MODIS has thirty-six spectral bands and GOES has five spectral bands. Imagery from GOES are themost suitable to track the time evolution of active and short lived dust storms because of their high temporalresolution.15 Also,there are three different MODIS sensors - Terra, Aqua and Aura - each on a different satellitecovering an area at different times of the day. So, as many as three MODIS images are available per day for thesame geographical region.

We experimented with three data sets for dust storms in the Chihuahuan desert region of southwest NorthAmerica. Most of our analysis was done using NOAA-AVHRR multispectral data of two dust events which haverecently been the subject of other investigations.4, 5 These events occurred on April 15th, 2003 (20:23 UTC),shown in Figure 6, and on December 15th 2003 (19:51 UTC), shown in Figure 9. A third, non-dust-storm imagedataset from April 1st, 2003 is shown in Figure 12 and was used for analyzing susceptibility to false dust stormdetections.

Next we show the results of the automation process which starts with power spectrum analysis on blocks ofthe cropped dust storm region. The power spectrum analysis of April 15, 2003 is shown in Figure 6. The blockswhich are marked in red indicate the presence of strong dust storm evidence. After binary thresholding usingOtsu’s method,13 the image closing operation is then performed (see Figure 7) in order to ease the extraction ofthe area and orientation information for the binary objects in the image, see the results in Figure 8. The resultsobtained from the automation technique for December 15, 2003 event are as shown in the next set of Figures,see Figures 9, 10, and 11.

To analyze false alarms, we study two data sets which have no dust storm evidence in them. The first one isthe April 1, 2003 image data. The second one is a cropped non-dust storm region taken from the original sizeDecember 15, 2003 raw image data. In this paper, only the April 1, 2003 image data is analyzed to investigatethe performance of the proposed technique on non-dust storm imagery. We seek to analyze false alarms thatcould occur in an automated method when processing imagery during the absence of dust storms. Band mathanalysis is done as a first step. The marked dark region in Figure 12 is not a dust storm but is the result of thisanalysis. Therefore, this suspected 512x512 region is cropped for false alarm analysis.

The suspected region from the April 1st, 2003 image is then subjected to k-means classification based seg-mentation. The resulting region does not agree well with the region found by band math analysis, see Figure13. Hence, we need not go further into the next steps of visualization and automation since we do not have theproper verification of the presence of a dust storm.



60

100

ISO

200

260

300

360

400

460

600

P0Ir!Jm 61°

66 100 160 200 260 300 360 462 4 600

4

4 4

4

Figure 6.

(Left) Cropped 512x512 band 4 of April 15, 2003 event. (Right) Power Spectrum Analysis of April 15, 2003event.

Figure 7. (Left) Binary thresholding of the power-spectrum of April 15, 2003 event. (Right) Morphological enhancementon the binary thresholded image of April 15, 2003 event.



9 32 15 1

00 -20.9101° 73.1550° 0°

1 224 413 9

005726201 -524462° 90°

437 221 580 1

21.87360 -78.4846° -5765 0°di:267 433 12 1

-79,3580° -87.5- 20.3006° 0°

Powe Spectrum 81°

66 lOU 160 W 260 306 392 400 460 600

Figure 8. Area and orientation computations of the binary objects in the morphologically enhanced image.

Figure 9. (Left) Band 4 of December 15, 2003 event. (Right) Power Spectrum Analysis of December 15, 2003 event.



25 29 33 7

90° -68,8632° 8405570 150

26 25 r 1' 344

-69.1832° 86.5444° 8.O329 2546S0

147 241 537 77

88,86230 85.9770° 87.3790° .17,52970-I,265 172 338 3

468450° -85.7150° 20.30O6 0°

Figure 10. (Left) Binary thresholding of the power-spectrum of December 15, 2003 event. (Right) Morphological enhance-ment on the binary thresholded image of December 15, 2003 event.

Figure 11. Area and orientation computations of the binary objects for December 15, 2003 event.



3

r.

4-A

çt/,

;.-

r:

Figure 12. Bandmath (Band subtraction) image of April 1, 2003 event.

Figure 13. (Left) Original 512x512 image of April 1, 2003. (Right) k-means classified 512x512 image of April 1, 2003.



7. CONCLUSION AND FUTURE WORK

The spatio-spectral processing technique proposed in this research is one possible approach to directional analysisof dust storm evidence in satellite imagery. Future work may targeted towards more quantitative and qualitativeresults demonstrating directionality information of the transport of dust storms. Further analysis on the enhanceddirectional streaks resulting from visualization technique could be done to see how accurately they correspond toactual dust plumes created at the sources of dust. In addition, it is of interest to see if their shape closely followsthe path of the dust transport over longer distances. Performance of the technique for monitoring volcanic ashcould also be considered. The behavior and properties of volcanic ash particles determine the significant bandsthat can be used for further analysis. Future work also includes investigation of more advanced multi-resolutiondirectional filtering approaches based on Wavelets, and Bamberger Pyramids.

Other possible future work includes the assessment of this processing technique on the GOES and MODISimages, which are more frequently available so that it could be used operationally to automatically detect duststorms. These processing techniques may also be evaluated on the images with no dust storm evidence and duringdifferent times of day to analyze false alarms. False alarms that might result from the textures of clouds, smoke,fire and air pollution may also be studied. Further investigation on supervised and unsupervised classificationtechniques is also suggested.

REFERENCES

[1] J.A. Lee, T.E. Gill, K. M. M. A. A. P., “Land use/land cover and point sources of the 15 December 2003dust storm in southwestern North America,” Geomorphology 105, 18–27 (2009).

[2] Ackerman, S. A., “Remote sensing aerosols using satellite infrared observations,” Journal of GeophysicalResearch 102, 17069–17079 (1997).

[3] H. El-Askary, M. Kafatos, X. L. T. E.-G., “Introducing new approaches for dust storms detection usingremote sensing technology,” Proc. of IEEE International Geoscience and Remote Sensing Symposium .

[4] N. I. Rivera Rivera, T.E. Gill, K. G. J. H. M. B. R. F., “Wind modeling of chihuahuan desert dust outbreaks,”Atmospheric Environment 43, 347– 354 (2009).

[5] Rivera, N. I. R., [Detection and characterization of dust source areas in the Chihuahuan desert, SouthwesternNorth America ], Masters Thesis, Environmental Science and Engineering, University of Texas at El Paso(December 2006).

[6] N. Khazenie, T. L., “Identification of aerosol features such as smoke and dust, in NOAA–AVHRR datausing spatial textures,” Proc. of IEEE International Geoscience and Remote Sensing Symposium (IGARSS)(1992).

[7] A. A. Green, M. Berman, P. S. and Craig, M. D., “A transformation for ordering multispectral data interms of image quality with implications for noise removal,” IEEE Transcations on Geoscience and RemoteSensing 26, 65–74 (1988).

[8] A. Agarwal, H. El-Askary, T. E.-G. M. K. J. L.-M., “Hierarchial PCA techniques for fusing spatial andspectral observations with application to MISR and monitoring dust storms,” IEEE Geoscience and RemoteSensing Letters 4, 678–682 (October 2007).

[9] I. Koren, J.H. Joseph, P. I., “Detection of dust plumes and their sources in northeastern libya,” Can. J.Remote Sensing 29, No. 6, 792–796 (2003).

[10] R. O. Duda, P. E. Hart, D. G. S., [Pattern Classification ], Wiley Inter-Science, Hoboken, NJ, second ed.(2001).

[11] W.T. Freeman, E. A., “The design and use of steerable filters,” IEEE Transactions on Pattern Analysisand Machine Intelligence 13 (1991).

[12] Schowengerdt, R. A., [Remote Sensing Models & Methods for Image Processing ], Academic Press, Burling-ton, MA, seventh ed. (1997).

[13] Otsu, N., “A threshold selection method from gray-level histograms,” IEEE Transactions on Systems, Man,and Cybernetics 9, No. 1, 62–66 (1979).

[14] (http://www.ncdc.noaa.gov/oa/pod-guide/ncdc/docs/klm/index.htm).[15] P. S. Chavez, D. J. Mackinnon, R. L. R. M. G. V., “Monitoring dust storms and mapping landscape

vulnerability to wind erosion using satellite and ground-based digital images,” Aridlands News Letter 51.

