Class Note for ECE 582 at UA-Comp Visn Dig Image Proc(2)
Class Note for ECE 582 at UA-Comp Visn Dig Image Proc(2)
Popular in Course
Popular in Department
This 15 page Class Notes was uploaded by an elite notetaker on Friday February 6, 2015. The Class Notes belongs to a course at University of Alabama - Tuscaloosa taught by a professor in Fall. Since its upload, it has received 25 views.
Reviews for Class Note for ECE 582 at UA-Comp Visn Dig Image Proc(2)
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 02/06/15
Image Spectral Enhancement IV Lecture 11 EE 482582 Digital Image Processing Spectral Filter Design Intuitionbased Design Highpass lowpass bandpass bandreject notchpass notch reject Implementation ideal Butterworth Gaussian Lapalacian Highboost Plomomorphic lter and highfrequencyemphasis er Modelbased Design Restoration Basic steps for filtering in the 5 frequency domain Frequency domain mienng npemnon v Filler inmse Fourier V V tr39insform mnmmquot FUN K Hm 39L tmnsrnnn Fquot V mu umn n We Postr processing processing my g v Input Enhanced image image noun 45 Basic sitn m l39llwring in lhc l39rt unenc unnnnn Basics of filtering in the frequency domain 1 multiply the input image by 1XY to center the transform to u M2 and v N2 if M and N are even numbers then the shifted coordinates will be integers computer Fuv the DFI39 of the image from 1 multiply Fuv by a filter function Huv compute the inverse DFI39 of the result in 3 obtain the real part of the result in 4 multiply the result in 5 by 1XY to cancel the multiplication of the input image 9915 Low pass filter high pass filter Mun u hmhinimmmihwmkhmi wnmwwihiiiimmmm Mimi mm mm iii m iiiiimximiiiii Add the 12 of filter height to F00 of the high pass filter FIGURE 43 Resuii oi highpass filtering his image in Fig 44m with he i39iiier in Fig 47c modificd by adding a constant uf mohair the filler height hi lhc filter funclkm Cumpm with Fig 44m Correspondence between filter in g spatial and frequency domains mm mm ii I I c a 7 s 39 incur w 39 I H I Smoothing Frequencydomain filters Ideal Lowpass filter HUL v I IUl y Dnm ahc noun 410 in Perspective plot of an ideal luupziss lilm transfer function h Fillcr displayed as an image 5 Film radial cross section image power circles l 3 aaaaaaaa u FIGURE 41 11 An image ul ic 5m x WU pik l And h its kuricr gtpctlrum Ihc gtupcrimpu ml circles nc rzulii nlucs ml 5 15 K1 NH and 230 which cnulusu 01M 041w unAJJannd WM M the Image pmwr rcspccliwl on 39U39 a ll Result of ILPF Min aaaaaaaa on quotIII a 48 UIIIIIH nu uauu u null III quota ll l Haaaaaa quotaaaaaaa FIGURE 412 1 Original im wu by Results nl39 idea Imxpnss l illcringy wim ruloff frequencies sol at mdii milk 01 5A 7 30 SH and 130 as shown in Fig 41 1h The rumor removed by lhcxc filters was x 1nd It quot u39 the mlal re pccliwly Example 3 b c a noun 4 m A requencydamam LPF of xadius s b Canespundmg spam mm mm me H39ngmgy 6 FM impulxes in me spanm damn mula ng he values nf ve pnels my Canmuuan n x u and m in me mum dumm Butterworch Lowpass Filter BLPF uquot a VG noun 414 n Pmpmixc plulnhlHutlcrwurlhlvmpuwfillertransferI39unrlinuIMHllcrLliV 1uLlnm imugc 1r Filler rudim mm wrlinn Murders lhmugh 4 ll at a l Example lllllllll ttaaaaaaa a tuld eauaaail 39a a Hill l ttaaaaaaa ttaaaaaaa tt tr Haunt us rttr Ultpmr lump rtrt ttt Rckmhul hlwnng wtm tu tttttu trtter 2 c d uh cul treatmm tt mutt t 5 li mm mm 170 M whmu m Hg 4 Hm t r tumpnrm rnhlul Spatial representation of BLPFs abcd mun m 21PM Spatial representation of BLPFs of nrdcr 11mm ZUde terresponatrtg graylme pretites through the center e1 the litters ten lters have a Culttfifrclllel1c nfi Note ttrttt ringing increase as a rttttetttm ol filter Order Gaussian Lowpass Filter GLPF Hum u ht am 1 FIGURE 417 H l crxpediw plnl M u FLI F mmle l unclinn h Fillcr dimluyed us un image c Filler mum mm gtucliun rm Valium mum at u Example 21 NH aaaaaaaa a llllllll a ll ll aaaaaaa a ll aaaaaaa L HHunwmv mm gm um Cumpuruwlh umm a ll aaaaaaaa an Rnwlh or mlcrmg MU mumquot mvprm n mdu wmnx m lt Hu mum h ahrvvm m h d r Example at FIGURE 419 a Sample 1ch HI poor resolution mole broken cha ems in nugnrricd vim n Result of lilluring mm a GLI F broken clmrddcr segments worm lninctl Historiially certain Computer programs were wrltten using only twa digits rather than four to deflne the applicable year Accordingly the company39s software may recognize a date uslng quot00quot as 1900 rather than tr 2000 Ea Historically certain computer programs were written using only two digits rather than four to de ne the applicable year Accordlngly the company39s software may recognize a date nsrng quot00quot as 1900 rather than the y r 2000 r h quotGUI Ma w Inner mm r lv39h x luknull ullllhnmmtlmlrl l39l mu a n nr ruhnlwuur m um lmnmllu nwumulmuum nlllrr mm 1 he FIGURE 41 1 Image lmwing pmmincnl czm lines In Result Musing n GLPF ilh Dn 3 u Rcwll m uaing n lLl lelh 1 m Original image man or NUAA Sharpening Frequency Domain Filter Ideal highpass lter M l ll L f l mu m 0 if Du V S DO r quotW 1 if Du v gt D0 7 Hm 1 l Huv Butterworth highpass filter 1 1 l H uv 1D0Duv2quot K 1 m Gaussian highpass lter 7D2 2D2 1Wquot Huv1 e W 0 I l J emu any mum 422 Tap mw Pelspecllve plul mage lepresenmtimL and cmss secnun of a typica ideal highpass ller dedls and banal mws39l391c same sequence m typical Butlerwnnh and Gaussian highpass lters 10 Spatial representation of Ideal Butterworth and Gaussian highpass g filters 21 ab mun 424 Resnlls iii rcspccliwly Problems win m 1g lhc image in Fig lllm with Dn 15 so and m it kvlklblll in a and h 11 Example result of BHPF 4 r ala aaflc ab new 425 Rosuns or ighpms mm g ha imagu in Fig mm using a BHPF of order 2 with 00 15 m and so rcspectiv lhesc rcsmls are mum smoulhcr than muse obtained with an ILPF Example rgsult of GHPF MHHHI r r r p r a a 11 1 L ubc FIGURE 41 Rusulh 1 highpzm mm 3n and m respuclivel 39ompam im 39 4 Hal using u 39rHI F 01 unlcr Z ih Du Sr 12 Laplacian in the Frequency domain 39 r a e I noun 1 21 m 391 pm rl Ia man n me quotname damnquot m hmgr lqurvmuhm ul m Lil aeranvn Wunmmmnammx Name rNDI Tu lv hlr7mimulnx1hnnlmrnumx Mm mums dpmhlrmmuglxmc mummummmunuwmsmmaa ExampleJap mum 423 21 Image of me Norm Pale cf the man by Laplaeaaa ltered mnge e laplncian image vealed d7 Image enhanced by using Eq 447127 Onginal amage eeunesy a NASA 13 Example highboost filter h c d mun Iquot mm uhp 34a hmuwm Mp Hn llmuzlm mu l39 m lupllx mm my Il 1mm I m 1 l In nhmml uung Lu NJ My mm 1 2 Mum n H mm mm A r um ulhwcwurln rl lrllnlmcl lull Imummm ul umw Examples a b c 0 FIGURE 430 a A chesl ery imaga b Rmu ofliullemorlh highpass filtering c Result ol highr lrequcncy emphasis fillering u Result ol performing histogram equalizaliun an r Original imaga Courtesy DnTlmmas R Cost Divigtinn ofAnzllnmiCal Sciences University of Michigan Medical School 14 Homomorphic Filter my gt m DFT mu 1 DFrr Hm 1 mu 7 FIGURE 43 Hmmmurpnic mm 39 1g uppmauh lnr image unhanccnmnl FIGURE 432 rms scclinn M u circularly s mmclric liner 1 mm the origin at the emeer trumlurm grin Result of Homomorphic filter a h FIGURE 433 3 Original in b magu processed by homummphic mmmm unto um side shelter SlockhanL 15
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'