ѧ
JOURNAL OF ASTRONAUTICS 
1999ꡡ204ڡvol20No.41999




СSARͼѹ
ۺϼ
ǿ־
ժҪĸһֻСѹۺϼ㣬ڱԭͼȵͬʱЧؽͰߵӰŵ㡣ҪǣѹͬʱУڶSAR״ͼĴȡõЧ
ʡСơͼѹ
THE WAVELET-BASED SYNTHETIC TECHNOLOGY 
OF DENOISING AND COMPRESSION SAR IMAGE
Li Qiang Wang Zhengzhi
(National University of Defense Technology . Changsha . 410073)
AbstractA Wavelet-Based compression and noise reduction synthetic technique is presented.The method is computional efficient and can significantly reduce the speckle while preserving the resolution of the original image.More importantly,it can integrate the noise reduction and data compression process,and yield good result on processing the SAR images.
Key wordsWavelet Noise reduction Image compression
1
ġʮӰңмչа͵ʷĿǰңмŸ߿ռֱʣ߹׷ֱԼѧ΢׶ңϵķչ
ңмЩչɵһңӴңݵѹܵӡңϢ޴ϸڽΪḻѹҪۺϿͼʱ()ͳԺƵƵԡС任һͬʱʱƵоֲԵı任ʮźŵѹʾԽ㷺ڸѹѹңһӦSARݵѹ1ݡ
΢ңԽԽܵΪṩѧңȫͬϢ⣬д͸ǿͲʱƵŵ㡣رǺϳɿ׾״(SAR)ڱЩŵͬʱ΢ңݵĿռֱʣʹ΢ңеӦôӴȫ۲⡢ʹ̽չ˾¡ҵؿ󡢺ͻȸ
SARңӦеһҪǰߵӰ졣ߵڳжͼһ·ԵҪģֳһָģʽԱΪһֵͼʵ2ݡͳİߵȥҪ֣ڰߵģ͵ȥͷֵȥСźŴΪߵȥṩһ˼·ͼͬʱͼϸϢ3ݡ
Сƺѹ涼õܡˣǿͨ÷ƺѹۺһͳһĹСͨSARݴʵ֤ǿжЧġ
2ڲƼ
һ㶼(1)ʽΪģͣy=x+e(1)
УxδӰźţyӰźţeƾҪһ(y)eźyƽͬʱźxeǲ(ռ䲻غx)Gaussianʱһͳƹ⡣þΪ۶(y)ĹֵӦʹСֱΪС(MSE)⡣
2.1SARͼİߵͳ
ף2ݷ˰ߵͳԣ֤˵ϵͳķֱ浥ԪĿϸںСұߵȾͼ㹻ʹһصĽʻ϶صĽʱߵģͻΪһֶArsenault4֤SARͼֵ任󣬰ߵΪGaussianSARͼİߵһСMSE⡣
2.2Сֵ
ٶģΪyi=f(ti)+zi,i=1,,n(2
fһδ֪tiǵȼĵ㣬ziǶֲͬı׼GaussianΪķСƷɷΪ
1)ݼ{yi}С任õСϵwj,k
2) 趨һֵ(nΪ)һõСϵֵ任
(3)
3)(2)ĽС任õԭźxĹ(y)
СֵƷص㣺С˷ͬźŵĹƲκйصĽṹڶڶƽԽϺõźţ÷õźŵĹƾнƵС۸÷ɱźŵıԵӾЧá
2.3㷨Ĺؼ
Ӧֵʱ漸Ҫǣ
СѡȡСѡҪעأ1)ԡСɶ߶ȷֽõĸӴݷֱ໥L2(R)2ӿռУʹӴԼСЧϺá˲ж߶ȷֽʱ׼ȷؽġġλ޳Ӧ˲ǲڵģʱһſΪ˫ʵ飬ڽʱ˫СЧһЩСıԵܵõСЧӦã2)(Regularity)СĹӦ˲鳤йأ˲ԽԽáСĹԶСǺҪģ̫ЧͲԣ̫ãƽͼϸڲ֣һѡ˲鳤Ϊ49ӦС
С任ȷΪ˽ͼߵ𿪣ҪһС任Ȼ任Ӧ˲鳤йءף5ѡȡ˳Ϊ4DaubechiesС任ȡ5ֻһҪС任ıԵЧӦѹһ˫Сұ任ȡ31ݡۺϿƺѹҪ󣬾ʵ飬ѡȡ˳Ϊ9˫С任ȡ3
۷ֵȷֵȷǱؼءһķδ֪ģйơף5ݸ˼ַһȡһСֽϵHHͼķΪķͼͳһķֵt=1.53ҽֵ任ף1ݶԴ˷˸Ľÿһͼֱ䷽i,b,iΪСֽļbΪÿһͼı(HHHLLH)ͼķֵȡΪ
ti,b=ˡi,bȡ0.30.6
֪С任༯HHͼķҲӦȡHHͼķǲַʵ֤Ƚϼ򵥶ЧáȡźyÿһάСֽϵHHͼķΪڸüķֽķͼķֵȡΪti,p=i,bi,HHЦi,HL=i,LH=.i,HHȡ1.051.35iHHiĵСΪÿһζ߶Сֽ⣬ƽһΡ⣬Զ߶ȷֽһLLͼֵ任Ᵽ֤ؽźŵľֵԭźš
3ͼ
ǲ˻С任ʸѹһΪϸĲ6ݣĲ׼صˡ̸̸ֻӦעķ档
ȣλ(Bits Allocation)ǻС任ıĹؼءС任ͼҪǲͬģҪͼҪ϶λ(Bits)ʾˣλһ£λС任ĹؼWi(bi)ʾԵiӴRiλʾʱӴʧ棬DΪʧ棬bΪҪﵽλRΪʵ

λʵϾҵһRi,i=1,,KʹRbʱDСһŻ⣬ͨճӷɽ

ЦiǸӴݵķϸƵοף67
ΣѹعͼʧǱȽϴġΪ˿˷һ⣬Niss8ݡWuGersho9ݵһϵеĲǽͼعʧ濴ͼعµģͨһ۵ƷȽԵظͼһ㣬Ľۣ10ݡ
4ϵͳ
ǰ˻СѹƼȻƹ̿ͼ1߿ʾʱ̽СΪͬʱСعͼʧЧӦѡʱ̢ڽлСֵ任ʵ֤ǺġǾͰѹͳһͼ1ʾһϵͳС

ͼ1SARͼѹۺϵͳ
˵ͼеͼعСۺָ任
5ʵ
ͼ2(a)еǷЧһSARͼͼĵľΪ52630EγΪ53145.7NͼֱȽԵĵ֡ݵغˮ塣ǽǰĻСֵֽƷӦڸͼǼֳõİߵ۱׼
ٱ׼ƫֵ(s/m)s/mںİߵСģ2ݡǼͼֵs/mֵڱ1
1ֵs/mֵ
ֲݵˮ
ԭͼ1.29450.64532.3178
ƴͼ1.01290.31241.7865
ƽͼ0.99140.31071.7903

ڶ׼ƫֱֵӰ׼˫㶨㷨(CFAR)ҪءֵĶ׼ƫֵڱ2
2ֵĶ׼ƫֵ
ֲݵˮ
ԭͼ4.89732.91743.1753
ƴͼ3.83211.32971.6493
ƽͼ3.83301.32981.6512

ɱɼߵƴs/mֵͶ׼ƫֵ½ˣҪİߵѱơͼ2ɼưߵͬʱͼϸõ

(a)ԭͼ(b)ǰƴͼ

(c)ƽعͼ(ʣ0.213bppPSNR37.15dB)
ͼ2С任ѹƽ
(ת24ҳ)
*оõӦصĿ֧
ߵλƼѧԶϵ302 . ɳ . 410073
ο
1Werness S A,Wei S CCarpinella R.Experiments with wavelets for compression of SAR data,IEEE Trans.on Geoscience and Remote Sensing,1994,32(1):197-210
2Goodman J W.Some fundamental properties of specklt.J.Opt.Soc.Am.,1976,66(11):1145-1150
3Guo H,Odegard J Eet al.Wavelet Based Speckle Reduction with Application to SAR Based ATD/R.Proc.ICIP,Austin,TX,Nov,1994
4Arsenault H HApril G.Properties of speckle integrated with a finite aperture and logarithmicaly transfomed.J.Opt.Soc.Am.,1976,66(11):1160-1163
5Donoho D L.De-noising by soft-thresholding.IEEE Trans.on Inform.Theory,1995,41(3)613-627
6Cosman P C,Gray R M,Vetterli M.Vector Quantization of Image Subbands:A Survey,IEEE Trans.On Image Processing,1996,5(2):202-225
7Shoham Y,Gersho A.Efficient Bit Allocation for an Arbitrary Set of Quantizers,IEEE Trans.On ASSP,1988,36(9):1445-1453
8Niss B.Prediction of AC Coefficients from the DC Values.ISO/IEC JTC1/SC2 WG8 N745,May 1988
9Wu S WGersho A.Enhancement of transform coding by nonlinear interpolative.Proc.SPIE Visual Commun.Image Processing,1991,1605:487-498
10ǿ־.СĽѹͼǿо
ո:1997714
޻:19971222
