I claim:
1. An image restoration method in a system having an image sensor and a digitizer, comprising the steps of:
(a) providing an imaging system described by an integral operator P;
(b) transmitting an ideal image f(x,y) through an image blurring and degrading transmission medium to provide a degraded image g(x,y) wherein the degraded image g(x,y) may be represented as Pf=g;
(c) receiving said degraded image g(x,y) from said transmission medium by said image sensor;
(d) digitizing said degraded image g(x,y) by said digitizer;
(e) transforming said degraded image g(x,y) to provide a time modified representation w(x,y,t)=P.sup.t f wherein w(x,y,0)=P.sup.0 f represents the ideal image f(x,y) at time t=0 prior to the operation of the integral operator P upon f and w(x,y,1)=Pf=g represents the degraded image g(x,y) when the image is received by said image sensor;
(f) requiring the magnitude of the difference between the ideal image f and a blurred version of f to be less than a preassigned tolerance value by minimizing .parallel.f-P.sup.s f.parallel. wherein P.sup.s f represents an image at time t=s as s approaches 0;
(g) determining a plurality of values of w(x,y,t) in accordance with said minimizing;
(h) adjusting said image in accordance with said determined plurality of values to provide a plurality of adjusted images;
(i) selecting an adjusted image of said plurality of adjusted images; and
(j) displaying said selected image.
2. The image restoration method of claim 1, comprising the further steps of:
displaying a sequence of partially restored images represented as w(x,y,t) for a plurality of values of t as t approaches 0; and
determining an optimum image of said plurality of partially restored images in accordance with said displayed sequence.
3. The image restoration method of claim 1, wherein said constraining comprises the step of imposing the constraint .parallel.f-P.sup.s f.parallel..ltoreq.K.epsilon., where K is a constant and .epsilon. is representative of at least one image restoration parameter.
4. The image restoration method of claim 3, comprising the further step of imposing the constraint .parallel.Pf-g.parallel..ltoreq..epsilon., wherein .epsilon. is representative of at least one image restoration parameter.
5. The image restoration method of claim 4, comprising the further step of imposing the constraint .parallel.f.parallel..ltoreq.M, wherein M>>.epsilon..
6. The image restoration method of claim 5, comprising the further step of determining a restored image f(x,y) which minimizes the quantity
{.parallel.Pf-g.parallel..sup.2 +.parallel.(.epsilon./M)f+(1/K)(f-P.sup.s f).parallel..sup.2 }.
7. The image restoration method of claim 6, wherein a plurality of images w(x,y,t)=P.sup.t f are determined for a corresponding plurality of values of the time t.
8. The image restoration method of claim 7, wherein said plurality of values of time comprises a sequence of progressively smaller values.
9. The image restoration method of claim 8, wherein the values of said restoration parameters are adjusted in accordance with said plurality of images for successively smaller values of time.
10. The image restoration method of claim 7, wherein each image w(x,y,t)=P.sup.t f is determined in the Fourier Transform domain by w(.xi.,.eta.,t)=p(.xi.,.eta.).sup.t f(.xi.,.eta.), wherein w(.xi.,.eta.,t) is the Fourier transform of w (x,y,t , p(.xi.,.eta.) is an optical transfer function of said system, and f(.xi.,.eta.) is the Fourier transform of f(x,y).
11. The image restoration method of claim 6, wherein said restored image f(x,y) is determined through algebraic operations performed in the Fourier Transform domain.
12. The image restoration method of claim 1, wherein f(.xi.,.eta.), the Fourier Transform of f(x,y), is determined by f(.xi.,.eta.)=g(.xi.,.eta.)/{p(.xi.,.eta.)[1+(.mu.K.vertline.p(.xi.,.eta.).vertline.).sup.-2 (1-.mu.p(.xi.,.eta.).sup.s).sup.2 ]}, wherein g(.xi.,.eta.) is the Fourier Transform of the degraded image g(x,y), p(.xi.,.eta.) is an optical transfer function of said system, .epsilon. is representative of at least one constant, K is a constant, .mu.=(1+K.epsilon./M).sup.-1, M>>.epsilon. and s is a substantially small value of time.