A general method for solving the inverse diffraction problem is presented. It is based on an identity of Bojarski which states that

and

are a Fourier transform pair. Here

is the characteristic function of the target (

inside the target,

outside),

is the frequency,

is a unit vector specifying the aspect, and

can be obtained by measurement of the backscattered electromagnetic far field at frequency

and aspect

. If data is obtained in any subset

of

space, the method yields partial or complete information about the target geometry. It is used to rederive earlier results very simply and to obtain a significant new solution, in which the target geometry is completely determined using frequencies only in a practical frequency band and aspects in a narrow cone.