Radar tomography using Doppler-based projections

Image reconstruction using tomography is based on inverse filtering a series of projections, collected at angular positions surrounding an object or scene, where the form of the projection is derived from the measurement geometry (monostatic/bistatic, point/plane transducer/sensor) as well as the probing phenomenology. Quality image reconstruction requires that sensors are positioned to adequately sample target/scene Fourier space. In a remote sensing application this is done by sequentially positioning a limited number of sensors in the half-space above the target region. Typically projections are formed by integration along the locus of points of constant delay (the line integration path). Such projections are defined purely spatially. In this paper we define and evaluate projections that are derived from the Doppler relationships of the measurement geometry and temporal integration. The circular aperture is subdivided into overlapping SAR subapertures. This has the effect of creating fan beam-like projections, in 2D the projections linearly diverge from the source point with the formation of an image referenced to scene center effects the receiving array. Images are reconstructed using back projection filtering. The Doppler-based phenomenology is demonstrated and point spread function(s) generated. The advantage of this technique is the production of high resolution 2D, and extendable to 3D, images using very low bandwidth. This is applicable to low frequency waveforms operating in spectrum challenged environments or where communications bandwidths do not support wideband data transfer.