Reduction of sidelobe and speckle artifacts in microwave imaging: the CLEAN technique

Large random thin arrays provide a high angular resolution microwave images but are plagued with artifacts (false targets and target breakup or speckle) caused by high sidelobe levels. The CLEAN algorithm for reducing the sidelobe-induced artifacts is extended to the coherent radiation field and the theory placed on a quantitative basis. The CLEAN technique decomposes the received echoes of a coherent multiple-target scene by iterative cancellation of the largest target found. At each step, cancellation information is used to create a target image. The image includes target intensities, phases, and directions. The process is designed for an imaging instrument consisting of a random thinned array. A condition is derived which, when satisfied, guarantees that all proper targets will be preserved in the cleaned image and all false targets discarded. An algorithm involving moving thresholds is derived to extract the target coordinates. It is shown that targets much weaker than the sidelobe level can be detected and displayed without the hazard of artifacts. The target dynamic range and the image contrast can be increased by up to twice the signal-to-noise ratio per element.<>