Accuracy Assessment of Photogrammetry Surface Reconstruction for Improving Microwave Imaging

We have conducted an accuracy assessment of the recently proposed 3-D photogrammetry-based surface reconstruction of the breast shape being imaged via microwave measurements. The technique features simplicity and high accuracy. The results for the reconstructed surfaces of various shapes show that this technique is very accurate for both canonical and random shapes. The maximum error is in the order of only 2 mm and occurs for surface containing near-singular points. The S-parameter measurement in the whole UWB range agrees very well with the forward model based on photogrammetry shapes and HFSS simulations. This is encouraging for our future efforts to employ photogrammetry in solving microwave imaging problems using both frequency-domain and time-domain forward models.