IMAGE RECONSTRUCTION BASED ON FILTERED BACK PROJECTION ALGORITHM FOR MICROWAVE TOMOGRAPHY

• Main Author: Anis Sekar Ningrum, Fauzia
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/67873
• Institution: Institut Teknologi Bandung

Microwave tomography is a method for reconstructing the image of an object by using microwaves. In this study, an image reconstruction is applied on the cross- section of the tree trunk as an object to observe its porosity using Filtered Back Projection (FBP) algorithm developed from Back Projection algorithm with the ability to eliminate blurring in the result. The image reconstruction process uses test image data of cross-sectional images of the tree in different pixel sizes. In addition, image reconstruction also uses measurement data by processing signals at the receiving and transmitting antennas around the object. The antenna used is a microstrip array antenna that works on the Ku-band frequency with a 2x4 configuration. Assessment of the result of image reconstruction is performed in 2 ways, namely quantitatively using the Mean Square Error (MSE) and Structural Similarity Index Measure (SSIM) for the test image, and qualitatively for measurement data. The results for the test image without interpolation with a pixel size of 180 x 180 using Shepp-Logan filter produce an MSE value of 0.000403 and an SSIM value of 0.964823. While the reconstruction with Cubic interpolation yields an MSE value of 0.000342 and an SSIM of 0.960507. For image reconstruction with measurement data, it is shown that the same type of filter can clearly reconstruct the cross-section of tree trunk indicated by the difference in colour intensity.