IMPLEMENTATION ITS MICROWAVE TOMOGRAPHY AND IMAGE RECONSTRUCTION USING COMPRESSIVE SENSING METHOD

• Main Author: Muliyani, Sri
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/67874
• Institution: Institut Teknologi Bandung

Microwave tomography is a technique for observing internal structure of an object using microwave emitted for various angles. In this research, image reconstruction using microwave is carried out for the detection of porosity in a tree trunk. The compressive sensing (CS) method is used in the image reconstruction by applying the least absolute shrinkage and selection operator (LASSO) function to increase the accuracy of linier regression model. This research consists of image reconstruction with test image using cross-sectional image of a tree trunk and image reconstruction using data measurement at the frequency band of 2 GHz – 4 GHz. Quality of assessment of the image reconstruction results in this study is performed objectively using image parameters based on mean square error (MSE) and structural similarity index measure (SSIM) for the test image, and subjectively by analyzing the visual image quality for the measurement data. From the reconstruction process, accurate test image reconstruction results are obtained for the image size of 128×128 pixels and an alpha of 0,001 with an MSE value of 0.0011 and an SSIM value of 0.9019. The image reconstruction using the measurement data at the frequency of 3 GHz is able to describe well the cross section of the tree trunk and can represent a clearly color intensity.