#TITLE_ALTERNATIVE#

• Main Author: ASHINWIJAYA (NIM: 13206087); Pembimbing : Prof. Dr. Ir. Tati Latifah Erawati Rajab Mengko , TRIGUNA
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/17460
• Institution: Institut Teknologi Bandung

Industrial tomography has recently been used to measure the strength of materials, defection analysis, analysis of internal parts, and reverse engineering. A system to perform object restoration in reverse engineering is designed in this final project. The designed restoration system is expected to reduce noise arising from the reconstructed object. Object restoration process is done in two stages. First stage is the restoration of the slice images by using soft thresholding technique. Soft thresholding is obtained by performing wavelet decomposition and analyze each sub-band wavelet decomposition. On the application of soft thresholding used several parameters to achieve optimal restoration, which is mother wavelet type, wavelet decomposition level, and the median estimator divider value. There were 12 types of mother wavelet type used, ie db1, db2, db3, db4, symlet1, symlet2, symlet3, symlet4, coiflet1, coiflet2, coiflet3, and coiflet4. Wavelet decomposition is performed on level1, level2, level3, level4, level5, and level6. The divider value of median estimator are 0.5; 0.6745; 1, 2, 4, and 8. Selection is then performed to pick six best restoration method using soft thresholding that has been applied to 15 slice image samples, which represents the 5 image slices from the top, front, and side. After selection, 6 best mehods of soft thresholding is applied to all 256 slice images of the top, front, and sides, and the reconstruction of the object is made. In an effort to maximize restoration system that is designed, logic operations AND and OR logic and arithmetic operations Mean is conducted on object reconstruction. The second phase is to implement morphological operations by using opening technique. In implementation of this thesis, three-dimensional opening run using several steps of erosion and dilation image analysis from three areas: front, side and top. <br /> <br /> <br /> This sistem has been successfully applied to the restoration of the projected image obtained from Pengembangan Sistem Simulated CT Untuk Keperluan Rekayasa Balik dan Rekonstruksi 3D Komponen Industri research. As optimization parameter of system that is designed, reference object is used to measure the number of error after the designed restoration is performed. Then the number of errors using the soft thresholding restoration method is compared with the number errors using the hard thresholding restoration method. The best restoration results are obtained by using soft thresholding with mother wavelet coiflet4, with depth of decomposition six levels and the divider of median estimator is 0.5 and opened by square-shaped structuring element with size of 11 pixels. The number of errors with that method is 481,177 pixels, while the best result on the hard thresholding method is obtained when the threshold value is 107 with 521,737 errors.