Pemrosesan data GPR (Ground Penetrating Radar) yang telah dilakukan di gedung A dan gedung B Bank Indonesia masih belum sempurna. Dari hasil pemrosesan data, masih terlihat ringing/background noise berupa terhapusnya amplitude yang menerus hingga lapisan yang lebih dalam. Akibat dari Ringing/background noise ini, kemenerusan/kontinuitas lapisan secara horisontal menjadi buram. Dalam penelitian ini, penulis membandingkan metode penghilangan ringing/background noise dengan berbagai metode yaitu background removal, bandpass filtering, bandpass-butterworth filtering, f-k filtering, dan radon transform. Hasilnya, metode yang lebih efektif yaitu menggunakan radon transform (Kim et all., 2007) yang terlihat dari kemenerusan/kontinuitas lapisan secara horisontal semakin baik. Setelah kemenerusan/kontinuitas lapisan secara horisontal semakin baik, penelitian diteruskan dengan melakukan pemodelan perlapisan bawah permukaan berupa pemodelan litologi dan pemodelan lapisan bedrock pada data GPR dengan mengacu kepada data kecepatan rata-rata GPR dari data WARR (Wide Angle Reflection Refraction), dibandingkan dengan data kedalaman litologi maupun data kedalaman SPT (Standard Penetrarion Test) sumur sehingga data twt (two way traveltime) dapat dihitung untuk dilakukan picking horizon pada batas litologi maupun batas bedrock. Kontras litologi maupun kontras lapisan bedrock pada data GPR sendiri tidak cukup kuat, namun dengan mengacu kepada kecepatan WARR dan kedalaman dari data litologi maupun data SPT sumur, posisi twt reflektor dapat ditentukan dan pemodelan perlapisan bawah permukaan dapat dilakukan.

---

GPR data processing (Ground Penetrating Radar) that has been conducted in building A and B of Bank Indonesia was still imperfect. The result of data processing still shows the ringing / background noise that constitutes the elimination of continuously infiltrated amplitude to the deeper layers. Due to this Ringing / background noise, the horizontal layer of continuity becomes opaque. In this study, the author compared the methods of removing ringing / background noise with different methods: background removal, bandpass filtering, bandpass-butterworth filtering, f-k filtering, and radon transform. The result shows that the more effective method is using radon transform (Kim et al., 2007) as seen from a better horizontal layer of continuity. After the horizontal layer of continuity became better, the study continued to do the modeling of the surface bottom layer in the form of lithology and bedrock layer modeling based on GPR data which refers to average velocity of GPR compared to WARR (Wide Angle Reflection refraction) data. If it is compared to the depth of lithology data as well as the depth data of SPT (Standard Penetrarion Test) wells, so the data of TWT (Two Way Traveltime) will be able to be calculated for picking horizon at the limit of both lithology and bedrock. Both of lithology contrast and bedrock layer contrast on GPR data are not strong enough, but referring to the velocity of WARR and the depth of both lithology data and SPT well, TWT reflector position can be determined and the modeling of surface bottom layer can be conducted.