Coherent and random noise attenuation using the curvelet transform

Neelamani, Ramesh; Baumstein, Anatoly; Gillard, Dominique; Hadidi, Mohamed T.; Soroka, William L.

doi:10.1190/1.2840373

Cited by 234 publications

(83 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Kelemahan pada metode ini (F-X Deconvolution) adalah filter koheren yang digunakan belum tentu mampu melemahkan noise acak yang tersebar pada penampang seismik dengan energi yang acak sehingga menyebabkan event seismik yang terlihat buatan dan membuat interpretasi menjadi tidak unik [3]. Tidak hanya itu, penggunaan teknik konvensional filter seperti F-X Deconvoluiton dan Median Filtering menghilangkan jumlah noise yang banyak namun juga membahayakan komponen sinyalnya [4] sehingga perlu suatu filter untuk memisahkan sinyal dan noise yang baik.…”

Section: Pendahuluanunclassified

Aplikasi Transformasi Curvelet untuk Denoising Random Noise : Studi Kasus Data Seismik Sintetik 2D Darat “Antiklin"

Nugraha

Bahri²,

Warnana³

2018

JTITS

View full text Add to dashboard Cite

B46Abstrak-Denoising (reduksi noise) merupakan hal umum yang dilakukan pada pengolahan data seismik. S ecara umum, kita telah mengenal berbagai filter konvensional seperti bandpass filter dan metode signal enhancemet seperti F-X Deconvolution. Metode tersebut telah lama digunakan dan memiliki kelemahan yakni belum mampu mengembalikan reflektor secara maksimal dengan baik sedangkan noise memiliki nilai frekuensi yang sama dengan nilai sinyal yang baik. Pada penelitian ini, penulis menggunakan data seismik sintetik. Noise yang digunakan adalah gaussian noise karena noise ini merupakan noise yang digunakan secara umum dalam dunia statistik. S etelah ditambahkan noise, maka data sintetik akan diolah menggunakan transformasi curvelet. Hasil dari pengolahan data adalah data seismik sintetik "Antiklin" memiliki nilai signal to noise ratio (S NR) yakni antara 27 hingga 28 dB. Data seismik sintetik ketika diberikan noise memiliki nilai S NR yakni antara 19 hingga 25 dB. Dengan demikian, dapat dikatakan bahwa transformasi curvelet mampu melemahkan dan menghapus random noise. S ementara itu, reflektor pada data masih memiliki bentuk yang hampir sama dengan data forward modelling.Kata Kunci-Denoising, Transformasi Curvelet, Signal to Noise Ratio. I. PENDAHULUANENGOLAHAN data seismik merupakan langkah untuk memperoleh data dengan sinyal yang baik. Akuisis i data seismik yang dilakukan di lapangan kadang memberikan sinyal yang buruk sehingga menampilkan data seismik yang kurang baik. Data seismik perlu dilakukan berbagai macam filter sehingga akan mempermudah proses langkah selanjutnya. Salah satu tantangan dalam pengolahan data adalah menghilangkan sinyal buruk (noise). Noise terjadi akibat gelombang seismik yang menjalar pada bawah permukaan melewati berbagai macam perlapisan dengan nilai densitas yang bermacam-macam serta akibat gangguan angina, instrumen dan aktivitas manusia. Akibatnya, pada data seismik menghasilkan dua jenis noise yaitu noise acak dan noise koheren. Noise acak merupakan noise yang diakibatkan oleh kegiatan manusia meliputi instrumen, dan angina [1]. Noise koheren disebabkan oleh adanya kondisi geologi permukaan sehingga gelombang seakan-akan terjebak, seperti multiple, gelombang permukaan, efek ghosting, difraksi dan lain-lain.Filter umum yang sering digunakan dalam pengolahan data seimsik yakni band pass filter, F-X Deconvolution [2] dan lain-lain merupakan metode yang menggunakan skala tunggal, yaitu pada frekuensi atau waktu. Kelemahan pada metode ini (F-X Deconvolution) adalah filter koheren yang digunakan belum tentu mampu melemahkan noise acak yang tersebar pada penampang seismik dengan energi yang acak sehingga menyebabkan event seismik yang terlihat buatan dan membuat interpretasi menjadi tidak unik [3]. Tidak hanya itu, penggunaan teknik konvensional filter seperti F-X Deconvoluiton dan Median Filtering menghilangkan jumlah noise yang banyak namun juga membahayakan komponen sinyalnya [4] sehingga perlu suatu filter untuk memisahkan sinyal dan noise yang baik. II. TRANSFORMASI CURVELETTransform...

show abstract

Section: Pendahuluanunclassified

Aplikasi Transformasi Curvelet untuk Denoising Random Noise : Studi Kasus Data Seismik Sintetik 2D Darat “Antiklin"

Nugraha

Bahri²,

Warnana³

2018

JTITS

View full text Add to dashboard Cite

show abstract

“…On the contrary, wavelet basis functions are well-localized and therefore they can describe seismic data in a sparser manner and provide better denoising than the nonspace methods (Foster et al, 1994). Curvelets can efficiently model the geometry of waveforms (Candés and Demanet, 2005), and they have proven to be some of the most suitable predefined functions to represent and denoise the seismic data (Hennenfent and Herrmann, 2006;Neelamani et al, 2008). However, using a dictionary that is constructed from a predefined transform leads to a global approach for seismic data representation -a single dictionary is used to represent all the seismic features.…”

Section: Introductionmentioning

confidence: 99%

A method of combining coherence-constrained sparse coding and dictionary learning for denoising

2017

View full text Add to dashboard Cite

We have addressed the seismic data denoising problem, in which the noise is random and has an unknown spatiotemporally varying variance. In seismic data processing, random noise is often attenuated using transform-based methods. The success of these methods in denoising depends on the ability of the transform to efficiently describe the signal features in the data. Fixed transforms (e.g., wavelets, curvelets) do not adapt to the data and might fail to efficiently describe complex morphologies in the seismic data. Alternatively, dictionary learning methods adapt to the local morphology of the data and provide state-of-the-art denoising results. However, conventional denoising by dictionary learning requires a priori information on the noise variance, and it encounters difficulties when applied for denoising seismic data in which the noise variance is varying in space or time. We have developed a coherence-constrained dictionary learning (CDL) method for denoising that does not require any a priori information related to the signal or noise. To denoise a given window of a seismic section using CDL, overlapping small 2D patches are extracted and a dictionary of patch-sized signals is trained to learn the elementary features embedded in the seismic signal. For each patch, using the learned dictionary, a sparse optimization problem is solved, and a sparse approximation of the patch is computed to attenuate the random noise. Unlike conventional dictionary learning, the sparsity of the approximation is constrained based on coherence such that it does not need a priori noise variance or signal sparsity information and is still optimal to filter out Gaussian random noise. The denoising performance of the CDL method is validated using synthetic and field data examples, and it is compared with the K-SVD and FX-Decon denoising. We found that CDL gives better denoising results than K-SVD and FX-Decon for removing noise when the variance varies in space or time.

show abstract

“…As outras duas figuras ilustram esta situação. Na figura lc, a ffeqüência crítica cresceu de/, para/2, o que faz com que o espectro possa agora ser aplainado até esta segunda freqüência sem que o ruído predomine (Neelamani, 2008). Uma outra possibilidade, muito utilizada atualmente no mundo do processamento sísmico, é a conhecida deconvolução FX para fins de discriminação entre o sinal e o ruído.…”

unclassified

Detecção de camadas delgadas usando sísmica de reflexão

Farias

Freitas

Tygel

2008

RBG

View full text Add to dashboard Cite

Resumo A melhoria da resolução vertical no dado sísmico tem se constituído em um dos maiores desafios para o método de reflexão sísmica. 0 grande problema está relacionado à maior perda das amplitudes das altas freqüências, quando comparada às baixas, durante a propagação da onda. Isto faz com que na composição final do pulso sísmico a contribuição das baixas freqüências seja maior do que a das altas. Em princípio, isto não deveria se constituir em obstáculo para a recuperação plena do espectro, na medida em que o processamento sísmico dispõe de várias ferramentas para a equalização de todas as freqüências presentes no volume sísmico. 0 que impede que essas ações logrem êxito é a presença de um ruído com comportamento randômico, o que faz com que exista uma ffeqüência limite, ffeqüência crítica, a partir da qual não é possível recuperar o sinal de forma a efetivamente contribuir para o encurtamento do pulso no tempo. Tentativas de contomar este problema, a partir da atenuação do ruído aleatório, têm se mostrado insuficientes. A alternativa a estas tentativas e o aumento da multiplicidade de empilhamento. Como a perda de amplitude com a freqüência é uma relação exponencial, é necessária multiplicidade extremamente elevada para que o ganho seja efetivo. A capacidade de utilização de altíssimas multiplicidades é exatamente a grande virtude da técnica CRS (Commo# jze/ecíz.o77 S#7/`/czcG). Neste trabalho propomos a combinação do método CRS com o chamado balanceamento espectral (£pGcZ7~¢/ wãg./G7?z.7cg), para a recuperação de altas fi-eqüências, com o objetivo de aumentar ganhos na resolução vertical e, como conseqüência, propiciar melhor discriminação de reservatórios delgados. Os primeiros testes em dados sintéticos e reais apresentados neste trabalho são bastante encorajadores e pemitem concluir que a estratégia proposta tem bom potencial de aplicação prática.Palayras-clwve: Resolução sísmica, método CRS, fi'eqüência crítica.Abstract 7%7¢-/cz);e7~ c7eíecJz.o72 aísz.72g 7ie/ccZz.o7"é'z.s77%.cs.Improvement of vertical resolution in seismic data is one of the major challenges for the seismic reflection method. The main problem is related to the larger loss of high fi.equencies relative to the lower ones during seismic wave propagation. As a consequence, the contribution of lower frequencies in the final section is significantly greater than the corresponding one due to higher fi.equencies. In principle, this situation should not constitute an obstacle to the ftill recovery of the spectrum, because of the powerftl tools available in seismic processing, designed to equalize all ftequencies that are present in the data volume. The main reason why these tools in general do not succeed is the presence of noise, for example random noise. Such situation gives rise to a critical frequency, such that all fi-equencies above it are impossible to recover and, as a consequence, the seismic pulse cannot be made short enough to allow for optimal resolution. Methods to solve the problem by means of random noise attenuation have been ...

show abstract

Coherent and random noise attenuation using the curvelet transform

Cited by 234 publications

References 0 publications

Aplikasi Transformasi Curvelet untuk Denoising Random Noise : Studi Kasus Data Seismik Sintetik 2D Darat “Antiklin"

Aplikasi Transformasi Curvelet untuk Denoising Random Noise : Studi Kasus Data Seismik Sintetik 2D Darat “Antiklin"

A method of combining coherence-constrained sparse coding and dictionary learning for denoising

Detecção de camadas delgadas usando sísmica de reflexão

Contact Info

Product

Resources

About