Conventional seismic imaging and data processing procedures are laborious and demand meticulous attention to detail in order to complete each stage of the process. High-power computers must handle vast amounts of raw seismic data using a variety of signal processing techniques in order to provide an interpretable seismic image of the subsurface. Each processing step usually applies different adjustments depending on the time inside the seismic trace, source event, source-receiver offset, and locations within the survey region. Because of this, the seismic processor typically needs to do a laborious analysis of the data set in order to choose the right parameter variables for each processing step. Reference books on seismic data processing have already provided step-by-step instructions (processing sequences) and their intended results. [1-4]. In this short communication, these steps would not be discussed in detail, although some key steps might be given cursory consideration in the course of the communication. The emphasis would be on three major principal goals for seismic reflection data processing and imaging.

Enhancing or improving the temporal resolution of the seismic data is the primary objective of seismic processing. A seismic trace is defined by the convolution of an input seismic signal (the source) with an earth model made up of discrete reflectors in the convolutional model of reflection seismology. Therefore, one goal of seismic processing is to extract the idealized impulse response of the earth model by subtracting the source's character from the data; this procedure is called deconvolution. If all goes well, each seismic trace is converted into a time series of impulses or spikes, whose amplitudes correspond to the reflectors' reflection coefficients and whose arrival timings indicate the primary reflection times of all reflectors, as well as amplitudes that show the reflectors' respective reflection coefficients. Reverberations of the source signal, source and receiver ghosts, or reflections from the air-water interface, and multiples in the seismic data from wave fronts that have undergone multiple subsurface reflections are some of the propagation effects that must be eliminated in order to get the data to resemble the impulse response model. Frequency filtering, spiking deconvolution, predictive deconvolution, and multichannel coherent noise filtering are a few of the deconvolution and multiple suppression processes used in a typical processing sequence [3,5].

Improving the low signal-to-noise quality characteristic of seismic data is a secondary objective in seismic data processing. This is the main reason for deploying as many receivers per shot as is practical and gathering somewhat redundant data when it comes to acquisition. Traditionally, the common midpoint, or CMP method, sometimes mistakenly referred to as the common depth point method, is used to process these "excess" data. Seismic traces are categorized into CMP gathers during CMP processing based on shared source-receiver midpoint bins. For a few reflection events in the gather, velocity functions are computed for a subset of CMP gathers depending on arrival time variations as a function of source-receiver offset for a few reflection events in the gather. CMP velocity functions are then interpolated throughout a survey area to construct a velocity model of the subsurface. This velocity model is used to perform normal move-out (NMO) corrections throughout the survey. NMO is a non-linear stretching of the seismic time axis to remove the travel time component due to source-receiver offset. NMO is applied to each trace in a gather so that the reflection travel times on all traces approximate that of a trace with zero source-receiver offset (a coincident source and receiver). After NMO, a CMP gather's traces can all be added together, or stacked. Reflection events on the several traces will sum constructively if the underlying geology does not significantly contradict the CMP method's assumptions. This will result in a single trace with a signal-to-noise ratio significantly higher than that of the individual prestack traces. Repeating this process for each CMP collect in the survey results in a substantially smaller poststack data file set with significantly better signal quality than the prestack data set, achieving the second crucial objective of seismic data processing and imaging. It is important to note that seismic data processing and acquisition have been extremely effective using the CMP approach. The CMP approach has been the foundation for almost all reflection seismic surveying since the early 1960s. Nevertheless, the CMP technique makes two very serious oversimplifications explicit: that all subsurface reflectors are horizontal and that the seismic velocity is constant. By adjusting for the impacts of dip, a processing technique known as dip move-out (DMO) preconditions prestack data for CMP processing. In regions where the geology deviates from the method’s assumptions, the accuracy and applicability of the CMP approach have been significantly increased thanks to DMO processing.

Enhancing the lateral resolution of the seismic data is the third objective of seismic data processing. Seismic stack sections or volumes with a single seismic trace representing each surface point are produced by CMP processing. The stack looks like a picture of geology when it is presented. However, as the seismic wave-field propagates through the subsurface, diffractions and spherical spreading generate distortions, which is why this initial image is not a true reflection of true geology (inaccurate). Seismic migration must be used to convert the seismic data into a subsurface image in order to create an image that is easier to interpret and such that it approximates the actual geology of the subsurface. Inverse wave scattering calculations called migration are used to move seismic reflections and diffractions back to their original locations. Unlike the primarily one-dimensional processes that were previously discussed, migration is intrinsically a 2D or 3D procedure since it distributes data laterally in the image region. When it comes to intricate three-dimensional structures, 3D migration yields a seismic image that is more precise than 2D migration. Naturally, 3D migration is limited to data collected in three-dimensional surveys. A large portion of the seismic energy recorded in the seismic profile during a two-dimensional surface over complex structures is actually reflected by out-of-plane reflectors (the plane being the vertical section underneath the profile). This energy can be eliminated from the profile and energy that was reflected from inside the plane but captured on a different plane can be restored using 3D migration if 3D data are available. 

Seismic traces that share the same source-receiver midpoint coordinate are added together to create each trace of a stacked seismic data set. Appropriate NMO functions must be applied to all common midpoint gathers in order for the delay timings of reflection events to line up for proper stacking. The uncorrected reflection events are typically assumed to lie along hyperbolic curves in NMO corrections. This is roughly true for layered media; it is strictly true only if the earth is a constant velocity medium above the reflector. Events in complex structures might not sit on a hyperbola, which prevents them from moving out and stacking correctly. Events with varying stacking velocities can also manifest on CMP gathers at the same travel time. In such cases, it is necessary to perform migration on the individual seismic traces before rather than after stacking. Prestack migration simultaneously improves both the lateral resolution and the signal-to-noise quality of seismic image; all the data contained in the individual traces are available during imaging, whereas stacking may destroy information that only appears at certain offsets. Prestack migration thereby replaces the functions of both stacking and poststack migration. In current practice, however, CMP processing is routinely applied before the data are reprocessed using prestack migration. Prestack migration has certain drawbacks as well as clear advantages over post stack migration. The magnitude of prestack data volume that needs to be migrated is a drawback of prestack migration. The data volume for prestack migration is 100 times bigger than the corresponding volume for poststack migration, for example, if a gather has 100 traces. The turnaround time for such methods is impacted by the significant demands the 3D data sets place on computer speed and memory due to their increased size. A salient feature of both prestack migration and the CMP technique is that a subsurface velocity model is generated as a byproduct of both processes. Although the derivation of the velocity model was not emphasized as a processing goal, it has obvious interpretational significance. In prestack migration, velocity analysis algorithms allow processors to improve their velocity models between migration iterations. Prestack migration programs which produce a migrated offset gather at each CMP location are particularly useful for certain velocity estimation strategies. This common image gathers (CIGs) allow the processor to perform velocity analysis using relatively familiar methods.

An attempt has been made to describe the basic principles of seismic data processing and imaging and, most importantly, the end-goals (or the intended deliverables from a typical processing and imaging assignment or project). This communication is not in any way exhaustive, as some common and important operations in seismic data processing like trace editing, muting, correction for spherical divergence, gain application, and static corrections were not mentioned. We conclude this short communication by saying that processing requirements for separate seismic surveys may differ greatly. A seismic data processor needs, therefore, to display a high level of creativity and flexibility in designing the processing sequences to address specific problems and goals in line with the peculiarities of the data set that he wishes to process to achieve desirable imaging outcomes.

There is no way one can write on this subject without leveraging existing knowledge sources. We acknowledge the pioneers of seismic data processing, imaging, and analysis of our generation (Jon. F. Claerbout and Oz Yilmaz). These are great men, and we grew up reading their reference books to understand and gain valuable insights on the act of seismic data processing, imaging, and analysis. Another geophysicist whom we draw a lot of inspiration from is Prof. Biondo L. Biondi (a SEG prestigious award recipient and a renowned 3D seismic imaging specialist). Other authors whose books, notes, briefs, and journals have been consulted in this buildup are all thankfully acknowledged.

1. Claerbout JF. Fundamentals of geophysical data processing. BMcGraw-Hill publisher, New York. 1976.
2. Claerbout JF. Fundamentals of geophysical data processing: with applications to petroleum prospecting. Blackwell scientific publications. 1985.
3. Yilmaz O. Seismic data processing, Society of Exploration Geophysicists – SEG reference book/manual. Tulsa, USA. 2001.
4. Biondi BL. 3D seismic imaging. Investigations in Geophysics Series No. 14. Society of Exploration Geophysicists – SEG, Tulsa, Oklahoma, USA. 2006.
5. Cary P. Reflections on the deconvolution of land seismic data. Recorder- Official publication of the Canadian Society of Exploration Geophysicist – CSEG. 2006; 31.