About This Item

Share This Item

The AAPG/Datapages Combined Publications Database

AAPG Bulletin


Volume: 49 (1965)

Issue: 7. (July)

First Page: 1085

Last Page: 1085

Title: Optical Methods and Equipment for Processing Geophysical Data: ABSTRACT

Author(s): Arthur L. Ingalls

Article Type: Meeting abstract


In recent years optical data processing has become popular for solving a number of problems. One of the biggest advantages of optical data processing is the high data rate which can be achieved. For example, a sq. in. of film may easily store as many as 1,000 elements in each of 2 directions. This gives a total of 106 elements per sq. in. The data may be recorded as 2-dimensional data or may be in the form of channelized data. Several operations are possible, such as spectral analysis filtering and cross correlation. The equipment for optical data processing has a high degree of versatility and stability. The versatility arises from the fact that one type of equipment may be converted to another type simply by rearrangement or addition of elements. In-put and o t-put equipment can be in electrical form, but more commonly is photographic in nature, the in-put being a transparency while the out-put is a picture.

In spectral analysis 2-dimensional data may be analyzed to obtain the power spectral density for the whole area of interest, or spectrum analysis may be obtained for a small area. This has the advantage of excluding information from the spectrum which is not of momentary interest in favor of a more precise and uncluttered spectrum of the area of interest. Such a small area may be moved in scanning fashion to search out areas of particular interest. The equations involved may be in either X, Y coordinates or R, ^Thgr coordinates, whichever lends itself to the best mathematical analysis.

LaserScan is a 2-dimensional filtering equipment in which 2 Fourier transforms are performed sequentially on the in-put transparency. The first transform produces a spectrum which may be filtered. The second produces a reconstruction of the original data after spatial filtering which permits photographing the filtered seismic section. Directional filtering may be performed to remove dips in undesirable directions, thereby improving the data of interest. Time filtering also is possible, either as band rejection with straight wires, or high-cut filters by means of knife-edge plates operating in the time direction. By this means, "ringing," or high-frequency noise, may be removed which is common to all channels.

It is also possible to perform one-dimensional spectral analysis simultaneously for many channels. For this configuration a cylindrical lens is added to the optical system to focus channels in the transparency into channels in the out-put plane. In this case, the final lens projects the Fourier transform of the data in each signal channel into a corresponding channel onto the final out-put film.

Cross-correlation also may be done by optical means. In general, this is one-dimensional, multi-channel instrument in which the recorded channels are interrogated by an expected signal. The out-put then becomes a display of the correlation functions obtained along each channel for all the channels in which the operation is taking place. In some cases it is preferable to cross-correlate filtered data. In such cases filtering and cross-correlation are performed on the same optical equipment by changing some of the optical elements. It may be that the signal function needs to be filtered. Conversely, in some cases it is desirable to cross-correlate record sections against themselves while first filtering one of the sections used as a reference function.

Various types of coded pulses can be used to obtain reference functions. These codes include Vibroseis, pseudo-random codes, and even dynamite shots which also can be classified as a coded impulse. It must be remembered that the coded signal may be altered for spectral content during its transit time in the earth's medium. In general, therefore, it is better to use an altered signal for the reference function than to use the initial source or code wave form. In general it is best to use coded pulses having a sharp auto-correlation function.

End_of_Article - Last_Page 1085------------

Copyright 1997 American Association of Petroleum Geologists