Texture And Structure Analysis Of Igneous Rock Based On FMI

As the development of petro exploration, more and more igneous hydrocarbon reservoir discovered around world. As an especial type of reservoir, it comes in for an great deal of attention. Log evaluation of igneous hydrocarbon reservoir a new and important field of study follows study of sandstone and carbonate rocks hydrocarbon reservoir. Identify lithology, lithofacies and lithofacies sequence accurately, is the important content of log evaluation of igneous reservoirs, it's foundation is lithologic identification, how to improve accuracy rate is the problem of lithologic identification in evaluation of igneous reservoirs.The fine differentiation of igneous, can accurately classify eruptive sequences, correlate formation, construct geology model, provide important facts for evaluation of reservoirs. At present, lithologic identification of igneous rocks is based on the composition of igneous rocks, such as crossplot method and microelement well logging, but never use the texture and structure information.FMI is promoted by Schlumberger in 1990s, it was widely used around the world. Resolution of FMI is up to 5 mm, it can provide image of resistivity that has fine detail, named as"underground stratigraphic microscope"by geologist. FMI reflect texture and structure information of rock that underground as an image, texture and structure of different rock is displayed as different image texture. By means of identify the texture that reflect texture and structure of igneous rocks, and associate these characters with others logging characters which can reflect composition of igneous rocks can identify lithology effectively, especially complex rock-types. For complex lithology (such as igneous rocks), identify rock-types only by composition is ineffective, by means of identify the texture and structure information from FMI data, is an important method of identify complex lithology.Digital image processing is originates from tow fields: first, in order to make the analysis of image information easily; second, for the reason that automatic understood by computer, and then storage, transfer, display the image. Texture is a common concept used in digital image processing. Texture analysis by statistics, texture is considered as quantificational measurement of the density distribution of image region. As use statistics in texture analysis of digital image, positional operator is the key factor. This paper use positional operators that rightward one pixel(Wr) and downward one pixel(Wd), finally, use the tow parameter(Wcr and Wcd) of these tow direction as descriptors. These tow descriptors are distinguishing obviously with igneous rocks.ANN is a net that consist of large numbers of process unit (neuron). It use device that realizable in physic or simulate structure and function of neural net of the organism by computer, it can conquer problem that can't solved by computer and others system, such as study, control, recognition, expert system etc.Texture statistic analysis based on FMI can give quantificational descriptors of the FMI, and these descriptors distinguishing obviously. It can be considered descriptors of igneous'texture and structure, use these descriptors and pattern recognized in ANN can improve the accuracy rate and efficiency of igneous lithologic identification.