An Applied Research Of Image Analysis Technology In Characteristic Parameter Extraction Based On Rock Slice Thin Images

The upper palaeozoic gas reservoir of PCOC (China petroleum Changqing OilfieldCompany) is one of the main exploration area in the Erdos basin, the upper palaeozoicsandstone has a characteristics of low porosity and low permeability, whose rock types,granularity, pore types and argillaceous content have a direct impaction on porosity,permeability and other important parameters, so, the study of reservoir pore structurecharacteristics in Erdos basin will contribute to predict the geological reserves and to point outthe development direction of oil and gas.This thesis focus on the pore identification of rock slice images. Starting with image analysis technology, de-noising pretreatment by adopting the method of spatial filtering and spectral filtering for image. The next step is to carry out two-dimensional entropy threshold segmentation and edge detection operation, in order to recognize and calibrate the red pore in the image. Then calculating characteristic parameters of related pore, such as pore number, pore size, pore perimeter, pore area, face rate, pore shapefactor. pore diameter etc. Finally, we can regard extracted characteristic parameters byapplying image processing as input, and the existing of porosity and permeability asoutput respectively, using Regression analysis tools to do data fitting. It is concluded that porosity and permeability parameters have a high relationship with Number of pore,pore area and average pore diameter, so the research area in reservoir pore structurewill help to predict geological reserves.