Automatic Recognition And Quantitative Evaluation Study On Electric Image Logging

As an advanced well logging technology, electric image logging shows us the 2D image of formation directly. Its appearance makes us learn about the downhole information more deeply and easily. But until now, about the identification of fractures and vugs, we accomplish that in the way of human-computer interaction, which is affected by subjective factors easily and not very effective. So in order to use computer more effectively, it is rather necessary to recognize fractures, vugs, lithology and so on automatically.According to electric image logging, the paper makes use of some computer image processing technology to preprocess images and identifies fractures vugs, then calculates their parameters to evaluate them quantitatively. Otherwise, according to the different characters of different lithology images, we could recognize the lithology automatically.In the respect of fracture identification, making use of Hough transform, the paper recognizes fractures and calculates their parameters such as dip, deviation, width, fracture porosity and so on. In addition, in order to identify the other kind of fractures and correct the automatic recognition results, human-computer interaction is used here.In the respect of vug identification, region marking, contour extraction and Freeman code are used here to recognize vugs and calculate their parameters automatically. For the purpose of recognize gravels and noddles, human-computer interaction is also used here.About lithology recognition, the paper extracts five feature curves such as average, variance etc. and makes use of correlation calculation and neural network to recognize lithology. Finally, the paper programs with the C# language and gain final calculation results.