A Statistics Methodolody Study On The Oil-Water Distribution In Rock Core Based On Regional Growth And Three-Dimensional Matching Algorithm

With the high-speed development of computer hardware and software,the related computer technology is widely applied in all walks of life.An effective combination of these technology and related industries can greatly reduce the workload and improve work efficiency;therefore,as the national lifeblood,oil industry makes full use of the computer technology.It is known that oil stores in the tiny pores in the rock,and most of the oil and water is mixed together and not easy to exploit,as a result,the water injection method is usually adopted to make the oil mix together and enhance oil recovery.The separation and the statistical distribution of oil and water,and analysis of the relations and movement rule between oil and water become crucial.Therefore,the statistics of the number and shape of oil and water in rock in today's oil exploration plays a very important role and becomes a research focus in the oil industry.Different algorithms are presented in this paper from two kinds of statistical perspectives: the number and the volume of oil and water and the shape of oil and water respectively.First is the statistics of the number and volume of oil and water in rock.A mass of oil or a drop of water can be considered as a full-bodied.If separate connecting sand bodies is labeled,it is called connected components labeling.The labeled number is the number of connected region,namely,is the number of oil or water droplets.Region growing method is applied in the experiment of this paper,that is to put,tracking from a pixel to all the pixels.Through comparing with and regional statistical algorithms,the statistics result of region growing method is superior to the regional statistics.The second is the statistics of the shape of oil and water in rock.This paper is based on the three-dimensional shape matching algorithm to complete the experiment.The cornerstone of shape matching algorithm is to determine the shape representation;the selection of feature descriptor is the basis of shape representation.Good feature descriptor leading to higher shape matching accuracy,therefore,the feature descriptor is a very important link in this paper.The author firstly predefines several kinds of three-dimensional shape of oil and water: cylinder,sphere,cuboid,film bulk,etc.if some special shapes can't match with the above predefined shapes,it will give special definition separately.Gradient direction histogram is selected as the experimental feature descriptor,then the shape of gradient direction histogram direction of each pixel in the three-dimensional data is calculated to determine the gradient direction histogram of each oil and water droplets.Through comparison,the shape of the oil and water is counted.By virtue of the experiment,it is found that the shape matching algorithm based on gradient direction histogram on matching algorithms in terms accuracy and speed is superior to other algorithms,but there are a lot of room for improvement.