RU2760105C1 - System, machine-readable media and method for core analysis by images - Google Patents

Abstract

FIELD: computer technology.SUBSTANCE: invention relates to a method for computer processing and description of the core image. In the method, a core image is obtained in daylight (DL) and ultraviolet (UV) light; solid segments are determined on the core image in daylight (DL) using the watershed method based on the specified segmentation settings; a vector of rock category features is formed for each previously defined image segment; the formed vectors of rock category features are transferred to a multiclass classifier of rocks (MC1), made with the ability to determine at least the following rocks: sandstone, siltstone, mudstone; the areas of luminescence on UV images of the core are determined; vectors of signs of oil saturation and carbonate content are formed for each specific area of luminescence; the formed vectors of signs of oil saturation and carbonate content are transferred to a multiclass classifier (MC2), made with the ability to determine oil saturation and carbonate content; the area of destruction of the DL image is determined and analytical conclusions about the degree of destruction are formed; a description of the core images is formed based on the results of determining the category of rock segments, determining the oil saturation and carbonate content of the glow areas in UV and analytical conclusions about the degree of destruction.EFFECT: increased accuracy of the core description from the images.8 cl, 5 dwg

Description

FIELD OF TECHNOLOGY

The technical solution relates to the field of computer technology, namely to methods and systems for processing core data, core images.

LEVEL OF TECHNOLOGY

A key way to obtain information about the characteristics of the underlying rocks and the presence of recoverable hydrocarbons in them is core drilling of exploration wells, followed by analysis of the core extracted from them. The core is a cylindrical sample several decimeters in diameter and several meters in length. Usually it consists of 3-5 broad layers of rock, alternating with narrow interlayers (layers) of inclusions. Recovered from a hydrocarbon reservoir, the core can include various oil-bearing sedimentary rocks, such as unconsolidated sand, limestone, dolomite, sandstone, siltstone, and clay layers that form the natural reservoir boundaries.

The core must be removed and transported intact so that it retains its original physical and mechanical properties: first of all, porosity (through and total) and associated gas and hydraulic permeability (they determine the fundamental possibility of oil production from the reservoir). Another important aspect is the exact indication of the initial orientation of the core (top / bottom), its binding to a specific borehole and depth, without which further research becomes meaningless. After extraction, marking and initial inspection, the extracted rock sample is delivered to the core storage for detailed laboratory tests. There, specialists from contracting service organizations (usually industry research institutes), invited in advance, first carry out qualitative lithological studies at the macroscopic scale, and then quantitative petrographic measurements at the micro level.

First of all, the core is cut into several parts, polished and photographed entirely in daylight and ultraviolet (UV) light. The latter is done for visual identification of oil: in daylight, oil inclusions (especially for samples with large pores / cracks and high saturation) are visible as dark oily spots on the surface; Under UV light, oil usually photoluminesces blue-white and sometimes brown-yellow, depending on the grade. Even if oil is not immediately detected in a sample, facies (layers with common characteristics) are still identified for it, the degree of fragmentation of the reservoirs is determined, and the lithotypes of the rocks composing the core are identified. This information is subsequently used to build a field model.

At the second stage, thin sections are prepared from the core zones of interest and point petrographic measurements are carried out using an optical / electron microscope. A number of chemical tests are also performed. In particular, they study:

1) Morphology and grain size, relationships between structural elements of the facies;

2) Chemical composition of rocks;

3) Filtration and reservoir properties of formations (porosity, permeability).

Usually, ~ 40 rock samples, each ~ 10 meters long, are removed from the well for analysis. Manual analysis of 1 meter of core takes ~ 15 minutes, which is 100 man-hours or 12.5 workdays for 400 meters of material extracted from just one well. Usually, several people are engaged in the analysis of the brought breed at the same time, which, taking into account the logistics, as well as the time for photography, takes almost one week in total.

The main problem for obtaining a quick and accurate report on the lithological properties of a sample is their rather general descriptive format, usually given in the form of a textual expert opinion on the obtained core photographs. This description turns out to be largely subjective, since it involves manual marking of facies boundaries and identification of rocks "by eye" by their color and texture. Because of this, the same layer may have a slightly different lithotype for different specialists. Another limiting factor is the limited ability of the human eye to distinguish close shades, which is why sometimes inclusions of similar color are missed. It should also be noted here that not only oil itself can luminesce in UV light, but also carbonates in the reservoir rock, whose color can range from orange-brown to, in rare cases, violet, which introduces additional difficulties in the implementation of marking. Finally, terminological discrepancies are possible between the opinions of different experts, since core analysis is usually performed by external contractors who can use their own designations.

The problems listed above lead to a very long and not always sufficiently accurate lithological analysis of the core.

The prior art discloses the publication of the PRC patent application CN110119753 A "Method for recognizing lithology through reconstructed textures", CHANGJIANG GEOTECHNICAL ENG CORPORATION; CHANGJIANG INST SURVEY PLANNING DESIGN & RES CO LTD, which discloses the technical solution used to determine the lithological structure of rocks by highlighting the characteristic features in the image of the rock sample. The technical solution allows you to determine the category of the breed by some characteristic features.

The prior art discloses the publication of US patent application US 20170286802A1, "Automated core description", Saudi Arabian Oil Co, which discloses a solution for the automatic description of a core sample using color, grain, orientation and texture of the samples.

The described technical solutions do not allow to fully automate the process of marking / describing the core from the image, they have a lower accuracy.

ESSENCE

The technical result is an increase in the efficiency and accuracy of the description of the core from images, a decrease in errors in the description of the core, an increase in the degree of automation of the process of description of the core image while maintaining high accuracy of the results and reducing the influence of the human factor on the analysis results.

The method of computer processing and description of the core image includes the following steps: obtaining an image of the core in daylight (DS) and ultraviolet (UV) light; determining whole segments in the daylight image of the core (DS) using the watershed method based on the specified segmentation settings; form a vector of characteristics of the category of rocks for each previously defined segment of the image; the generated vectors of signs of the category of rocks are transmitted to the multiclass classifier of rocks (MK1), made with the ability to determine at least the following rocks - sandstone, siltstone, mudstone; determine the areas of luminescence on the UV images of the core; form vectors of signs of oil saturation and carbonate content for each specific area of the glow; the generated vectors of oil saturation and carbonate indicators are transmitted to a multiclass classifier (MK2), configured to determine oil saturation and carbonate content; determine the area of destruction of the image of the DS and form analytical conclusions about the degree of destruction; form a description of the images of the core based on the results of determining the category of rocks of the segments, determining the oil saturation and carbonate content of the glow areas in the UV and analytical conclusions about the degree of destruction.

In some embodiments, the segmentation settings include at least: the standard deviation for the Gaussian kernel, the value of the structuring element for the morphology, the trimming factor of the side regions in the core images, the factor value for the binarization of the image.

In some embodiments, the segmentation removes noise from the image using a Gaussian filter, uses the Sobel operator and the Otsu method to isolate cracks and boundary transitions between interlayers, and label foreground objects using morphology methods.

In some implementations, the edges of the image are further cropped.

In some implementations, a (multiclass) classifier (MK1 and / or MK2) is trained using Random Forest or Extra Trees or Decision Tree methods.

In some embodiments, vectors (for MK1 and MK2) include at least the following features: lbp feature [1..24]; moments HU [1..7]; minimum intensity of image pixels in a segment; average intensity of image pixels in a segment; maximum intensity of image pixels in a segment; contrast; difference; homogeneity; energy; correlation; quadratic energy (ASM).

In some implementations, some of the method steps are performed using / using a GPU.

In some implementations, the technical solution can be implemented in the form of a system that includes at least one processor (CPU and / or GPU), RAM and is capable of performing / processing / executing steps / steps of the previously described method of computer processing and description of a core image.

In some embodiments, the technical solution may be implemented in the form of a computer-readable medium containing computer-readable instructions for executing by a processor or other calculator / processor a method for computer processing and description of a core image.

In some embodiments, a computer processing and description system for a core image includes at least one processor, random access memory, and computer-readable instructions executed by said processors using random-access memory, the computer-readable instructions comprising the steps of a computer processing and description of a core image.

In some embodiments, the computer-readable medium comprises computer processing method instructions and descriptions of a core image.

BRIEF DESCRIPTION OF THE FIGURES

FIG. 1 shows the distribution of the significance of the features used in the classifiers.

FIG. 2 shows an illustrative example of a block diagram of a method for computer processing and description of a core image.

FIG. 3a shows an illustrative example of a sequence of actions for enriching layer data with the Destruction Degree attribute / characteristic.

FIG. 3b shows an illustrative example of a verbal description of the degree of destruction of the layer with indication of intervals of destruction.

FIG. 4 shows an illustrative example of a system for implementing the described technical solution.

DETAILED DESCRIPTION

Used in the present description of the present technical solution, the terms "module", "component", "element" and the like are used to refer to computer entities that can be hardware / equipment (for example, a device, tool, apparatus, hardware, part of a device, for example , processor, microprocessor, integrated circuit, printed circuit board, including electronic printed circuit board, development board, motherboard, etc., microcomputer and so on), software (for example, executable program code, compiled application, software module, part of software or program code, and so on) and / or firmware (in particular, firmware). So, for example, a component can be a process running on a processor (processor), an object, executable code, program code, file, program / application, function, method, (software) library, subroutine, coroutine, and / or computing device (for example, microcomputer or computer) or a combination of software or hardware components. In some implementations, a neural network module can be implemented in software and / or hardware in the form of a SoC containing a microprocessor and / or a neural processor (Neural Processing Unit, NPU or AI accelerator), which is a specialized class of microprocessors and coprocessors ( often an application specific integrated circuit), used for hardware acceleration of algorithms for artificial neural networks, computer vision, voice recognition, machine learning, and other artificial intelligence methods.

Some of the terms will be defined below.

Moments of the image (eng. Image moments) in computer vision, image processing and related areas - some particular weighted average (moment) of the intensities of the pixels of the image, or a function of such moments.

The method of computer processing and description of the core image includes the following steps:

get an image of the core in daylight (DS) and ultraviolet (UV) light, the depth of sampling, the beginning and end of the images relative to the interval of coring in meters;

In some embodiments, additional information is obtained about the well from which the coring was performed.

In some embodiments, the data (from the laboratory) (images / photos of the core in DS and UV) comes in the form of a set of graphic files (for example, in the -JPG format). In some implementations, a set of files is associated with a specific well and project (the user who uploads the photos understands which project and well the received photos belong to and selects this project and the well in the user interface before loading the data), as well as with the coring interval (slotting interval ).

In some embodiments, each file (or portion of the acquired data) contains an image (photograph) of one meter of core (or less, the remainder).

In some embodiments, together with the image of the core in daylight (DS) and ultraviolet (UV) light, the size of the roof is obtained (received, transmitted) in meters and cm, the size of the foot in meters and centimeters is applied to each image.

In some implementations, the data can be indicated in the name of the corresponding core image file, for example, "3203_40_3204_40_DAL.jpg" - the core sample in this photo contains information from 3203.40 to 3204.40 meters of depth in daylight, "1900 00 1900 90 DAL.jpg" - sample the core in this photo contains information from 1900.00 to 1900.90 centimeters of depth in daylight (not a full meter).

Further, when receiving (uploading a package) images / photographs for one well and one sampling interval, the data (for example, from the name) will be automatically interpreted as the depth of this sample and will then be placed on the ruler (virtual ruler) of the depth in the appropriate place by meter, in reference to the sampling interval, well and project.

In some embodiments, sampling depths are automatically tied to a metering interval.

In some implementations, core images are obtained from external data sources such as databases, archives, or directly from connected equipment.

determining whole segments in the daylight image of the core (DS) using the watershed method based on the specified segmentation settings;

In some implementations, the portion between the start and end of the images relative to the coring interval in meters is used to define the segments.

Before using the WaterShed method, follow these steps:

1. Blur the image with a Gaussian filter;

2. define the boundaries on a blurred image using the Sobel filter;

3. determine the binarization threshold using the Otsu (Otsu) method;

4. form a mask using the threshold binarization method;

5. closure of the binary image (mask) with a structural element is performed.

6. on the binarized image (mask), remove small internal "holes";

7. remove the side areas of the mask that will not be segmented.

8. form a matrix of distances by determining the distance from non-zero (ie, non-background) points to the nearest zero (ie, background) point on the blurred image;

9. on the distance matrix, determine the points that are most distant from the background points;

Then the watershed method is fed a blurred image after processing with the Sobel filter (an image with boundaries), the points found in step 9 are given as the starting points from which the method (markers) began to work, and as a mask, that is, the area in which segmentation will occur (the rest of the areas will be ignored) feed the mask obtained in step 7.

In some embodiments, a disk is used as a structural element, and the parameters of the disk (diameter) are set by the user or are determined automatically.

Segmentation parameters:

1. the standard deviation for the Gaussian kernel (the larger, the stronger the smoothing);

2. the size of the structuring element (disk) for morphology (the higher the value, the more “thick” boundaries / cracks are highlighted to define solid segments);

3. window size for the Otsu (Otsu) method;

4. the value of the coefficient (from 0% to 50%) for trimming the side regions in the images of the core.

5. The value of the coefficient for the Otsu method for binarization of the image (the larger, the less detailed the image - fewer segments). In some implementations, the value of the coefficient is determined using the Otsu method.

In the case of DS - the destroyed areas do not fall into the mask, so that the degree / area of destruction will be determined / calculated from them in the future. In the case of UV (will be described later), the black areas without glow do not fall into the mask.

In some implementations, the parameters can be set by the user.

In some implementations, parameters may be predefined (eg, in configuration files or otherwise).

form a vector of features of the category of rocks for each previously defined segment of the core DS image;

For the task of determining the breed, at least the following features were used based on: adjacency matrix, local binary patterns, global image description.

Adjacency matrix - a matrix of frequencies of pairs of pixels of a certain brightness located on the image in a certain way relative to each other (https://en.wikipedia.org/wiki/Co-occurrence_matrix).

Based on this matrix, additional characteristics / features are determined that describe the texture of the image: contrast, dissimilarity, homogeneity, energy, correlation, quadratic energy (eng. ASM).

Local binary patterns represent a description of the neighborhood of an image pixel in binary form. The LBP operator, applied to a pixel in an image, uses eight pixels of the neighborhood, taking the center pixel as a threshold. Pixels that have a value greater than (or equal to) the central pixel take on the values 1, and those that are less than the central one take on the values 0. Thus, an eight-bit binary code is obtained that describes the neighborhood of the pixel. Then the histogram of the distribution of the obtained eight-bit codes is determined / calculated / formed and used as a feature vector. For extreme pixels, the histogram is not detected / generated. They are taken into account when working with neighboring pixels, but do not become centers themselves.

24 (twenty four) lbp [1..24] features are formed based on local binary patterns, since a neighborhood of radius 3 (three) is considered around the pixel and 24 points are considered along the perimeter of this circle, which characterize the area. Then this is reduced to a histogram (distribution of repetitions) with the number of divisions 24, thus the object is described by 24 features (the number of certain local binary patterns found).

Global image description

For image segments, the maximum, average and minimum intensities in the region were calculated, as well as the moments of pixel intensities (MOMENTS HU) of the image.

Features are formed based on the moments of the images.

In some implementations, MOMENTS HU (Rotation invariants, https://en.wikipedia.org/wiki/Image_moment) are used as image moments.

In some embodiments, features [1..7] are generated using the above Moments HU (Rotation invariants).

The presented features are determined / calculated for all segments obtained after image processing and segmentation by the watershed method.

In some embodiments, a breed category vector is generated that includes at least the following features (or a subset of them):

• lbp feature 1… lbp feature 24 (local binary patterns);

• moments HU 1… moments HU 7;

• minimum intensity of image pixels in the area (segment);

• average intensity of image pixels in the area (segment);

• maximum intensity of image pixels in the area (segment);

• contrast (contrast);

• difference (dissimilarity);

• homogeneity;

• energy (energy);

• correlation;

• Quadratic Energy (ASM).

In some embodiments, the breed category vector is generated for each segment (the features used in the vector are defined for each segment separately).

the generated vectors of signs of the category of rocks are transmitted to a multiclass classifier of rocks, made with the ability to determine at least the following rocks - sandstone, siltstone, mudstone;

In some implementations, a previously trained classifier, such as RandomForestClassifier or ExtraTreesClassifier or DecisionTreeClassifier, is used as the classifier.

In some implementations, StratifiedKFold cross-validation is used to evaluate the classifier.

When using machine learning based models / classifiers, there is a chance that the model / classifier will ignore the minority classes. In some implementations, the training sample data is balanced before training.

In some implementations, balancing is performed taking into account the different geographic locations of the wells from which the sample is taken. One of the difficulties in processing the sample was that the data were obtained during surveying with different equipment and under different conditions; it was necessary to put into the model the ability to predict the type of rock from images, taking into account different shooting conditions presented in the sample.

In some implementations, balancing is done as follows:

1) images of DS with the type of rock are selected from among the necessary ones (sandstone, mudstone, siltstone);

2) discard images of small size (for example, where the total image area is less than 1 million pixels, but not limited to);

3) determine / calculate the histogram of the distribution of the selected images by fields and wells;

4) determine the maximum number of examples for each class and select the minimum, and for classes with a large number of examples, mixing is carried out;

5) evenly select samples from each well until a sample of the required size is formed.

Thus, in the training sample, there were necessarily samples from various wells / fields.

In some embodiments, each segment is associated with a sampling depth and / or distance from the start of the coring interval. In some embodiments, the segment size is determined by recalculating the sampling interval into pixels (taking into account the image resolution) and calculating the size (height, width) of the segment in relation to the real dimensions of the core.

At the output (after the transmission of the feature vector) (multiclass) the classifier determines the breed category and this result will be used further.

determine the areas of luminescence on the UV images of the core;

To determine the glow areas, use the watershed method and preprocessing described earlier (the step is determined by the whole segments in the image in daylight using the watershed method based on the specified segmentation settings). Thus, only bright (compared to the whole image) areas are detected / obtained.

form vectors of signs of oil saturation and carbonate content for each specific area of the glow;

In some embodiments, a vector of oil saturation and carbonate features is generated, including at least the following features (or a subset of them):

• lbp feature 1… lbp feature 24 (local binary patterns);

• moments HU 1… moments HU 7 (from the first to the seventh moments calculated by the Hu moments method);

• minimum intensity of image pixels in the area (segment);

• average intensity of image pixels in the area (segment);

• maximum intensity of image pixels in the area (segment);

• contrast (contrast);

• difference (dissimilarity);

• homogeneity;

• energy (energy);

• correlation;

• Quadratic Energy (ASM).

In some embodiments, vectors of oil saturation and carbonate indicators are generated for each segment (the features used in the vector are defined for each segment separately).

the generated vectors of oil saturation and carbonate indicators are transmitted to a multiclass classifier adapted to determine the oil saturation and carbonate content;

In some implementations, a previously trained classifier, such as RandomForestClassifier or ExtraTreesClassifier or DecisionTreeClassifier, is used as the classifier.

In some implementations, StratifiedKFold cross-validation is used to evaluate the classifier.

When using machine learning based models / classifiers, there is a chance that the model / classifier will ignore the minority classes. In some implementations, the training sample data is balanced before training.

At the output (after transmission of the feature vector), the classifier determines the oil saturation and / or carbonate content of the segment.

determine the area of destruction using a pixel-by-pixel calculation of the DS image and form analytical conclusions about the degree of destruction

The degree of destruction is determined using the results of the step of forming the mask using the method of binarization by the threshold for the image of the DS, since in the case of DS, the destroyed areas do not fall into the mask. Thus, all other areas that did not fall into the mask are areas of destruction / absence of core. Next, the pixel-by-pixel area of these areas is determined and the ratio of the areas (the total area of the image to the sum of the areas of all destroyed areas) is determined.

In some embodiments, the degree of destruction can be one of the following: destroyed, partially destroyed, not destroyed.

In some embodiments, fracture analysis determines the percentage representation of the ratio of areas (total image area to the sum of areas of destroyed areas).

We consider the areas "allowed", determine their per-pixel area and determine the ratio of the areas (the total area of the image to the sum of the areas of all destroyed areas).

form a description of the images of the core based on the results of determining the category of rocks of the segments, determining the oil saturation and carbonate content of the glow areas in the UV and analytical conclusions about the degree of destruction.

To form the description, each obtained image segment is depicted (taking into account the binding of the DS and UV segments to the sampling depth) in UV light is taken as the layer described below, and for each layer the description is formed as follows:

1.The parameters of oil saturation and carbonate content are mutually exclusive, i.e. if the layer is oil-saturated, then it is not carbonate;

2. oil saturation or carbonate content is the primary criterion for identifying a separate layer, then the type of rock is tied to this layer; If several segments with rocks in DS light fall into one segment (layer) in the UV light, then all rock types falling into it are attached to the layer.

3.the parameter of destruction has a minimum weight, therefore this characteristic is tied to the final layer, selected on the basis of items 1 and 2

4. If the layer with the parameters of destruction completely coincides with the resulting layer - the degree of destruction is transferred as it is, 0 - 0, 1 - 1, 2 - 2.

5. If within the resulting layer there are several layers with different degrees of destruction (for example, 0 - not destroyed and 1 - partially destroyed or any other combinations) - set the resulting layer to 1 - partially destroyed.

The data of the layer is enriched with the attribute / characteristic "Degree of destruction" (Fig. 3a):

• if the layer with the parameters of destruction completely coincides with the resulting layer - the degree of destruction is set (transferred) as it is, 0 - 0, 1 - 1, 2-2;

• if within the resulting layer there are several layers with different degrees of destruction (for example, 0 - not destroyed and 1 - partially destroyed or any other combinations) - the resulting layer is set to 1 - partially destroyed.

Form a verbal description of the degree of destruction of the layer indicating the intervals of destruction (Fig. 3b).

In some implementations, a method for computer processing and description of a core image is performed by a server and invoked through an API (eg, REST).

In some embodiments, based on the generated description of the core, an optimal research program is formed, samples are marked for subsequent production, and numerous problems that geological services face are solved.

In some embodiments, according to the presence and depth of oil-saturated and dense intervals (the nature of the glow that evaluates the method), the depths of the test intervals and the prospects of the target interval as a whole are promptly specified.

FIG. 4 shows an example of a general-purpose computer system used to implement the described method - a personal computer or server 20 containing a central processor 21, system memory 22 and a system bus 23, which contains various system components, including memory associated with the central processor 21. System bus 23 is implemented as any bus structure known in the art, which in turn contains a bus memory or bus memory controller, a peripheral bus and a local bus that is capable of interfacing with any other bus architecture. System memory contains read-only memory (ROM) 24, random access memory (RAM) 25. The main input / output system (BIOS) 26 contains basic procedures that transfer information between the elements of the personal computer 20, for example, at the time of loading the operating room systems using ROM 24.

The personal computer 20, in turn, contains a hard disk 27 for reading and writing data, a magnetic disk drive 28 for reading and writing to removable magnetic disks 29 and an optical drive 30 for reading and writing to removable optical disks 31, such as CD-ROM, DVD -ROM and other optical media. The hard disk 27, the magnetic disk drive 28, and the optical drive 30 are connected to the system bus 23 via the hard disk interface 32, the magnetic disk interface 33, and the optical drive interface 34, respectively. Drives and corresponding computer storage media are non-volatile storage media for computer instructions, data structures, program modules and other data of a personal computer 20.

The present description discloses an implementation of a system that uses a hard disk 27, but it should be understood that it is possible to use other types of computer storage media that are capable of storing data in a computer readable form (solid state drives, flash memory cards, digital disks, random access memory (RAM), etc.), which are connected to the system bus 23.

Computer 20 has a file system 36, where the recorded operating system 35 is stored, as well as additional software applications 37, other program modules 38 and program data 39. The user has the ability to enter commands and information into the personal computer 20 through input devices (keyboard 40, manipulator " mouse "42). Other input devices may be used (not shown): microphone, joystick, game console, scanner, etc. Such input devices, as usual, are connected to the computer system 20 via the USB interface 46, which in turn is connected to the system bus, but can be connected in another way, for example, using a parallel port, a game port. A monitor 47 or other type of display device is also connected to the system bus 23 via an interface such as a video adapter 48. In addition to the monitor 47, the personal computer may be equipped with other peripheral output devices (not shown).

The personal computer 20 is capable of operating in a networked environment using a network connection with other or more remote computers 49. The remote computer (or computers) 49 are the same personal computers or servers that have most or all of the elements mentioned earlier in the description of the entity. the personal computer 20 shown in FIG. 4. In a computer network, there may also be other devices, such as routers, network stations, peer-to-peer devices, or other network nodes.

Network connections can form a local area network (LAN) 50 and a wide area network (WAN). Such networks are used in corporate computer networks, internal networks of companies and, as a rule, have access to the Internet. In LAN or WAN networks, the personal computer 20 is connected to the local network 50 via a network adapter or network interface 51. When using networks, the personal computer 20 may use a router 54 or other means of providing communication with a wide area network, such as the Internet. Router 54, which is an internal or external device, is connected to the system bus 23 via a USB port 46. It should be noted that the network connections are only approximate and are not required to reflect the exact configuration of the network, i. E. in fact, there are other ways of establishing a connection by technical means of communication of one computer with another.

Claims (17)

1. The method of computer processing and description of the core image includes the following steps: - get an image of the core in daylight (DS) and ultraviolet (UV) light; - define whole segments on the core image in daylight (DS) using the watershed method based on the specified segmentation settings; - form a vector of characteristics of the category of rocks for each previously defined segment of the image; - the generated vectors of attributes of the category of rocks are transmitted to the multiclass classifier of rocks (MK1), made with the ability to determine at least the following rocks - sandstone, siltstone, mudstone; - determine the areas of luminescence on the UV images of the core; - generates vectors of oil saturation and carbonate indicators for each specific glow area; - the generated vectors of oil saturation and carbonate indicators are transmitted to the multiclass classifier (MK2), made with the ability to determine the oil saturation and carbonate content; - determine the area of destruction of the image of the DS and form analytical conclusions about the degree of destruction; - form a description of the images of the core based on the results of determining the category of rocks in the segments, determining the oil saturation and carbonate content of the glow areas in the UV and analytical conclusions about the degree of destruction. 2. The method according to claim 1, in which the segmentation settings include at least: standard deviation for the Gaussian kernel, the size of the structuring element for the morphology, the trimming factor of the side regions in the core images, the factor value for the binarization of the image. 3. The method according to claim 1, in which the segmentation removes noise from the image using a Gaussian filter, the Sobel operator and the Otsu method are used to highlight cracks and boundary transitions between interlayers, and foreground objects are marked using morphological methods. 4. The method of claim 3, further comprising cropping the edges of the image. 5. The method of claim 1, wherein the multiclass classifier is trained using Random Forest or Extra Trees or Decision Tree methods. 6. The method according to claim 1, wherein the vectors include at least the following features: lbp feature [1..24]; moments HU [1..7]; minimum intensity of image pixels in a segment; average intensity of image pixels in a segment; maximum intensity of image pixels in a segment; contrast; difference; homogeneity; energy; correlation; quadratic energy (ASM). 7. A system for computer processing and description of a core image, including at least one processor, random access memory and computer-readable instructions executed by said processors using random access memory, the computer-readable instructions comprising the steps of the method for computer processing and description of a core image according to claim 1. 8. Computer-readable medium containing instructions for the computer processing method and description of the core image according to claim 1.

Images