Lithology identification method and device based on fusion of rock debris fluorescence image and raman spectrum and electronic equipment

Abstract

The present application relates to a kind of lithology identification technical field, it is a kind of lithology identification method, device and electronic equipment based on cutting fluorescence image and Raman spectrum fusion, including corresponding cutting sample is non-petroleum hydrocarbon fluorescence cutting fluorescence image, application Gaussian weighting method dynamic segmentation fluorescence area, obtain high, medium, low fluorescence intensity area image and no fluorescence area image, and determine the gray value range of each area image and fluorescence area proportion, and bring into mineral type discrimination rule and match, obtain corresponding mineral species and mineral content;Again with lithology identification rule and match, obtain corresponding lithology identification result.The present application can be without any experimental instrument in the lithology identification, directly through the gray value range of reaction fluorescence intensity, combined with the mapping relationship of established fluorescence intensity and mineral composition directly obtain mineral species and mineral content, both can effectively save drilling cost, and quickly, accurately identify lithology.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of lithology identification, and is a lithology identification method and device based on fusion of a cutting fluorescence image and a Raman spectrum, and an electronic device. BACKGROUND

[0002] Lithology identification is a key link in oil and gas resource exploration and development, but the wide application of PDC drill bits poses a severe challenge to cutting identification in geological logging. Specifically, the PDC drill bit breaks rocks through shearing and cutting, resulting in extremely small cutting particle sizes. In particular, in poor lithology formation, the cutting is mostly in a paste or dispersed granular state, and the mineral composition and structure are difficult to identify with the naked eye, which reduces the accuracy of lithology identification.

[0003] With the rapid development of computer technology, a large number of researchers have begun to analyze cutting images using intelligent algorithms such as machine learning to identify lithology, or combine image analysis results with cutting mineral and element data to jointly identify lithology. For example:

[0004] Existing patent document one, CN111709423B, discloses a PDC drill bit condition cutting lithology identification method based on lithology feature library matching, which includes: 1. drilling under the condition of a PDC drill bit, collecting fine cutting particle white light original images and fluorescence images; 2. converting the white light original images and fluorescence images from RGB color space images to HSV color space images; 3. segmenting the fluorescence images using a threshold segmentation formula; 4. segmenting the white light original images using an image segmentation algorithm based on watershed and near neighbor region merging to segment all cutting particles in the entire white light original image; 5. extracting the color and texture features of the cutting particles at the corresponding positions to construct a lithology feature library; 6. calculating the Bhattacharyya similarity distance between the lithology feature library features and the matching cutting features, calculating the area ratio of the characteristic cutting according to the matching results obtained from the similarity distance, and completing the identification of the cutting lithology.

[0005] Existing patent document two, CN119478669A, discloses a hyperspectral image lithology identification method and device based on spatial and spectral features, relating to the technical field of image information processing. The steps of the hyperspectral image lithology identification method based on spatial and spectral features mainly include: preprocessing the hyperspectral data of the study area to obtain a hyperspectral sample data set, dividing the hyperspectral sample data set according to a preset ratio and performing data enhancement processing to obtain a training set, a validation set, and a test set; constructing a lithology identification model using a high-level convolutional neural network model; training and verifying the lithology identification model using the training set and the validation set to obtain a trained lithology identification model; and obtaining a lithology identification result using the trained lithology identification model according to the test set.

[0006] The existing published patent document three, the publication number is CN110031493B, discloses a lithology intelligent identification system and method based on image and spectrum technology, including a rock block shape analysis system for collecting shape information of a test sample and preselecting multiple XRF detection surfaces according to the shape information of the test sample, and determining a test sample grinding position according to the grinding workload of different detection surfaces, and transmitting the test sample grinding position and grinding surface flatness to a central analysis control system; the central analysis control system controls the sample processing system to grind the sample according to the determined test sample grinding position until the flatness requirement of X-ray fluorescence analysis is met, and the ground rock block is preliminarily judged by the image recognition system; after the sample processing system grinds the sample to meet the requirements, the debris generated in the grinding process of the sample is ground, and the image recognition system is used to judge whether the rock powder meets the requirements of X-ray diffraction analysis on the size of the rock particles; the spectrum analysis system performs X-ray diffraction analysis and X-ray fluorescence analysis on the rock powder with the required particle size and the sample with the required flatness respectively, and transmits the respective analysis results to the central analysis control system; the central analysis control system determines the final lithology of the sample according to the rock block identification result transmitted by the image recognition system and the analysis result transmitted by the spectrum analysis system.

[0007] The above method has the following problems:

[0008] (1) The existing method of combining cutting image and mineral data to realize lithology identification mostly depends on XRD and XRF instrument equipment, and the XRD and XRF sample analysis period is long, which does not meet the requirement of on-site rapid analysis of logging, resulting in delayed decision-making.

[0009] (2) Only the lithology can be determined, and the mineral species and content cannot be quantified. SUMMARY

[0010] The present application provides a lithology identification method, device and electronic equipment based on the fusion of cutting fluorescence image and Raman spectrum, which overcomes the shortcomings of the prior art. The present application effectively solves the problem that most existing lithology identification methods rely on XRD, XRF and other instrument equipment, and the XRD and XRF sample analysis period is long, time-consuming and high-cost.

[0011] One of the technical solutions of the present application is realized by the following measures: a lithology identification method based on the fusion of cutting fluorescence image and Raman spectrum, comprising:

[0012] obtaining a cutting fluorescence image, and the fluorescence type of the cutting sample corresponding to the cutting fluorescence image is non-hydrocarbon fluorescence;

[0013] The Gaussian weighting method is applied to dynamically segment the fluorescent region, high, medium and low fluorescent intensity region images and non-fluorescent region images are obtained according to different fluorescent intensity ranges, and the gray value range and fluorescent area proportion of each region image are determined, wherein the fluorescent area proportion is the proportion of the number of fluorescent pixels in the total number of image pixels;

[0014] The gray value range and fluorescent area proportion of each region image are brought into the mineral type discrimination rule for matching to obtain the corresponding mineral species and mineral content, wherein the mineral type discrimination rule is obtained by combining in-situ Raman spectrum analysis to analyze the mineral composition of the rock debris samples in different fluorescent intensity regions of a plurality of historical rock debris fluorescent images, and the mineral content is the sum of the fluorescent area proportions of the region images corresponding to the same mineral species;

[0015] The mineral species and mineral content are brought into the lithology identification rule for matching to obtain the corresponding lithology identification result, wherein the lithology identification rule comprises:

[0016] (1) determining the basic name

[0017] When the highest mineral content is greater than 50%, the corresponding mineral type is taken as the basic name of the rock;

[0018] When all mineral contents are less than 50%, the mineral type with the highest mineral content is taken as the basic name of the rock;

[0019] (2) determining the additional noun

[0020] When the remaining mineral content is between 50% and 25%, the corresponding mineral type is added in the form of XX quality before the basic name;

[0021] When the remaining mineral content is between 25% and 10%, the corresponding mineral type is added in the form of containing XX before the basic name;

[0022] When the remaining mineral content is less than 10%, the corresponding mineral type does not participate in rock naming.

[0023] The following is a further optimization or / and improvement of the above technical solutions of the invention:

[0024] The above application of the Gaussian weighting method to dynamically segment the fluorescent region, the high, medium and low fluorescent intensity region images and the non-fluorescent region images are obtained according to different fluorescent intensity ranges, and the gray value and fluorescent area proportion of each region image are determined, comprising:

[0025] The rock debris fluorescent image is converted into a gray scale image and preprocessed;

[0026] The Gaussian weighting method is applied to dynamically segment the gray scale image to obtain the corresponding fluorescent region, and the fluorescent region is further processed by morphological optimization and connected region extraction in turn;

[0027] Based on the reprocessed fluorescent region, high, medium and low fluorescent intensity region images and no fluorescent region image are obtained according to different fluorescent intensity ranges;

[0028] The gray value and the fluorescent area proportion of each region image are determined, and the calculation formula of the fluorescent area proportion is as follows:

[0029]

[0030] Wherein, is the fluorescent area proportion; is the number of fluorescent pixels of the fluorescent region; is the total number of pixels of the rock debris fluorescent image.

[0031] The establishment of the above mineral type identification rule includes:

[0032] A plurality of historical rock debris fluorescent images are obtained, and the fluorescent type of the rock debris sample corresponding to each historical rock debris fluorescent image is not petroleum hydrocarbon fluorescence;

[0033] The fluorescent region is dynamically segmented by using the Gaussian weighting method, high, medium and low fluorescent intensity region images and no fluorescent region image are obtained according to different fluorescent intensity ranges, and the gray value range and the fluorescent area proportion of each region image are determined, wherein the fluorescent area proportion is the proportion of the number of fluorescent pixels in the total number of pixels.

[0034] At least one rock debris sample is obtained in each region image, and the mineral composition of each rock debris sample is obtained by using in-situ Raman spectrum analysis.

[0035] Based on the gray value range of each region image and the mineral composition of all rock debris samples, the gray value range of each mineral type is determined, and the gray abnormal value of each mineral type is removed by using the 3sigma principle, and finally the gray value range of each mineral type is obtained.

[0036] The above rock debris fluorescent image is obtained, and the fluorescent type of the rock debris sample corresponding to the rock debris fluorescent image is non-petroleum hydrocarbon fluorescence, which includes:

[0037] Any rock debris fluorescent image is obtained, and the three-dimensional fluorescence spectrum analysis result of the rock debris sample corresponding to the rock debris fluorescent image is combined;

[0038] If no fluorescent characteristic region is displayed in the three-dimensional fluorescence spectrum analysis result, lithology identification is performed;

[0039] If the fluorescent characteristic region is displayed in the three-dimensional fluorescence spectrum analysis result, the fluorescence type rejudgment process is triggered;

[0040] obtain a three-dimensional fluorescence spectrum of the rock sample corresponding to the rock debris fluorescence image and a three-dimensional fluorescence spectrum of the drilling fluid, and detect the fluorescence feature regions in the two three-dimensional fluorescence images respectively, and extract the corresponding fluorescence feature region features, wherein the fluorescence feature region features include contour circumscribed rectangle parameters and contour mask region color features;

[0041] perform similarity analysis on the fluorescence feature region features in the two three-dimensional fluorescence images to obtain a comprehensive similarity;

[0042]

[0043] wherein, is the comprehensive similarity; is the region shape similarity, which is obtained by weighting the region number similarity, the position distribution similarity, and the shape similarity; is the color distribution similarity;

[0044] If > 0.8, the fluorescence is caused by drilling fluid additive pollution, and lithology identification is performed, and if < 0.8, the fluorescence type is oil hydrocarbon fluorescence, and lithology identification is not performed.

[0045] The above similarity analysis on the fluorescence feature region features in the two three-dimensional fluorescence images to obtain a comprehensive similarity includes:

[0046] For the three-dimensional fluorescence spectrum of the rock sample and the three-dimensional fluorescence spectrum of the drilling fluid, the hue histograms of the fluorescence feature regions of the two are obtained respectively, the Bhattacharyya distance between the two is determined, and the color distribution similarity is obtained;

[0047]

[0048] wherein, is the color distribution similarity; is the Bhattacharyya distance;

[0049] The region shape similarity is obtained by weighting the region number similarity, the position distribution similarity, and the shape similarity, including:

[0050] determine the region number similarity;

[0051]

[0052] wherein, is the region number similarity; , are the number of fluorescence feature regions in the two three-dimensional fluorescence images respectively;

[0053] determine the position distribution similarity, including:

[0054] (1) Obtain the three-dimensional fluorescence spectrum point set of the rock sample and the three-dimensional fluorescence spectrum point set of the drilling fluid , determine the directed Hausdorff distance between the point sets, wherein the three-dimensional fluorescence spectrum point set is a set of geometric center points of each fluorescence feature region in the three-dimensional fluorescence spectrum;

[0055]

[0056] wherein, is the directed Hausdorff distance from to ; is the directed Hausdorff distance from to ; is one of the geometric center points in ; is one of the geometric center points in ; is the Euclidean distance between and ;

[0057] (2) Take the maximum value as the final distance :

[0058]

[0059] (3) Normalize the distance and convert it to a position distribution similarity;

[0060]

[0061] wherein, is the position distribution similarity; is the maximum value of the size of the two images; is the distance normalized according to the image size;

[0062] Determine the shape similarity, obtain the largest area fluorescence feature region in the three-dimensional fluorescence spectrum of the rock sample and the three-dimensional fluorescence spectrum of the drilling fluid respectively, determine the corresponding Hu moment, calculate the cosine similarity between the two Hu moments , and get the shape similarity after normalization;

[0063] ;

[0064] Get the regional shape similarity based on weighting;

[0065]

[0066] wherein, is a region shape similarity; is a region number similarity; is a position distribution similarity; is a shape similarity.

[0067] The second technical solution of the present application is realized by the following measures: a lithology identification device based on fusion of a cutting fluorescence image and a Raman spectrum, comprising:

[0068] An original image acquisition unit acquires a cutting fluorescence image, and the fluorescence type of a cutting sample corresponding to the cutting fluorescence image is a non-petroleum hydrocarbon fluorescence;

[0069] An image analysis unit applies a Gaussian weighting method to dynamically segment a fluorescence region, obtains high, medium, and low fluorescence intensity region images and a non-fluorescence region image according to different fluorescence intensity ranges, and determines a gray value range and a fluorescence area proportion of each region image, wherein the fluorescence area proportion is a proportion of a fluorescence pixel number in a total pixel number of the image;

[0070] A mineral analysis unit inputs the gray value range and the fluorescence area proportion of each region image into a mineral type discrimination rule for matching to obtain corresponding mineral species and mineral content, wherein the mineral type discrimination rule is obtained by combining in-situ Raman spectrum analysis to analyze mineral compositions of cutting samples in different fluorescence intensity regions of a plurality of historical cutting fluorescence images, and the mineral content is a sum of the fluorescence area proportions of the region images corresponding to the same mineral species;

[0071] A lithology identification unit inputs the mineral species and the mineral content into a lithology identification rule for matching to obtain a corresponding lithology identification result, wherein the lithology identification rule comprises:

[0072] (1) determining a basic name

[0073] When the highest mineral content is greater than 50%, the corresponding mineral type is taken as the basic name of the rock;

[0074] When all the mineral contents are less than 50%, the mineral type with the highest mineral content is taken as the basic name of the rock;

[0075] (2) determining an additional noun

[0076] When the remaining mineral contents are between 50% and 25%, the corresponding mineral type is added in the form of XX quality before the basic name;

[0077] When the remaining mineral contents are between 25% and 10%, the corresponding mineral type is added in the form of containing XX before the basic name;

[0078] If the remaining mineral content is less than 10%, the corresponding mineral type does not participate in the rock naming.

[0079] The following is a further optimization or / and improvement of the above technical solutions:

[0080] The above image analysis unit comprises:

[0081] The pre-processing module converts the cutting fluorescence image into a gray-scale image and performs pre-processing;

[0082] The fluorescence region extraction module applies the Gaussian weighting method to the gray-scale image for dynamic segmentation to obtain the corresponding fluorescence region, and sequentially performs fluorescence region reprocessing through morphological optimization and connected region extraction;

[0083] The fluorescence region reprocessing module, based on the reprocessed fluorescence region, obtains high, medium and low fluorescence intensity region images and no fluorescence region image according to different fluorescence intensity ranges;

[0084] The fluorescence region quantification module determines the gray-scale value and fluorescence area proportion of each region image, wherein the calculation formula of the fluorescence area proportion is as follows:

[0085]

[0086] Wherein, is the fluorescence area proportion; is the number of fluorescence pixels in the fluorescence region; is the total number of pixels of the cutting fluorescence image.

[0087] The above original image acquisition unit comprises:

[0088] The fluorescence type acquisition module acquires any cutting fluorescence image and combines the three-dimensional fluorescence spectrum analysis result of the cutting sample corresponding to the cutting fluorescence image;

[0089] The first fluorescence type analysis module performs lithology identification if there is no fluorescence characteristic region displayed in the three-dimensional fluorescence spectrum analysis result;

[0090] The second fluorescence type analysis module triggers the fluorescence type rejudgment process if there is a fluorescence characteristic region displayed in the three-dimensional fluorescence spectrum analysis result, including:

[0091] Obtain the three-dimensional fluorescence spectrum of the cutting sample corresponding to the cutting fluorescence image and the three-dimensional fluorescence spectrum of the corresponding drilling fluid, and detect the fluorescence characteristic regions in the two three-dimensional fluorescence images respectively, and extract the corresponding fluorescence characteristic region features, wherein the fluorescence characteristic region features include the contour circumscribed rectangle parameters and the contour mask region color features;

[0092] The similarity of the fluorescence characteristic region features in the two three-dimensional fluorescence images is analyzed to obtain the comprehensive similarity;

[0093]

[0094] wherein, is the comprehensive similarity; is the regional shape similarity, obtained by weighting the regional number similarity, position distribution similarity and shape similarity; is the color distribution similarity;

[0095] If > 0.8, the fluorescence is caused by drilling fluid additive pollution, and lithology identification is performed, and if < 0.8, the fluorescence type is petroleum hydrocarbon fluorescence, and lithology identification is not performed.

[0096] The third technical solution of the present application is realized by the following measures: an electronic device comprising a processor and a memory, the memory storing a computer program, the computer program being loaded and executed by the processor to realize the steps in the lithology identification method based on the fusion of the fluorescence image of the rock debris and the Raman spectrum.

[0097] The beneficial effects of the present application include:

[0098] The present application combines in-situ Raman spectrum analysis to analyze the mineral composition of the rock debris samples in different fluorescence intensity regions in a plurality of historical rock debris fluorescence images, establishes a mapping relationship between the fluorescence intensity and the mineral composition, so that in the lithology identification, without the participation of any experimental instrument, the mineral type and the mineral content can be directly obtained by combining the established mapping relationship between the fluorescence intensity and the mineral composition and the gray value range of the reaction fluorescence intensity, which not only effectively saves the drilling cost, but also quickly and accurately identifies the lithology, and further, in the present embodiment, no conventional XRD, XRF and other instrument equipment are used for sample analysis, avoiding the problems of long sample analysis period, not meeting the on-site rapid analysis demand of logging, and leading to decision lag;

[0099] When the fluorescence region characteristic display is shown in the three-dimensional fluorescence spectrum analysis result of the rock debris sample corresponding to the rock debris fluorescence image, the present application comprehensively evaluates the comprehensive similarity between the two three-dimensional fluorescence images based on the regional shape similarity and the color distribution similarity of the fluorescence characteristic region in the three-dimensional fluorescence spectrum of the rock debris and the drilling fluid, and determines whether the display is caused by drilling fluid additive pollution or petroleum hydrocarbon fluorescence, compared with the method of judging whether the fluorescence type is mineral fluorescence by a single factor, the judgment result is more accurate, and a reliable guarantee is provided for the lithology identification method. BRIEF DESCRIPTION OF DRAWINGS

[0100] The Figure 1 is the lithology identification method flowchart provided by the present embodiment 1.

[0101] The Figure 2The schematic diagram of the acquisition process of the fluorescence image of the cuttings is provided for the embodiment 2 of the present application.

[0102] The schematic diagram of the acquisition process of the fluorescence image of the cuttings is provided for the embodiment 2 of the present application. Figure 3 The schematic diagram of the acquisition process of the fluorescence image of the cuttings is provided for the embodiment 2 of the present application.

[0103] The schematic diagram of the acquisition process of the fluorescence image of the cuttings is provided for the embodiment 2 of the present application. Figure 4 The schematic diagram of the acquisition process of the fluorescence image of the cuttings is provided for the embodiment 2 of the present application.

[0104] The schematic diagram of the acquisition process of the fluorescence image of the cuttings is provided for the embodiment 2 of the present application. Figure 5 The schematic diagram of the acquisition process of the fluorescence image of the cuttings is provided for the embodiment 2 of the present application. Figure 5 a is the three-dimensional fluorescence spectrum of the cuttings, and Figure 5 b is the three-dimensional fluorescence spectrum of the cuttings, and Figure 5 c is the three-dimensional fluorescence spectrum of the drilling fluid, and Figure 5 d is the three-dimensional fluorescence spectrum of the drilling fluid, and Figure 5 e is the contrast diagram of the characteristic region contour.

[0105] The schematic diagram of the acquisition process of the fluorescence image of the cuttings is provided for the embodiment 2 of the present application. Figure 6 The schematic diagram of the acquisition process of the fluorescence image of the cuttings is provided for the embodiment 2 of the present application. Figure 6 a is the fluorescence image of the cuttings, Figure 6 b is the preprocessed gray-scale image, Figure 6 c is the high-fluorescence-intensity region image, Figure 6 d is the medium-fluorescence-intensity region image, Figure 6 e is the low-fluorescence-intensity region image, Figure 6 f is the non-fluorescence region image.

[0106] The schematic diagram of the acquisition process of the fluorescence image of the cuttings is provided for the embodiment 2 of the present application. Figure 7 The characteristic Raman spectrum of the minerals in the cuttings is provided for the embodiment 5 of the present application.

[0107] The characteristic Raman spectrum of the minerals in the cuttings is provided for the embodiment 5 of the present application. Figure 8 The confusion matrix display diagram is provided for the embodiment 5 of the present application.

[0108] The confusion matrix display diagram is provided for the embodiment 5 of the present application. Figure 9 The schematic diagram of the structure of the lithology identification device is provided for the embodiment 6 of the present application. DETAILED DESCRIPTION

[0109] The present application is not limited by the following embodiments, and the specific implementation manners can be determined according to the technical solutions of the present application and the actual situation.

[0110] Those skilled in the art can understand that, unless specifically stated, the "module" or "unit" in the embodiments of the present application refers to a computer program or a part of a computer program with a predetermined function, and works together with other related parts to achieve a predetermined target, and can be implemented entirely or partially by using software, hardware (such as a processing circuit or a memory) or a combination thereof. Similarly, one processor (or multiple processors or memories) can be used to implement one or more modules or units. In addition, each module or unit can be a part of an integral module or unit that includes the functions of the module or unit.

[0111] In addition, "a plurality of" in the embodiments of the present application refers to two or more, and "first" and "second" and the like are used for differentiation description and cannot be understood as implying relative importance.

[0112] The embodiments of the present application provide a lithology identification method and device based on fusion of rock debris fluorescence images and Raman spectra and electronic equipment. The lithology identification device based on fusion of rock debris fluorescence images and Raman spectra can be integrated in a computer device, which can be a server, a terminal, or the like. It can also be executed by a terminal and a server together, and the above examples should not be understood as limiting the present application.

[0113] The terminal described above can include a mobile phone, a wearable smart device, a tablet computer, a notebook computer, a personal computer (PC), and a vehicle-mounted computer, and the present application does not limit this. The present application does not limit the number of terminal devices.

[0114] The server described above can be a standalone physical server, a server cluster or a distributed system composed of multiple physical servers, a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDNs, and basic cloud computing services such as big data and artificial intelligence platforms, and the present application does not limit this.

[0115] For example, a computer device acquires a fluorescence image of rock fragments, and the fluorescence type of the rock fragment sample corresponding to this image is non-petroleum hydrocarbon fluorescence. The fluorescence region is dynamically segmented using a Gaussian weighted method, and images of high, medium, and low fluorescence intensity regions, as well as images of non-fluorescent regions, are obtained based on different fluorescence intensity ranges. The grayscale value range and fluorescence area ratio of each region are determined, where the fluorescence area ratio is the proportion of fluorescent pixels in the total number of pixels in the image. The grayscale value range and fluorescence area ratio of each region are then matched with mineral type discrimination rules to obtain the corresponding mineral type and mineral content. The mineral type discrimination rules are obtained by combining in-situ Raman spectroscopy analysis of rock fragment samples from different fluorescence intensity regions in several historical rock fragment fluorescence images. The mineral content is the sum of the fluorescence area ratios of the regions corresponding to the same mineral type. The mineral type and mineral content are then matched with lithology identification rules to obtain the corresponding lithology identification results.

[0116] Based on this, the technical solution of the present invention will be described and explained below with reference to several examples.

[0117] Example 1: As shown in the attached document Figure 1 As shown, this embodiment of the invention discloses a lithology identification method based on the fusion of rock debris fluorescence images and Raman spectroscopy, including:

[0118] Step S110: Obtain a fluorescence image of rock cuttings, and the fluorescence type of the rock cutting sample corresponding to the fluorescence image is non-petroleum hydrocarbon fluorescence;

[0119] Step S120: Apply Gaussian weighted method to dynamically segment the fluorescent region, obtain images of high, medium and low fluorescence intensity regions and non-fluorescent regions according to different fluorescence intensity ranges, and determine the gray value range and fluorescence area ratio of each region image, where the fluorescence area ratio is the proportion of the number of fluorescent pixels in the total number of pixels in the image.

[0120] Step S130: Input the gray value range and fluorescence area ratio of each region image into the mineral type discrimination rule for matching to obtain the corresponding mineral type and mineral content. The mineral type discrimination rule is obtained by combining in-situ Raman spectroscopy analysis to analyze the mineral composition of rock debris samples in different fluorescence intensity regions in several historical rock debris fluorescence images. The mineral content is the sum of the fluorescence area ratios of the region images corresponding to the same mineral type.

[0121] Step S140: Input the mineral type and mineral content into the lithology identification rules for matching, and obtain the corresponding lithology identification results.

[0122] The lithology identification rules include:

[0123] (1) Determine the basic name

[0124] When the highest mineral content is > 50%, the corresponding mineral type is taken as the basic name of the rock;

[0125] When all the mineral contents are < 50%, the mineral type with the highest mineral content is taken as the basic name of the rock;

[0126] (2) Determine the additional noun

[0127] When the remaining mineral content is between 50% and 25%, the corresponding mineral type is added in the form of XX quality before the basic name;

[0128] When the remaining mineral content is between 25% and 10%, the corresponding mineral type is added in the form of containing XX before the basic name;

[0129] When the remaining mineral content is < 10%, the corresponding mineral type does not participate in the naming of the rock.

[0130] The fluorescence types of the rock debris fluorescence images include petroleum hydrocarbon fluorescence and mineral fluorescence, but only mineral fluorescence can reflect the mineral type, so the fluorescence type of the rock debris fluorescence image used in the embodiment of the present application must be non-petroleum hydrocarbon fluorescence before lithology identification is performed.

[0131] The mineral type determination rule in the above step S130 is obtained by analyzing the mineral composition of the rock debris samples in different fluorescence intensity regions of a plurality of historical rock debris fluorescence images by in-situ Raman spectrum analysis, that is, the rock debris samples in different fluorescence intensity regions of a plurality of historical rock debris fluorescence images are analyzed by Raman spectrum analysis, a mapping relationship between fluorescence intensity and mineral composition is established, then based on the corresponding relationship between different fluorescence intensity regions and the gray value range, a mapping relationship between the gray value range and the mineral composition is obtained, which is used for subsequent inversion of the mineral type according to different gray value ranges in the rock debris fluorescence image.

[0132] The lithology identification rule in the above step S140 is set according to the rock three-level naming rule, which can include:

[0133] (1) Determine the basic name

[0134] When the highest mineral content is > 50%, the corresponding mineral type is taken as the basic name of the rock;

[0135] When all the mineral contents are < 50%, the mineral type with the highest mineral content is taken as the basic name of the rock;

[0136] (2) Determine the additional noun

[0137] When the remaining mineral content is between 50% and 25%, the corresponding mineral type is added in the form of XX quality before the basic name;

[0138] If the remaining mineral content is between 25% and 10%, the corresponding mineral type is added to the basic name in the form of XX-containing;

[0139] If the remaining mineral content is less than 10%, the corresponding mineral type does not participate in the rock naming.

[0140] For example, if the calcite content is greater than 50%, the rock is named limestone; if the dolomite / ferrodolomite content is greater than 50%, the rock is named dolomite.

[0141] For example, if the calcite content is greater than 50% and the dolomite / ferrodolomite content is between 50% and 25%, the rock is named dolomitic limestone; if the calcite content is greater than 50% and the dolomite / ferrodolomite content is between 25% and 10%, the rock is named dolomite-containing limestone.

[0142] The embodiment of the present application discloses a lithology identification method based on the fusion of the cutting fluorescence image and the Raman spectrum, combines the in-situ Raman spectrum analysis to perform mineral composition analysis on the cutting samples in different fluorescence intensity regions of a plurality of historical cutting fluorescence images, establishes a mapping relationship between the fluorescence intensity and the mineral composition, so that in the lithology identification, no experimental instrument is needed to participate, the mineral type and the mineral content can be directly obtained by combining the mapping relationship between the fluorescence intensity and the mineral composition and the gray value range of the reaction fluorescence intensity, the drilling cost is effectively saved, the lithology can be quickly and accurately identified, and further, in the embodiment, no conventional XRD, XRF and other instrument equipment are used for sample analysis, the problems of long sample analysis period, failure to meet the rapid analysis demand of the logging site and lagging decision are avoided.

[0143] Embodiment 2: as shown in the accompanying Figure 2 The embodiment of the present application is a further optimization of the above-mentioned embodiment, wherein the cutting fluorescence image is acquired, and the acquisition includes:

[0144] Step S210: acquiring any cutting fluorescence image, and combining the three-dimensional fluorescence spectrum analysis result of the cutting sample corresponding to the cutting fluorescence image;

[0145] Step S220: if no fluorescence characteristic region is displayed in the three-dimensional fluorescence spectrum analysis result, performing lithology identification;

[0146] Step S230: if the fluorescence characteristic region is displayed in the three-dimensional fluorescence spectrum analysis result, triggering the fluorescence type re-judgment process;

[0147] It should be noted that if no fluorescence characteristic region is displayed in the three-dimensional fluorescence spectrum of the cutting fluorescence image, the fluorescence type is mineral fluorescence, if the fluorescence characteristic region is displayed, the fluorescence type may be petroleum hydrocarbon fluorescence or caused by drilling fluid additive pollution, and the fluorescence type needs to be re-judged.

[0148] Step S240, obtain the three-dimensional fluorescence spectrum of the rock debris sample corresponding to the rock debris fluorescence image and the three-dimensional fluorescence spectrum of the drilling fluid, and detect the fluorescence feature region in each three-dimensional fluorescence image and extract the corresponding fluorescence feature region feature, wherein the fluorescence feature region feature includes the contour circumscribed rectangle parameter and the contour mask region color feature;

[0149] Specifically, detecting the fluorescence feature region in the two three-dimensional fluorescence images includes:

[0150] (1) Convert both three-dimensional fluorescence spectra from BGR color space to HSV space using the cv2.cvtColor function of the OpenCV library;

[0151] (2) Define the HSV range of the fluorescence feature region in the two three-dimensional fluorescence spectra, generate a binary mask through the cv2.inRange() function, and mark the pixels in the fluorescence feature region in each three-dimensional fluorescence spectrum;

[0152] (3) Morphological optimization to smooth the boundary of the fluorescence feature region in each three-dimensional fluorescence spectrum and improve connectivity, specifically using a 5x5 elliptical kernel to perform mask merging, i.e., 2 times of closed operation to fill holes and gaps in the feature region, and 1 time of open operation to eliminate noise;

[0153] (4) Contour extraction and filtering, for the fluorescence feature region in each three-dimensional fluorescence spectrum, detect all independent connected regions in the mask through the cv2.findContours() function, eliminate small regions with an area <100 pixels, and only keep significant regions, and iterate through all contours to obtain the number N of fluorescence feature regions.

[0154] Specifically, extracting the fluorescence feature region feature includes:

[0155] (1) Obtain the contour circumscribed rectangle parameter, for the fluorescence feature region in each three-dimensional fluorescence spectrum, extract the coordinates (x, y) of the contour circumscribed rectangle, the width w, and the height h through the cv2.boundingRect() function, wherein (x, y) is the top-left corner coordinate of the contour circumscribed rectangle, and w and h are the width and height of the contour circumscribed rectangle, respectively;

[0156] (2) Obtain the contour mask region color feature, calculate the average BGR value of the pixels in the fluorescence feature region using the cv2.mean() function.

[0157] Step S250, perform similarity analysis on the fluorescence feature region features in the two three-dimensional fluorescence images to obtain a comprehensive similarity;

[0158]

[0159] wherein, For comprehensive similarity; The region shape similarity is obtained by weighting the region quantity similarity, location distribution similarity, and shape similarity. This represents the color distribution similarity.

[0160] The above This indicates that the two three-dimensional fluorescence images are identical. A value >0.8 indicates that the two 3D fluorescence images are highly similar. A value greater than 0.6 indicates that the two three-dimensional fluorescence images are moderately similar, while other values ​​indicate that the two three-dimensional fluorescence images are significantly different.

[0161] Specifically, a similarity analysis is performed on the fluorescence feature regions in the two three-dimensional fluorescence images, including:

[0162] Step S251: The region shape similarity is obtained by weighting the region quantity similarity, location distribution similarity, and shape similarity, including:

[0163] Step S2511: Determine the similarity of the number of regions;

[0164]

[0165] in, For the similarity of the number of regions; , These represent the number of fluorescent feature regions in the two three-dimensional fluorescence images, respectively.

[0166] here The time interval indicates the highest similarity in the number of regions, meaning that the number of fluorescent feature regions is exactly the same in both three-dimensional fluorescence images. This indicates the region with the lowest similarity.

[0167] Step S2512, determine the similarity of location distribution, including:

[0168] (1) Obtain the three-dimensional fluorescence spectrum point set of rock cuttings samples Three-dimensional fluorescence spectrum point set of drilling fluid Determine the directed Hausdorff distance between point sets, where the point set of the three-dimensional fluorescence spectrum is the set of geometric center points of each fluorescence feature region in the three-dimensional fluorescence spectrum;

[0169]

[0170] in, for arrive The directed Hausdorff distance; for arrive the directed Hausdorff distance; is a geometric center point in the is a geometric center point in the is and the Euclidean distance between them;

[0171] It should be noted that the coordinates of the geometric center point of the fluorescent feature region are (x+w / 2, y+h / 2).

[0172] (2) Take the maximum value as the final distance :

[0173]

[0174] (3) Normalize the final distance and convert it to a position distribution similarity;

[0175]

[0176] wherein, is the position distribution similarity; is the maximum value of the image size; is the distance normalized by the image size.

[0177] Step S2513, determine the shape similarity, obtain the largest area fluorescent feature region in the three-dimensional fluorescence spectrum of the rock sample and the three-dimensional fluorescence spectrum of the drilling fluid respectively, determine the corresponding Hu moment, calculate the cosine similarity between the two Hu moments , and get the shape similarity after normalization, which maps the range from [-1, 1] to [0, 1];

[0178] ;

[0179] It should be noted that the determination process of the Hu moment can include: calculating the contour moment of the three-dimensional fluorescence spectrum by cv2.moments(), converting the contour moment to the Hu moment by the cv2.Humoments() function, and performing logarithmic transformation on the Hu moment to compress the numerical range and enhance the discrimination.

[0180] Step S2514, obtain the regional shape similarity based on weighting;

[0181]

[0182] wherein, is the regional shape similarity; is the regional number similarity; is a color distribution similarity; is a shape similarity.

[0183] Step S252, for the three-dimensional fluorescence spectrum of the rock sample and the three-dimensional fluorescence spectrum of the drilling fluid, the hue histogram of the fluorescence feature region of the two is obtained respectively, the Bhattacharyya distance between the two is determined, and the color distribution similarity is obtained;

[0184]

[0185] wherein, is a color distribution similarity; is a Bhattacharyya distance.

[0186] It should be noted that the hue histogram (only H channel is counted) of the fluorescence feature region can be realized by the cv2.calcHist() function. After obtaining the hue histogram of the fluorescence feature region of the two, the histogram is further normalized by the cv2.normalize() function, and then the Bhattacharyya distance between the two is calculated by the cv2.compareHist() function. bd ).

[0187] The above indicates that the color distribution similarity is maximum.

[0188] Step S260, if > 0.8, the fluorescence is caused by drilling fluid additive pollution, and lithology identification is performed, and if < 0.8, the fluorescence type is petroleum hydrocarbon fluorescence, and lithology identification is not performed.

[0189] When the rock fluorescence image corresponding to the three-dimensional fluorescence spectrum analysis result of the rock sample has a fluorescence region feature display, the rock sample and the drilling fluid three-dimensional fluorescence spectrum image are based on the region shape similarity and the color distribution similarity of the fluorescence feature region, the comprehensive similarity between the two three-dimensional fluorescence images is comprehensively evaluated, and it is determined whether it is caused by drilling fluid additive pollution or petroleum hydrocarbon fluorescence. Compared with the method of judging whether the fluorescence type is mineral fluorescence by a single factor, the judgment result is more accurate, and a reliable guarantee is provided for the lithology identification method.

[0190] Embodiment 3: as shown in the accompanying Figure 3 The embodiment of the present application is a further optimization of the above-mentioned embodiment, wherein the Gaussian weighting method is applied to dynamically segment the fluorescence region, the high, medium and low fluorescence intensity region images and the no fluorescence region image are obtained according to different fluorescence intensity ranges, and the gray value and the fluorescence area proportion of each region image are determined, including:

[0191] Step S310, the rock fluorescence image is converted into a gray image and preprocessed;

[0192] Specifically, the cv2.cvtColor() function is called to convert the debris fluorescence image (i.e., the RGB fluorescence image) into a grayscale image, and a Gaussian filter (cv2.GaussianBlur) can be used to suppress high-frequency noise, with the filter parameters set to a kernel size ksize=(5, 5) and a standard deviation σ=1.1, effectively eliminating the reflection artifacts on the surface of the debris.

[0193] In step S320, a Gaussian weighted method is applied to the grayscale image for dynamic segmentation to obtain the corresponding fluorescence region, and the fluorescence region is reprocessed by morphological optimization and connected region extraction in sequence.

[0194] The parameter settings in the above Gaussian weighted method can include, but are not limited to, setting the neighborhood size BlockSize=7 and the threshold offset C=5 to enhance the robustness of the algorithm to uneven lighting and noise.

[0195] The above reprocessing of the fluorescence region by morphological optimization and connected region extraction in sequence includes:

[0196] (1) Morphological optimization: 3x3 rectangular structural elements (kernel=np.ones((3,3), np.uint8)) are used for iterative erosion (cv2.erode, 3 iterations), to eliminate isolated noise points in the binary image, and an expansion operation (cv2.dilate, 3 iterations) is performed, with the same structural elements as above, to fill the broken fluorescence regions caused by threshold segmentation.

[0197] (2) Connected region analysis: the cv2.findContours() function is used to extract the connected region contours, and the area threshold (<50 pixels²) is used to filter out interference regions, to achieve the rejection of fragmented regions.

[0198] In step S330, based on the reprocessed fluorescence region, high, medium, and low fluorescence intensity region images and a non-fluorescence region image are obtained according to different fluorescence intensity ranges.

[0199] In step S340, the gray values and fluorescence area proportions of each region image are determined, and the calculation formula of the fluorescence area proportion is as follows:

[0200]

[0201] wherein, is the fluorescence area proportion; is the number of fluorescence pixels in the fluorescence region; is the total number of pixels in the debris fluorescence image.

[0202] Example 4: as shown in the accompanying Figure 4As shown, the embodiments of the present invention are further optimizations of the above embodiments, wherein the establishment of mineral type discrimination rules includes:

[0203] Step S410: Acquire several historical rock cutting fluorescence images, and the fluorescence type of the rock cutting sample corresponding to each historical rock cutting fluorescence image is not petroleum hydrocarbon fluorescence.

[0204] Step S420: Apply Gaussian weighted method to dynamically segment the fluorescent region, obtain images of high, medium and low fluorescence intensity regions and non-fluorescent regions according to different fluorescence intensity ranges, and determine the gray value range and fluorescence area ratio of each region image, where the fluorescence area ratio is the proportion of the number of fluorescent pixels in the total number of pixels in the image.

[0205] The specific implementation process of this step is the same as that in Example 3, and will not be repeated here.

[0206] Step S430: Obtain at least one rock fragment sample from each region image and use in-situ Raman spectroscopy to obtain the mineral composition of each rock fragment sample.

[0207] Step S440: Based on the gray value range of the images of each region and the mineral composition of all rock fragment samples, determine the gray value range of each mineral type, and use the 3sigma principle to remove gray value outliers of each mineral type, and finally obtain the gray value range of each mineral type.

[0208] Furthermore, an update time can be set for the mineral type identification rules, and the mineral type identification rules can be updated according to the update time.

[0209] Example 5: The above embodiments are verified using the embodiments of the present invention, as detailed below:

[0210] (i) Obtain a fluorescence image of rock cuttings, and the fluorescence type of the rock cutting sample corresponding to the rock cutting fluorescence image is non-petroleum hydrocarbon fluorescence;

[0211] (1) Identification of rock cutting fluorescence types, including:

[0212] Step 1.1: Obtain any rock fragment fluorescence image and combine it with the three-dimensional fluorescence spectral analysis results of the corresponding rock fragment sample;

[0213] Step 1.2: If a fluorescent characteristic region is displayed in the 3D fluorescence spectrum, it is necessary to determine whether this characteristic region is caused by drilling fluid additives. Read the 3D fluorescence spectrum of the cuttings and the corresponding 3D fluorescence spectrum of the drilling fluid (as attached). Figure 5 (a, 5c) Call the cv2.cvtColor function of the OpenCV library to convert the three-dimensional fluorescence spectrum from the BGR color space to the HSV color space, which facilitates the detection of fluorescence feature regions.

[0214] Step 1.3. Define the HSV range of the fluorescent feature area, see Appendix Figure 5 a. The fluorescent feature area color in 5c includes red and orange-yellow. The red HSV range is set as (H: 0-10, S: 150-255, S: 150-255), and the orange-yellow HSV range is (H: 15-30, S: 140-255, S: 150-255). A binary mask is generated by the cv2.inRange() function to mark the pixels in the feature peak area in the image.

[0215] Step 1.4. Morphological optimization to smooth the area boundary and improve connectivity. A 5x5 elliptical kernel is used to perform 2 times of closed operation to fill holes and gaps in the feature peak area, and 1 time of open operation to eliminate noise.

[0216] Step 1.5. Contour extraction and filtering. All independent connected domains in the mask are detected by the cv2.findContours() function, and small areas with an area <100 pixels are removed, leaving only significant areas. The contour of the feature area in the three-dimensional fluorescent image of the rock debris is shown in Appendix Figure 5 b. The contour of the feature area in the three-dimensional fluorescent image of the drilling fluid is shown in Appendix Figure 5 d. The number of contour areas of both is 1, see Appendix Figure 5 e. The contour comparison of the feature areas of both is shown in Appendix.

[0217] (2) The region shape similarity is obtained by weighting the region number similarity, position distribution similarity, and shape similarity, including:

[0218] Step 2.1. Determine the region number similarity.

[0219]

[0220] wherein, ;

[0221] Step 2.2. Determine the position distribution similarity.

[0222] (a) The final distance ;

[0223]

[0224] (b) The distance is normalized and converted to the position distribution similarity.

[0225]

[0226] wherein, is ;

[0227] Step 2.3. Determine the shape similarity.

[0228]

[0229] Step 2.4, determining the region shape similarity;

[0230]

[0231] (3) For the three-dimensional fluorescence spectrum of the rock sample and the three-dimensional fluorescence spectrum of the drilling fluid, the hue histogram of the fluorescence characteristic region of the two is obtained respectively, the Bhattacharyya distance between the two is determined, and the color distribution similarity is obtained;

[0232]

[0233] Wherein, the Bhattacharyya distance ;

[0234] (4) determining the comprehensive similarity;

[0235]

[0236] (5) because , it is indicated that the fluorescence is caused by drilling fluid additive pollution, and lithology identification is performed.

[0237] (II) Establishing a mineral type discrimination rule

[0238] (1) Quantitative characterization of the rock sample image, load the rock sample fluorescence image (such as attached Figure 6 a), convert it into a gray scale image and perform pretreatment (such as attached Figure 6 b), use the method disclosed in Example 3 to obtain the gray scale value range and the fluorescence area ratio of the high, medium and low fluorescence intensity region images and the non-fluorescence region image (such as attached Figure 6 c to 6f), wherein the gray scale value range of each region image is 198-255, 107-198, 46-107 and 0-46 in turn.

[0239] (2) In-situ Raman spectrum analysis, in-situ Raman spectrum analysis is performed on the rock sample in the fluorescence region (such as the red cross mark points in attached Figure 6 c to 6f, sample numbers A-H) of the rock sample fluorescence image to determine the mineral composition, and the results are shown in attached Figure 7 and Table 1. The mineral Raman characteristic peak of the high-medium fluorescence intensity region (see attached Figure 6 c and 6d, gray scale value 107-255) is identified as calcite, and the content is 25.21%. The content of calcite can be determined from the fluorescence area ratio of each region image in step (1), and the mineral content is the sum of the fluorescence area ratios of the region images corresponding to the same mineral type. Figure 6e, gray value 46-107) mineral identification as dolomite, content is 42.34%; Non-fluorescent area (see attached Figure 6 f, gray value 0-46) mineral identification as ankerite, content is 32.45%.

[0240] Table 1 Raman characteristic peak displacement data table of carbonate rock minerals

[0241] .

[0242] (3) Obtain 10 historical cutting fluorescence images, repeat the above steps (1) to (2), and count the gray value range of all calcite, dolomite, ankerite and other carbonate rock minerals, and use the 3sigma principle to eliminate the gray value of each mineral, and finally obtain the mineral type discrimination rule, which includes: the gray value range of calcite is 102 to 255, the gray value range of dolomite is 41 to 102, and the gray value range of ankerite is 0 to 41.

[0243] (Three) Obtain 80 cutting fluorescence images of the Ordovician system of X well, identify the lithology based on the method disclosed in the application, and perform Raman spectrum surface scanning on the 80 cutting samples corresponding to the cutting fluorescence image to determine the true lithology, and the results show that: the coincidence rate of the lithology identification results of the application and the Raman surface scanning analysis results reaches 92.5%, and the lithology comparison results are shown in the confusion matrix graph (as shown in the attached Figure 8 ), the method disclosed in the application realizes the rapid and accurate identification of lithology.

[0244] Example 6: as shown in the attached Figure 9 , the application discloses a lithology identification device based on cutting fluorescence image and Raman spectrum fusion, which comprises:

[0245] The original image acquisition unit acquires the cutting fluorescence image, and the fluorescence type of the cutting sample corresponding to the cutting fluorescence image is non-petroleum hydrocarbon fluorescence;

[0246] The image analysis unit applies the Gaussian weighting method to dynamically segment the fluorescence region, obtains high, medium and low fluorescence intensity region images and non-fluorescence region images according to different fluorescence intensity ranges, and determines the gray value range and fluorescence area ratio of each region image, wherein the fluorescence area ratio is the ratio of the number of fluorescence pixels to the total number of image pixels;

[0247] The mineral analysis unit inputs the gray value range and fluorescence area ratio of each region image into the mineral type discrimination rule for matching to obtain the corresponding mineral type and mineral content, wherein the mineral type discrimination rule is obtained by combining in-situ Raman spectrum analysis to analyze the mineral composition of the cutting sample in the different fluorescence intensity regions of a plurality of historical cutting fluorescence images, and the mineral content is the sum of the fluorescence area ratios of the region images corresponding to the same mineral type.

[0248] The lithology identification unit brings in the mineral type and the mineral content to the lithology identification rule for matching to obtain a corresponding lithology identification result, wherein the lithology identification rule comprises:

[0249] (1) Determine the basic name

[0250] When the highest mineral content is > 50%, the corresponding mineral type is taken as the basic name of the rock;

[0251] When all mineral contents are < 50%, the mineral type with the highest mineral content is taken as the basic name of the rock;

[0252] (2) Determine the additional noun

[0253] When the remaining mineral content is between 50% and 25%, the corresponding mineral type is added in the form of XX quality before the basic name;

[0254] When the remaining mineral content is between 25% and 10%, the corresponding mineral type is added in the form of containing XX before the basic name;

[0255] When the remaining mineral content is < 10%, the corresponding mineral type does not participate in the naming of the rock.

[0256] The image analysis unit comprises:

[0257] The preprocessing module converts the cutting fluorescence image into a gray-scale image and performs preprocessing;

[0258] The fluorescence region extraction module applies the Gaussian weighting method to dynamically segment the gray-scale image to obtain the corresponding fluorescence region, and sequentially performs fluorescence region reprocessing through morphological optimization and connected region extraction;

[0259] The fluorescence region reprocessing module obtains high, medium and low fluorescence intensity region images and no fluorescence region image according to different fluorescence intensity ranges based on the reprocessed fluorescence region;

[0260] The fluorescence region quantification module determines the gray-scale value and fluorescence area proportion of each region image, wherein the calculation formula of the fluorescence area proportion is as follows:

[0261]

[0262] Wherein, is the fluorescence area proportion; is the number of fluorescence pixels of the fluorescence region; is the total number of pixels of the cutting fluorescence image.

[0263] The original image acquisition unit comprises:

[0264] The fluorescence type acquisition module acquires any fluorescence image of the drill cutting, and combines the three-dimensional fluorescence spectrum analysis result of the drill cutting sample corresponding to the fluorescence image of the drill cutting;

[0265] The first fluorescence type analysis module performs lithology identification if no fluorescence characteristic region is displayed in the three-dimensional fluorescence spectrum analysis result.

[0266] The second fluorescence type analysis module triggers the fluorescence type re-judgment process if the fluorescence characteristic region is displayed in the three-dimensional fluorescence spectrum analysis result, including:

[0267] The three-dimensional fluorescence spectrum of the drill cutting sample corresponding to the fluorescence image of the drill cutting and the three-dimensional fluorescence spectrum of the drilling fluid are acquired, and the fluorescence characteristic regions in the two three-dimensional fluorescence images are detected respectively, and the corresponding fluorescence characteristic region features are extracted, wherein the fluorescence characteristic region features include the contour circumscribed rectangle parameters and the contour mask region color features.

[0268] The fluorescence characteristic region features in the two three-dimensional fluorescence images are subjected to similarity analysis to obtain a comprehensive similarity.

[0269]

[0270] Wherein, is the comprehensive similarity; is the region shape similarity, which is obtained by weighting the region number similarity, the position distribution similarity and the shape similarity; is the color distribution similarity;

[0271] If > 0.8, the fluorescence is caused by drilling fluid additive pollution, and lithology identification is performed, and if < 0.8, the fluorescence type is oil hydrocarbon fluorescence, and no lithology identification is performed.

[0272] In an embodiment of the present application, an electronic device is disclosed, which comprises a processor and a memory, the memory stores a computer program, the computer program is loaded and executed by the processor to realize the lithology identification method based on the fusion of the fluorescence image of the drill cutting and the Raman spectrum.

[0273] The processor can be a central processing unit (CPU), a general-purpose processor, a digital signal processor (DSP), an ASIC, an FPGA or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It can realize or execute various exemplary logical blocks, modules and circuits described in combination with the disclosure. It can also be a combination of computing functions, such as one or more microprocessor combinations, DSP and microprocessor combinations, etc. The memory can include but is not limited to: U disk, read-only memory, mobile hard disk, magnetic disk or optical disk and various computer program storage media.

[0274] Those skilled in the art will appreciate that embodiments of the present application can be readily used as a method, a system or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer-readable program code. The specification and drawings are, accordingly, to be regarded as illustrative merely.

[0275] The present application is described in relation to flow diagrams and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the present application. It is to be understood that each block of the flow diagrams and / or block diagrams, and combinations of blocks in the flow diagrams and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processing element or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flow diagrams and / or block diagrams block or blocks. Figure 1 one or more functions specified in one or more of the flow diagram or block diagram block or blocks. Figure 1 means for performing the function specified by the flow diagram or block diagram block or blocks.

[0276] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the flow diagram or block diagram block or blocks. Figure 1 one or more functions specified in one or more of the flow diagram or block diagram block or blocks. Figure 1 means for performing the function specified by the flow diagram or block diagram block or blocks.

[0277] The above merely provides a specific implementation of the present application, which has strong adaptability and implementation effects. However, the protection scope of the present application is not limited to this. Any skilled person in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, equivalent changes made according to the claims of the present application are still within the scope covered by the present application.

Claims

1. A lithology identification method based on fusion of a cutting fluorescence image and a Raman spectrum, characterized by, The method comprises the following steps: obtaining a fluorescence image of the cuttings, and the fluorescence type of the cuttings sample corresponding to the fluorescence image is non-petroleum hydrocarbon fluorescence; applying a Gaussian weighting method to dynamically segment the fluorescence region, obtaining high, medium and low fluorescence intensity region images and no fluorescence region image according to different fluorescence intensity ranges, and determining the gray value range and fluorescence area ratio of each region image, wherein the fluorescence area ratio is the ratio of the number of fluorescence pixels to the total number of image pixels; inputting the gray value range and fluorescence area ratio of each region image into the mineral type discrimination rule for matching to obtain the corresponding mineral type and mineral content, wherein the mineral type discrimination rule is obtained by combining in-situ Raman spectrum analysis to analyze the mineral composition of the cuttings sample in different fluorescence intensity regions of a plurality of historical fluorescence images of the cuttings, and the mineral content is the sum of the fluorescence area ratios of the region images corresponding to the same mineral type; inputting the mineral type and mineral content into the lithology identification rule for matching to obtain the corresponding lithology identification result, wherein the lithology identification rule comprises: (1) determining the basic name when the highest mineral content is greater than 50%, the corresponding mineral type is taken as the basic name of the rock; when all mineral contents are less than 50%, the mineral type with the highest mineral content is taken as the basic name of the rock; (2) determining the additional noun when the remaining mineral content is between 50% and 25%, the corresponding mineral type is added in the form of XX quality before the basic name; when the remaining mineral content is between 25% and 10%, the corresponding mineral type is added in the form of containing XX before the basic name; when the remaining mineral content is less than 10%, the corresponding mineral type does not participate in the naming of the rock. 2.The lithology identification method based on fusion of lithodetritus fluorescence image and Raman spectrum according to claim 1, characterized in that, The application of the Gaussian weighting method to dynamically segment the fluorescence region, the obtaining of high, medium and low fluorescence intensity region images and no fluorescence region image according to different fluorescence intensity ranges, and the determination of the gray value and fluorescence area ratio of each region image comprise: converting the fluorescence image of the cuttings into a gray image and performing pretreatment; applying the Gaussian weighting method to the gray image for dynamic segmentation to obtain the corresponding fluorescence region, and sequentially performing fluorescence region reprocessing through morphological optimization and connected region extraction; based on the reprocessed fluorescence region, obtaining high, medium and low fluorescence intensity region images and no fluorescence region image according to different fluorescence intensity ranges; determining the gray value and fluorescence area ratio of each region image, wherein the calculation formula of the fluorescence area ratio is as follows: wherein, is the fluorescent area ratio; is the number of fluorescent pixels of the fluorescent area; is the total number of pixels of the rock debris fluorescent image. 3.The lithology identification method based on fusion of a debris fluorescence image and a Raman spectrum according to claim 1, characterized in that, The establishment of the mineral type discrimination rule comprises: obtaining a plurality of historical fluorescence images of the cuttings, and the fluorescence type of the cuttings sample corresponding to each historical fluorescence image is not petroleum hydrocarbon fluorescence; applying a Gaussian weighting method to dynamically segment the fluorescence region, obtaining high, medium and low fluorescence intensity region images and no fluorescence region image according to different fluorescence intensity ranges, and determining the gray value range and fluorescence area ratio of each region image, wherein the fluorescence area ratio is the ratio of the number of fluorescence pixels to the total number of image pixels; obtaining at least one cuttings sample in each region image, and obtaining the mineral composition of each cuttings sample by in-situ Raman spectrum analysis; Based on the gray value range of each area image and the mineral composition of all the rock debris samples, the gray value range of each mineral type is determined, and the abnormal gray value of each mineral type is removed by using the 3sigma principle, and finally the gray value range of each mineral type is obtained.

4. The lithology identification method based on fusion of debris fluorescence image and Raman spectrum according to claim 1 or 2 or 3, characterized in that, The rock debris fluorescence image is obtained, and the fluorescence type of the rock debris sample corresponding to the rock debris fluorescence image is non-petroleum hydrocarbon fluorescence. Any rock debris fluorescence image is obtained, and the three-dimensional fluorescence spectrum analysis result of the rock debris sample corresponding to the rock debris fluorescence image is combined. If no fluorescence characteristic region is displayed in the three-dimensional fluorescence spectrum analysis result, lithology identification is performed. If the fluorescence characteristic region is displayed in the three-dimensional fluorescence spectrum analysis result, the fluorescence type re-judgment process is triggered. The three-dimensional fluorescence spectrum of the rock debris sample corresponding to the rock debris fluorescence image and the three-dimensional fluorescence spectrum of the drilling fluid are obtained, and the fluorescence characteristic regions in the two three-dimensional fluorescence images are detected respectively, and the corresponding fluorescence characteristic region features are extracted, wherein the fluorescence characteristic region features include contour circumscribed rectangle parameters and contour mask region color features. The similarity of the fluorescence characteristic region features in the two three-dimensional fluorescence images is analyzed to obtain a comprehensive similarity. wherein, is a comprehensive similarity; is a region shape similarity, which is weighted by a region number similarity, a position distribution similarity, and a shape similarity; is a color distribution similarity; If > 0.8, the fluorescence is caused by drilling fluid additive contamination, and lithology identification is performed, if < 0.8, the fluorescence type is oil hydrocarbon fluorescence, and lithology identification is not performed. 5.The lithology identification method based on fusion of lithodetritus fluorescence image and Raman spectrum according to claim 4, characterized in that, The similarity of the fluorescence characteristic region features in the two three-dimensional fluorescence images is analyzed to obtain a comprehensive similarity, including: The hue histogram of the fluorescence characteristic region of the three-dimensional fluorescence spectrum of the rock debris sample and the three-dimensional fluorescence spectrum of the drilling fluid is obtained respectively, the Bhattacharyya distance between them is determined, and the color distribution similarity is obtained. wherein, is a color distribution similarity; is a Bhattacharyya distance; The region shape similarity is obtained by weighting the region number similarity, the position distribution similarity and the shape similarity, including: The region number similarity is determined. wherein, is a region quantity similarity; , are the number of fluorescent feature regions in the two three-dimensional fluorescent images, respectively. The position distribution similarity is determined, including: (1) obtaining a three-dimensional fluorescence spectrum point set of a rock sample and a three-dimensional fluorescence spectrum point set of a drilling fluid , determining a directed Hausdorff distance between the point sets, wherein the three-dimensional fluorescence spectrum point set is a set of geometric center points of each fluorescence feature region in the three-dimensional fluorescence spectrum; wherein, is to a directed Hausdorff distance; is to a directed Hausdorff distance; is a geometric center point in is a geometric center point in is and an Euclidean distance; (2) take the maximum value of both as the final distance : (3) the distance Normalization is done and converted into a position distribution similarity; wherein, is a position distribution similarity; is a maximum of two image sizes; is a distance normalized by image size; Determine shape similarity, respectively acquire the largest area fluorescent characteristic region in the three-dimensional fluorescence spectrum of the drilling fluid and the three-dimensional fluorescence spectrum of the drilling fluid, determine the corresponding Hu moment, and calculate the cosine similarity between the two Hu moments , and get shape similarity after normalization ; ； The region shape similarity is obtained by weighting. wherein, is a region shape similarity; is a region number similarity; is a position distribution similarity; is a shape similarity.

6. A lithology identification device based on fusion of debris fluorescence image and Raman spectrum using the method according to any one of claims 1 to 5, characterized in that, It includes: The rock debris fluorescence image is obtained, and the fluorescence type of the rock debris sample corresponding to the rock debris fluorescence image is non-petroleum hydrocarbon fluorescence. The image analysis unit applies the Gaussian weighting method to dynamically segment the fluorescence region, obtains high, medium and low fluorescence intensity region images and non-fluorescence region image according to different fluorescence intensity ranges, and determines the gray value range and fluorescence area proportion of each region image, wherein the fluorescence area proportion is the proportion of the number of fluorescence pixels in the total number of image pixels. The mineral analysis unit inputs the gray value range and fluorescence area proportion of each region image into the mineral type discrimination rule for matching to obtain the corresponding mineral species and mineral content, wherein the mineral type discrimination rule is obtained by combining the in-situ Raman spectrum analysis to analyze the mineral composition of the rock debris samples in different fluorescence intensity regions of a plurality of historical rock debris fluorescence images, and the mineral content is the sum of the fluorescence area proportions of the region images corresponding to the same mineral species. The lithology identification unit inputs the mineral species and mineral content into the lithology identification rule for matching to obtain the corresponding lithology identification result, wherein the lithology identification rule includes: (1) determining the basic name When the highest mineral content is greater than 50%, the corresponding mineral type is taken as the basic name of the rock; When all mineral contents are less than 50%, the mineral type with the highest mineral content is taken as the basic name of the rock; (2) determine the additional noun If the remaining mineral content is between 50% and 25%, the corresponding mineral type is added in the form of XX quality before the basic name; If the remaining mineral content is between 25% and 10%, the corresponding mineral type is added in the form of containing XX before the basic name; If the remaining mineral content is less than 10%, the corresponding mineral type does not participate in rock naming. 7.The lithology identification device based on fusion of lithodetritus fluorescence image and Raman spectrum according to claim 6, characterized in that, The image analysis unit comprises: A preprocessing module converts the cutting fluorescence image into a gray-scale image and performs preprocessing; A fluorescence region extraction module applies a Gaussian weighting method to the gray-scale image for dynamic segmentation to obtain corresponding fluorescence regions, and sequentially performs fluorescence region reprocessing through morphological optimization and connected region extraction; A fluorescence region reprocessing module obtains high, medium and low fluorescence intensity region images and a non-fluorescence region image according to different fluorescence intensity ranges based on the reprocessed fluorescence regions; A fluorescence region quantification module determines the gray-scale values and fluorescence area proportions of each region image, wherein the calculation formula of the fluorescence area proportion is as follows: wherein, is the fluorescent area proportion; is the number of fluorescent pixels of the fluorescent area; is the total number of pixels of the rock debris fluorescent image. 8.The lithology identification device based on fusion of lithodetritus fluorescence image and Raman spectrum according to claim 6 or 7, characterized in that, The original image acquisition unit comprises: A fluorescence type acquisition module acquires any cutting fluorescence image and combines the three-dimensional fluorescence spectrum analysis result of the cutting sample corresponding to the cutting fluorescence image; A first fluorescence type analysis module performs lithology identification if no fluorescence characteristic region is displayed in the three-dimensional fluorescence spectrum analysis result; A second fluorescence type analysis module triggers a fluorescence type rejudgment process if a fluorescence characteristic region is displayed in the three-dimensional fluorescence spectrum analysis result, including: Acquiring the three-dimensional fluorescence spectrum of the cutting sample corresponding to the cutting fluorescence image and the three-dimensional fluorescence spectrum of the corresponding drilling fluid, detecting the fluorescence characteristic regions in the two three-dimensional fluorescence images, and extracting the corresponding fluorescence characteristic region features, wherein the fluorescence characteristic region features include contour circumscribed rectangle parameters and contour mask region color features; Performing similarity analysis on the fluorescence characteristic region features in the two three-dimensional fluorescence images to obtain a comprehensive similarity; wherein, is a comprehensive similarity; is a region shape similarity, which is weighted by a region number similarity, a position distribution similarity, and a shape similarity; is a color distribution similarity; If > 0.8, the fluorescence is caused by drilling fluid additive contamination, and lithology identification is performed, if < 0.8, the fluorescence type is oil hydrocarbon fluorescence, and lithology identification is not performed.

9. An electronic device, comprising: The device comprises a processor and a memory, and the memory stores a computer program which is loaded and executed by the processor to realize the steps in the method of any one of claims 1 to 5. The device comprises a processor and a memory, and the memory stores a computer program which is loaded and executed by the processor to realize the steps in the method of any one of claims 1 to 5.

Images