CN114140415A - Ultrasonic logging image crack extraction method and device based on deep learning - Google Patents
Ultrasonic logging image crack extraction method and device based on deep learning Download PDFInfo
- Publication number
- CN114140415A CN114140415A CN202111414115.1A CN202111414115A CN114140415A CN 114140415 A CN114140415 A CN 114140415A CN 202111414115 A CN202111414115 A CN 202111414115A CN 114140415 A CN114140415 A CN 114140415A
- Authority
- CN
- China
- Prior art keywords
- network model
- neural network
- logging
- convolutional neural
- real
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000605 extraction Methods 0.000 title claims abstract description 59
- 238000013135 deep learning Methods 0.000 title claims abstract description 42
- 238000012549 training Methods 0.000 claims abstract description 76
- 238000003384 imaging method Methods 0.000 claims abstract description 73
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 65
- 238000000034 method Methods 0.000 claims abstract description 24
- 238000012545 processing Methods 0.000 claims description 10
- 238000012795 verification Methods 0.000 claims description 9
- 238000004422 calculation algorithm Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 7
- 238000002604 ultrasonography Methods 0.000 claims description 7
- 230000006870 function Effects 0.000 claims description 4
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 9
- 239000011159 matrix material Substances 0.000 description 7
- 238000004891 communication Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 239000002245 particle Substances 0.000 description 4
- 230000011218 segmentation Effects 0.000 description 4
- 238000010521 absorption reaction Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000012937 correction Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000001413 cellular effect Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 239000012530 fluid Substances 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 230000008439 repair process Effects 0.000 description 2
- 239000011435 rock Substances 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 239000012855 volatile organic compound Substances 0.000 description 2
- 230000003416 augmentation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000002002 slurry Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10132—Ultrasound image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
- G06T2207/30184—Infrastructure
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The application discloses an ultrasonic logging image crack extraction method and device based on deep learning, wherein the method comprises the following steps: creating an initial convolutional neural network model; acquiring a logging imaging graph, and adding the marked logging imaging graph into a training data set; inputting the training data set into the initial convolutional neural network model for iterative training to obtain a convolutional neural network model with complete training; and acquiring a real-time logging imaging graph, and extracting a fracture image according to the real-time logging imaging graph and the completely trained convolutional neural network model. According to the method, the dense connection module and the SE attention module are added to the convolutional neural network model based on the jump connection, so that the learning capacity of the network model on the local characteristics of the cracks is improved, and the performance of the network model on crack extraction is improved; the influence of the attenuation of the amplitude of the sound wave energy on the crack extraction is eliminated through the sound wave attenuation compensation.
Description
Technical Field
The invention relates to the technical field of oil and gas resource exploration, in particular to an ultrasonic logging image crack extraction method and device based on deep learning, electronic equipment and a computer readable storage medium.
Background
During the exploration and development of oil and gas resources, well logging is an indispensable ring, and the geophysical characteristics of rock strata are utilized to measure geophysical parameters so as to obtain geological and engineering technical data. In recent years, with the development of downhole ultrasonic imaging, a large number of visual ultrasonic well logging images of the well wall and the vicinity of the well wall can be obtained, and the well logging images can reflect lithological characteristics of the well wall, particularly the distribution and specific structure of cracks around the well bore.
The fracture is a geological structure which changes when the rock is stressed, has great influence on the distribution condition of oil gas and water in the stratum, and has guiding significance on oil gas exploration and development. Therefore, accurate identification and quantitative analysis of fractures are very important, and at the same time, it is a difficult point of imaging logging technology. The existing fracture extraction method of imaging logging data usually utilizes the traditional image processing method to extract fractures, and the fractures are usually extracted by image segmentation algorithms such as watershed segmentation, threshold segmentation, wavelet transformation and the like.
However, the existing method still has some defects, and the traditional segmentation method is very dependent on the selection of the threshold, so that over-segmentation or under-segmentation is easily caused. Therefore, the existing methods need to adjust the parameters for many times or meet certain specific application conditions. In addition, due to non-ideality of the imaging process and the like, the crack often has more or less defects, which are apparent imperfections in the image, and this will seriously affect the effective identification and extraction of the crack.
Disclosure of Invention
In view of the above, it is necessary to provide a method, an apparatus, an electronic device and a computer-readable storage medium for extracting a fracture of an ultrasonic logging image based on deep learning, so as to solve the problems of excessive parameter settings and insufficient stability in the existing method for extracting a fracture of a logging image.
In order to solve the above problems, the present invention provides a deep learning-based ultrasonic well logging image fracture extraction method, which includes:
creating an initial convolutional neural network model;
acquiring a logging imaging graph, and adding the marked logging imaging graph into a training data set;
inputting the training data set into the initial convolutional neural network model for iterative training to obtain a convolutional neural network model with complete training;
and acquiring a real-time logging imaging graph, and extracting a fracture image according to the real-time logging imaging graph and the completely trained convolutional neural network model.
Further, a convolutional neural network model based on hopping connections, which incorporates a dense connection module and an SE attention module.
Further, inputting the training data set to the initial convolutional neural network model for iterative training comprises: and setting the size of a training block, the initial learning rate, the loss function, the optimizer, the weight attenuation and the momentum in the initial convolutional neural network model.
Further, inputting the training data set to the initial convolutional neural network model for iterative training comprises:
dividing the training data set into a verification set and a training set, training the neural network by using the training set, and performing performance verification on the network model by using the verification set;
judging whether the trained network model reaches a preset performance standard or not, and if not, continuing to train the network model; and if the preset performance standard is reached, outputting the network model as a completely trained convolutional neural network model.
Further, acquiring a real-time logging imaging map comprises:
acquiring a real-time logging signal;
and processing the real-time logging signals to obtain a real-time logging imaging graph.
Further, processing the real-time logging signal to obtain a real-time logging imaging graph comprises:
compensating the attenuation of the real-time logging signal according to the attenuation characteristic of sound waves in mud, and extracting the effective head wave signal amplitude in the real-time logging signal;
and obtaining a real-time logging imaging graph according to the effective head wave signal amplitude.
Further, the method further comprises the step of repairing the crack according to the extracted crack image by using a preset algorithm to obtain an optimized crack extraction result.
The invention also provides an ultrasonic logging image crack extraction device based on deep learning, which comprises:
the model creating module is used for creating an initial convolutional neural network model;
the data set generating module is used for acquiring a logging imaging graph and adding the marked logging imaging graph into a training data set;
the model optimization module is used for inputting the training data set into the initial convolutional neural network model for iterative training to obtain a convolutional neural network model with complete training;
and the extraction module is used for acquiring a real-time logging imaging graph and extracting a fracture image according to the real-time logging imaging graph and the completely trained convolutional neural network model.
The invention also provides an electronic device, which comprises a memory and a processor, wherein the memory is stored with a computer program, and when the computer program is executed by the processor, the method for extracting the ultrasonic well logging image crack based on the deep learning is realized.
The invention also provides a computer readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method for extracting the ultrasonic well logging image crack based on the deep learning according to any one of the above technical solutions is realized.
Compared with the prior art, the invention has the beneficial effects that: firstly, establishing an initial convolutional neural network model; secondly, acquiring a well logging imaging graph, adding the marked well logging imaging graph into a training data set, and training an initial network model by using the training data set to obtain a completely trained convolutional neural network model; and finally, acquiring a real-time logging imaging graph, and extracting a fracture image according to the real-time logging imaging graph and the completely trained convolutional neural network model. According to the method, the dense connection module and the SE attention module are added to the convolutional neural network model based on the jump connection, so that the learning capacity of the network model on the local characteristics of the cracks is improved, and the performance of the network model on crack extraction is improved; the influence of the attenuation of the amplitude of the sound wave energy on the crack extraction is eliminated through the sound wave attenuation compensation.
Drawings
FIG. 1 is a schematic view of an application scenario of an ultrasonic well logging image fracture extraction device based on deep learning according to the present invention;
FIG. 2 is a schematic flow chart of an embodiment of a deep learning-based ultrasonic well logging image fracture extraction method provided by the invention;
FIG. 3 is a schematic structural diagram of a convolutional neural network model based on a hopping connection provided in an embodiment of the present invention;
FIG. 4(a) is a waveform diagram of a log signal prior to compensation as provided in an embodiment of the present invention;
FIG. 4(b) is a waveform diagram of a compensated log signal provided in an embodiment of the present invention;
FIG. 5(a) is a schematic representation of a real-time log image provided in an embodiment of the present invention;
FIG. 5(b) is a fracture image obtained through a network model provided in an embodiment of the present invention;
FIG. 5(c) is a crack image after crack repair provided in an embodiment of the present invention;
FIG. 6 is a block diagram of an embodiment of an ultrasonic well-logging image fracture extraction device based on deep learning according to the present invention;
fig. 7 is a block diagram of an embodiment of an electronic device provided in the present invention.
Detailed Description
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate preferred embodiments of the invention and together with the description, serve to explain the principles of the invention and not to limit the scope of the invention.
The invention provides an ultrasonic logging image crack extraction method and device based on deep learning, an electronic device and a computer readable storage medium, which are respectively described in detail below.
An embodiment of the present invention provides an application system of an ultrasonic logging image fracture extraction method based on deep learning, and fig. 1 is a scene schematic diagram of an embodiment of an application system of an ultrasonic logging image fracture extraction method based on deep learning provided by the present invention, where the system may include a server 100, and an ultrasonic logging image fracture extraction device based on deep learning, such as the server in fig. 1, is integrated in the server 100.
The server 100 in the embodiment of the present invention is mainly used for:
creating an initial convolutional neural network model;
acquiring a logging imaging graph, and adding the marked logging imaging graph into a training data set;
inputting the training data set into the initial convolutional neural network model for iterative training to obtain a convolutional neural network model with complete training;
and acquiring a real-time logging imaging graph, and extracting a fracture image according to the real-time logging imaging graph and the completely trained convolutional neural network model.
In this embodiment of the present invention, the server 100 may be an independent server, or may be a server network or a server cluster composed of servers, for example, the server 100 described in this embodiment of the present invention includes, but is not limited to, a computer, a network host, a single network server, a plurality of network server sets, or a cloud server composed of a plurality of servers. Among them, the Cloud server is constituted by a large number of computers or web servers based on Cloud Computing (Cloud Computing).
It is to be understood that the terminal 200 used in the embodiments of the present invention may be a device that includes both receiving and transmitting hardware, i.e., a device having receiving and transmitting hardware capable of performing two-way communication over a two-way communication link. Such a device may include: a cellular or other communication device having a single line display or a multi-line display or a cellular or other communication device without a multi-line display. The specific terminal 200 may be a desktop, a laptop, a web server, a Personal Digital Assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, an embedded device, and the like, and the type of the terminal 200 is not limited in this embodiment.
Those skilled in the art will understand that the application environment shown in fig. 1 is only one application scenario of the present invention, and does not constitute a limitation on the application scenario of the present invention, and that other application environments may further include more or fewer terminals than those shown in fig. 1, for example, only 2 terminals are shown in fig. 1, and it is understood that the application system of the deep learning based ultrasound well log image fracture extraction method may further include one or more other terminals, which is not limited herein.
In addition, as shown in fig. 1, the application system of the deep learning-based ultrasonic well logging image fracture extraction method may further include a memory 200 for storing data, such as an initial convolutional neural network model, a vibration center position, and the like.
It should be noted that the scene schematic diagram of the application system of the ultrasonic logging image fracture extraction based on the deep learning shown in fig. 1 is only an example, the application system and the scene of the ultrasonic logging image fracture extraction method based on the deep learning described in the embodiment of the present invention are for more clearly illustrating the technical solution of the embodiment of the present invention, and do not form a limitation on the technical solution provided in the embodiment of the present invention, and it can be known by those skilled in the art that the technical solution provided in the embodiment of the present invention is also applicable to similar technical problems with the evolution of the application system of the ultrasonic logging image fracture extraction method based on the deep learning and the appearance of a new service scene.
The embodiment of the invention provides an ultrasonic logging image crack extraction method based on deep learning, a flow schematic diagram of the method is shown in figure 2, and the ultrasonic logging image crack extraction method based on deep learning comprises the following steps:
step S201, an initial convolutional neural network model is established;
step S202, obtaining a logging imaging graph, and adding the marked logging imaging graph into a training data set;
step S203, inputting the training data set into the initial convolutional neural network model for iterative training to obtain a convolutional neural network model with complete training;
and S204, acquiring a real-time logging imaging graph, and extracting a fracture image according to the real-time logging imaging graph and the completely trained convolutional neural network model.
In the embodiment of the invention, firstly, an initial convolutional neural network model is created; secondly, acquiring a well logging imaging graph, adding the marked well logging imaging graph into a training data set, and training an initial network model by using the training data set to obtain a completely trained convolutional neural network model; and finally, acquiring a real-time logging imaging graph, and extracting a fracture image according to the real-time logging imaging graph and the completely trained convolutional neural network model. The method can improve the learning ability of the network model to the local characteristics of the cracks and improve the extraction performance of the network model to the cracks; and the influence of the attenuation of the amplitude of the sound wave energy on the crack extraction is eliminated through the sound wave attenuation compensation.
As a preferred embodiment, in step S201, the initial convolutional neural network model includes: a convolutional neural network model based on hopping connections that incorporates a dense connection module and an SE attention module.
Specifically, the convolutional neural network model structure based on the jump connection is shown in fig. 3, and the addition of a Dense connection module (Dense Block) improves the efficiency of feature transfer between layers, so that the learned features can be more effectively utilized; the SE Attention module (Attention Block) is added, so that the learning capability of the network on the local characteristics of the crack is improved, and the crack extraction performance of the network is further improved.
As a specific embodiment, in step S202, 200 logging imaging charts are first acquired, and image augmentation is performed by processing methods such as rotation, folding, scaling, and the like, so as to increase the data amount required by training. Wherein each logging imaging graph generates 5 images (including original graphs), and the total number of the logging imaging graphs is 1000, and on the basis, the marking contains 2 types of targets through the marking of VOC (volatile organic compound) format: fracture and background, add the marked 1000 log images to the training dataset.
As a preferred embodiment, in step S203, inputting the training data set to the initial convolutional neural network model for iterative training includes: and setting the size of a training block, the initial learning rate, the loss function, the optimizer, the weight attenuation and the momentum in the initial convolutional neural network model.
Specifically, the training block size is set to 8, the initial learning rate is 0.01, the loss function is crossentry, the optimizer is SGD, the weight attenuation is set to 0.0001 and the momentum is 0.9, and the OneCycle is the learning rate scheduler.
As a preferred embodiment, inputting the training data set to the initial convolutional neural network model for iterative training includes:
dividing the training data set into a verification set and a training set, training the neural network by using the training set, and performing performance verification on the network model by using the verification set;
judging whether the trained network model reaches a preset performance standard or not, and if not, continuing to train the network model; and if the preset performance standard is reached, outputting the network model as a completely trained convolutional neural network model.
Specifically, in order to improve the training effect of the model, cross validation is performed in the training process, the training data set is divided into seven parts, one part is taken as a validation set each time, the rest are taken as training sets, and the model is trained. In the training process, when the performance of the network is not improved any more for 5 continuous periods, the training is stopped, and the finally trained complete convolutional neural network model is obtained.
In step S204, acquiring a real-time log imaging map includes:
acquiring a real-time logging signal;
and processing the real-time logging signals to obtain a real-time logging imaging graph.
As a specific embodiment, the ultrasonic transducer is used for transmitting pulses to the well wall and receiving echo signals to obtain echo signal values of different depths and different orientations, namely real-time logging signals which are represented by S.
As a preferred embodiment, the processing the real-time logging signal to obtain a real-time logging imaging graph includes:
compensating the attenuation of the real-time logging signal according to the attenuation characteristic of sound waves in mud, and extracting the effective head wave signal amplitude in the real-time logging signal;
and obtaining a real-time logging imaging graph according to the effective head wave signal amplitude.
As a specific embodiment, the processing the real-time logging signal comprises the following steps:
step S401, sound wave attenuation compensation:
and compensating the attenuation of the real-time acoustic imaging logging signal by combining the attenuation characteristic of the sound wave in the mud, thereby extracting the amplitude of the effective head wave signal in the echo signal of each depth imaging point.
The sound wave propagates in the mud, the attenuation of the sound intensity is related to the propagation distance and the attenuation coefficient, and the compensation attenuation adopted by the embodiment is as follows:
J＝J0e2cd (1)
wherein J is the post-compensation sound intensity, J0To compensate for the pre-acoustic intensity, a is the attenuation coefficient and l is the propagation distance, in mud the attenuation coefficient a is dominated by the absorption coefficient a1And scattering coefficient a2Two-part, i.e. a ═a1+a2. Wherein, a1Mainly because when the sound wave propagates in the fluid, overcome fluid internal friction and absorb some energy, its calculation formula is:
wherein f is the frequency of the sound wave, c is the speed of the sound wave in the mud, r is the density of the mud, and h is the viscosity coefficient of the mud, and the parameters can be easily obtained, so that the absorption attenuation coefficient can be accurately obtained. In addition, a2Mainly, when the sound wave propagates in the mud, the original propagation direction is changed due to the scattering of the suspended particles, so that the energy is weakened. The calculation formula is as follows:
where f is the frequency of the acoustic wave, c is the velocity of the acoustic wave in the slurry, and a is the diameter of the scattering particles assumed in advance, and in this method, the number of particles per volume n can be calculated assuming that a is 2 μm.
On the basis, the absorption coefficient a can be calculated by the formula (2) and the formula (3)1And scattering coefficient a2. Thereby deriving the attenuation coefficient a. Finally, attenuation compensation is carried out on the logging data by the formula (1).
Although the diameter of the scattering particles is an estimated value and the calculated attenuation coefficient is different from the real attenuation coefficient, the attenuation compensation can still effectively compensate the logging data. The compensation effect can be seen in fig. 4(a) and 4(b), and it can be seen that the logging signal is effectively compensated.
S402, obtaining the amplitude of the head wave signal;
for the compensated logging data (as shown in FIG. 4 (b)), S is usedCompensationIt is shown that, since the first 20 points are the transmitted signals, the transmitted signals should be filtered when calculating the amplitude of the head wave signal, and therefore, the amplitude calculation formula of the head wave signal is:
A[i]＝max(Scompensation) (4)
Wherein A is an amplitude matrix formed by all imaging points, i represents a coordinate value and is measured by current logging data SCompensationIs translated into the coordinate position of (c). That is, each sampling point obtains a set of log data, and then obtains an amplitude value ai from the set of log data]. And finally, forming an amplitude matrix A by amplitude values obtained by all sampling points.
Step S403, azimuth correction;
according to the instrument orientation curve data, carrying out orientation correction on the amplitude extracted by each depth imaging point, and adjusting the amplitude to an orientation coordinate system of a real space, wherein the correction formula is as follows:
wherein A is an amplitude matrix formed by all imaging points, i is a coordinate value, Z is an instrument azimuth angle, the amplitude matrix can be obtained from azimuth curve data, N is the total number of scanning points, and percent is a remainder operation.
S404, normalizing and imaging the amplitude matrix;
all amplitude values are normalized, and the normalization formula is as follows:
wherein A is an amplitude matrix, i is a coordinate value, max (A), and min (A) is respectively a maximum value and a minimum value of A, and all amplitude values are converted into 0-1 through a formula (6) to obtain a real-time logging imaging graph.
As a specific example, as shown in fig. 5(a) and 5(b), fig. 5(a) is a real-time logging imaging graph, the size of the real-time logging imaging graph is scaled to 280 × 280 pixels, and a well-trained convolutional neural network model is input to perform fracture image extraction; the fracture image output by the network model is shown in fig. 5 (b).
As a preferred embodiment, the deep learning-based ultrasound well logging image fracture extraction method further includes: and repairing the crack according to the extracted crack image by using a preset algorithm to obtain an optimized crack extraction result.
Specifically, because partial fracture response of the imaging log is overlapped with background pixels, the model cannot be completely extracted as a complete fracture, and some breakpoints sometimes exist in the segmentation result of the model. In order to obtain a better extraction effect, the minimum spanning tree algorithm is used for connecting the cracks, so that the condition that the cracks have breakpoints can be improved, and the method specifically comprises the following steps:
firstly, obtaining an end point set { T ] by taking end points of all local cracks in the graph1,T2,…,Ti}. For the resulting set of endpoints { T }1,T2,…,TiAnd constructing a special adjacency matrix L.
And G ═ V, E > is a connected weighted graph of the post-endpoint fracture image, V is the obtained endpoint set { T1, T2, …, Ti }, and E is the edge set.
The initial state is: u ═ U1,u2}，V＝{v1,v2…, TE { }. Wherein u is1，u2To take the 2 endpoints of the longest local fracture unit in the fracture image after endpoint, V1, V2 are the endpoints in endpoint set V.
Finding an edge (U, V) with the minimum weight value in all the edges (U, V) belonging to U and V-U0,v0) The edge is added to the set TH, while the other vertex v of the edge is added0And v0Another point v belonging to the same fracture1And incorporate U.
If U is equal to V, the algorithm is ended; otherwise, repeating the step (2). At the end of the algorithm, TH includes the n/2-1 edge of G. All the edges selected through the above steps just form a minimum spanning tree of the graph G, and the minimum spanning tree is connected with the disconnected part in the crack.
As shown in fig. 5(b), the crack image extracted from the network model is optimized, and the crack extraction result after the crack repair is optimized is shown in fig. 5 (c). It can be seen that the fracture extraction results are significantly improved.
The embodiment of the invention provides an ultrasonic logging image crack extraction device based on deep learning, which has a structural block diagram, as shown in fig. 6, the ultrasonic logging image crack extraction device 600 based on deep learning comprises:
a model creating module 601, configured to create an initial convolutional neural network model;
a data set generating module 602, configured to obtain a logging imaging graph, and add the marked logging imaging graph to a training data set;
a model optimization module 603, configured to input the training data set to the initial convolutional neural network model for iterative training, so as to obtain a completely trained convolutional neural network model;
and an extraction module 604, configured to obtain a real-time logging imaging graph, and perform fracture image extraction according to the real-time logging imaging graph and the completely trained convolutional neural network model.
As shown in fig. 7, the invention further provides an electronic device, which can be a mobile terminal, a desktop computer, a notebook, a palm computer, a server and other computing devices, according to the method for extracting a crack in an ultrasonic logging image based on deep learning. The electronic device comprises a processor 10, a memory 20 and a display 30.
The storage 20 may in some embodiments be an internal storage unit of the computer device, such as a hard disk or a memory of the computer device. The memory 20 may also be an external storage device of the computer device in other embodiments, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), etc. provided on the computer device. Further, the memory 20 may also include both an internal storage unit and an external storage device of the computer device. The memory 20 is used for storing application software installed in the computer device and various data, such as program codes installed in the computer device. The memory 20 may also be used to temporarily store data that has been output or is to be output. In one embodiment, the memory 20 stores a deep learning-based ultrasound well logging image fracture extraction method program 40, and the deep learning-based ultrasound well logging image fracture extraction method program 40 can be executed by the processor 10, so as to implement the deep learning-based ultrasound well logging image fracture extraction method according to the embodiments of the present invention.
The processor 10 may be, in some embodiments, a Central Processing Unit (CPU), microprocessor or other data Processing chip for executing program code stored in the memory 20 or Processing data, such as performing a deep learning-based ultrasonic well log image fracture extraction program.
The display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch panel, or the like in some embodiments. The display 30 is used for displaying information at the computer device and for displaying a visual user interface. The components 10-30 of the computer device communicate with each other via a system bus.
The embodiment also provides a computer readable storage medium, on which a program of the deep learning based ultrasonic well logging image fracture extraction method is stored, and when the program is executed by a processor, the deep learning based ultrasonic well logging image fracture extraction method is implemented according to any one of the technical solutions described above.
The invention discloses an ultrasonic logging image crack extraction method based on deep learning, a device, electronic equipment and a computer readable storage medium, wherein firstly, an initial convolution neural network model is created; secondly, acquiring a well logging imaging graph, adding the marked well logging imaging graph into a training data set, and training an initial network model by using the training data set to obtain a completely trained convolutional neural network model; and finally, acquiring a real-time logging imaging graph, and extracting a fracture image according to the real-time logging imaging graph and the completely trained convolutional neural network model.
According to the technical scheme, the dense connection module and the SE attention module are added to the convolutional neural network model based on the jump connection, so that the learning capacity of the network model on the local characteristics of the cracks is improved, and the performance of the network model on crack extraction is improved; the influence of the attenuation of the amplitude of the sound wave energy on the crack extraction is eliminated through the sound wave attenuation compensation.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention.
Claims (10)
1. An ultrasonic logging image crack extraction method based on deep learning is characterized by comprising the following steps:
creating an initial convolutional neural network model;
acquiring a logging imaging graph, and adding the marked logging imaging graph into a training data set;
inputting the training data set into the initial convolutional neural network model for iterative training to obtain a convolutional neural network model with complete training;
and acquiring a real-time logging imaging graph, and extracting a fracture image according to the real-time logging imaging graph and the completely trained convolutional neural network model.
2. The deep learning-based ultrasonic well logging image fracture extraction method according to claim 1, wherein the initial convolutional neural network model comprises: a convolutional neural network model based on hopping connections that incorporates a dense connection module and an SE attention module.
3. The deep learning based ultrasonic well logging image fracture extraction method of claim 1, wherein inputting the training data set to the initial convolutional neural network model for iterative training comprises: and setting the size of a training block, the initial learning rate, the loss function, the optimizer, the weight attenuation and the momentum in the initial convolutional neural network model.
4. The deep learning based ultrasonic well logging image fracture extraction method of claim 1, wherein inputting the training data set to the initial convolutional neural network model for iterative training comprises:
dividing the training data set into a verification set and a training set, training the neural network by using the training set, and performing performance verification on the network model by using the verification set;
judging whether the trained network model reaches a preset performance standard or not, and if not, continuing to train the network model; and if the preset performance standard is reached, outputting the network model as a completely trained convolutional neural network model.
5. The deep learning-based ultrasonic well logging image fracture extraction method according to claim 1, wherein obtaining a real-time well logging imaging map comprises:
acquiring a real-time logging signal;
and processing the real-time logging signals to obtain a real-time logging imaging graph.
6. The deep learning-based ultrasonic well logging image fracture extraction method according to claim 5, wherein the processing the real-time well logging signals to obtain a real-time well logging imaging map comprises:
compensating the attenuation of the real-time logging signal according to the attenuation characteristic of sound waves in mud, and extracting the effective head wave signal amplitude in the real-time logging signal;
and obtaining a real-time logging imaging graph according to the effective head wave signal amplitude.
7. The deep learning-based ultrasonic well logging image fracture extraction method according to claim 1, further comprising:
and repairing the crack according to the extracted crack image by using a preset algorithm to obtain an optimized crack extraction result.
8. The utility model provides an ultrasonic logging image crack extraction element based on degree of depth study which characterized in that includes:
the model creating module is used for creating an initial convolutional neural network model;
the data set generating module is used for acquiring a logging imaging graph and adding the marked logging imaging graph into a training data set;
the model optimization module is used for inputting the training data set into the initial convolutional neural network model for iterative training to obtain a convolutional neural network model with complete training;
and the extraction module is used for acquiring a real-time logging imaging graph and extracting a fracture image according to the real-time logging imaging graph and the completely trained convolutional neural network model.
9. An electronic device comprising a memory and a processor, the memory having stored thereon a computer program that, when executed by the processor, implements the deep learning-based ultrasound log image fracture extraction method of any of claims 1-7.
10. A computer-readable storage medium, having stored thereon a computer program which, when executed by a processor, implements the method of depth-learning based ultrasound log image fracture extraction according to any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111414115.1A CN114140415B (en) | 2021-11-25 | 2021-11-25 | Ultrasonic logging image crack extraction method and device based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111414115.1A CN114140415B (en) | 2021-11-25 | 2021-11-25 | Ultrasonic logging image crack extraction method and device based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114140415A true CN114140415A (en) | 2022-03-04 |
| CN114140415B CN114140415B (en) | 2024-10-25 |
Family
ID=80391843
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111414115.1A Active CN114140415B (en) | 2021-11-25 | 2021-11-25 | Ultrasonic logging image crack extraction method and device based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114140415B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115576014A (en) * | 2022-10-26 | 2023-01-06 | 江苏科技大学 | Intelligent identification method for fractured reservoir based on acoustic wave remote detection imaging |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110208859A (en) * | 2019-05-07 | 2019-09-06 | 长江大学 | Oil-base mud well crack quantitative parameter intelligence computation method based on ultrasonic imaging |
| US20190290210A1 (en) * | 2018-03-22 | 2019-09-26 | Shenzhen Mindray Bio-Medical Electronics Co., Ltd. | Predictive use of quantitative imaging |
| CN111583291A (en) * | 2020-04-20 | 2020-08-25 | 中山大学 | Layer segmentation method and system for retina layer and effusion region based on deep learning |
| US20210192712A1 (en) * | 2019-12-18 | 2021-06-24 | Schlumberger Technology Corporation | Methods for transmitting data acquired downhole by a downhole tool |
-
2021
- 2021-11-25 CN CN202111414115.1A patent/CN114140415B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190290210A1 (en) * | 2018-03-22 | 2019-09-26 | Shenzhen Mindray Bio-Medical Electronics Co., Ltd. | Predictive use of quantitative imaging |
| CN110208859A (en) * | 2019-05-07 | 2019-09-06 | 长江大学 | Oil-base mud well crack quantitative parameter intelligence computation method based on ultrasonic imaging |
| US20210192712A1 (en) * | 2019-12-18 | 2021-06-24 | Schlumberger Technology Corporation | Methods for transmitting data acquired downhole by a downhole tool |
| CN111583291A (en) * | 2020-04-20 | 2020-08-25 | 中山大学 | Layer segmentation method and system for retina layer and effusion region based on deep learning |
Non-Patent Citations (1)
| Title |
|---|
| 罗会兰;易慧;: "基于迭代训练和集成学习的图像分类方法", 计算机工程与设计, no. 05, 16 May 2020 (2020-05-16) * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115576014A (en) * | 2022-10-26 | 2023-01-06 | 江苏科技大学 | Intelligent identification method for fractured reservoir based on acoustic wave remote detection imaging |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114140415B (en) | 2024-10-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9064066B2 (en) | Methods and systems for well planning based on a complex fracture model | |
| CA2823710C (en) | Methods and systems regarding models of underground formations | |
| US10760405B2 (en) | Apparatus and method for semi-three-dimensional rock formation imaging along the borehole sidewall | |
| US11592590B2 (en) | Well log channel matching | |
| CN115201902B (en) | Fault intelligent identifying method and system based on deep learning | |
| CN114140415B (en) | Ultrasonic logging image crack extraction method and device based on deep learning | |
| US11971515B2 (en) | Deep learning model with dilation module for fault characterization | |
| CN112147688B (en) | Reservoir oil gas identification method and device | |
| CN111859709B (en) | Geologic statistics simulation method and device for aquifer structure variation transition probability | |
| US10565467B2 (en) | Extracting well log data | |
| CN113435059B (en) | Model establishing method, crack initiation event diagnosis method and device | |
| CN115079257A (en) | Q value estimation and seismic attenuation compensation method based on fusion network | |
| CN104314558B (en) | Method for judging reservoir effectiveness by Stoneley wave energy loss degree | |
| CN113550735A (en) | Method for calculating oil content of shale oil reservoir by using logging information | |
| US11199642B2 (en) | Adaptive signal decomposition | |
| US12072460B2 (en) | System and method for determining a set of first breaks of a seismic dataset | |
| CN117823041B (en) | Geosteering model building method and device, electronic equipment and storage medium | |
| CN116757096B (en) | Determination method, device, equipment and medium for deep, thin and compact sandstone gas reservoir development well | |
| US11965407B2 (en) | Methods and systems for wellbore path planning | |
| Wang et al. | Study on a New Rock Fracability Evaluation Model of Shale Gas Reservoir | |
| CN105700018A (en) | Earthquake attribute optimizing method and earthquake attribute optimizing device | |
| CN117312975A (en) | Reservoir type classification method and device, electronic equipment and storage medium | |
| CN118426055A (en) | Porosity inversion method, device, equipment and medium based on support vector machine | |
| CN118778133A (en) | Crack and micro-crack identification method and device, electronic equipment and storage medium | |
| CN117988714A (en) | Method, device, equipment and medium for determining drilling track of horizontal well |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |