Drill core digital mapping method and system based on mechanical rotation scanning

Abstract

The invention relates to the technical field of drilling core imaging, in particular to a drilling core digital mapping method and system based on mechanical rotation scanning. A digital mapping system may include a detection device, a sensing device, and a mapping device. The detection equipment comprises a scanning lens which is lowered into a well, can rotate freely and is used for collecting a drilled core image. The sensing device is installed in the scanning lens for direction positioning. And the mapping equipment is electrically connected with the detection equipment, and generates a rock core environment simulation image according to the image acquired by the scanning lens. The mapping equipment is provided with a deep learning module; a mapping result of mapping equipment can be identified, and a bad area in the mapping result is confirmed; and the scanning lens can be controlled to recollect the image of the bad area, and feature compensation is performed on the image of the bad area to generate a complete and clear rock core environment simulation image, so that the problem of bad acquisition of the bad area in the prior art is solved, and the integrity of rock core image acquisition and the quality of the image are enhanced.

Description

Technical Field

[0001] The invention relates to the technical field of borehole core imaging, and in particular to a borehole core digital mapping method and system based on mechanical rotation scanning. Background Art

[0002] Existing technologies usually use borehole imagers to collect images of borehole cores. The working principle of borehole imagers is to analyze high-definition images of the borehole wall to identify and record geological structure information. Specifically, the device uses a camera close to the borehole wall to shoot, and then uses computer software to process and analyze the images to generate a series of high-quality three-dimensional images of stratum lithology, brittleness, cracks, mud height, etc., so as to accurately determine the structural properties of the rock formation.

[0003] However, the existing technology has many blind spots and poor areas for the imaging of borehole cores, that is, the images collected by the borehole imager have problems such as blur, dark spots, and ghosting, which seriously affects the quality of borehole core imaging;

[0004] For example, the patent with publication number CN209525257U discloses a borehole core imaging spectral scanning system. This patent realizes the technical effect of "the system can realize and accurately control the rotation of the borehole core around the axis, ensure that the imaging spectrometer can accurately focus on the core surface, ensure that the scanning direction of the imaging spectrometer is parallel to the core axis and the illumination of the scanned core is uniform" through the technical scheme of "a borehole core imaging spectral scanning system, the system includes an integral frame, a core rotation device, a core rotation control platform, a horizontal-rotation scanning coordination device and a lighting lamp group, the core rotation device is installed in the bottom of the integral frame, the core rotation control platform is installed in the middle of the integral frame, the horizontal-rotation scanning coordination device is installed on the top of the beam of the integral frame, and the lighting lamp group is installed on the horizontal-rotation scanning coordination device".

[0005] However, this technology only compensates for bad features by enhancing illumination. It still cannot solve the problem of hard-to-detect bad areas caused by occlusion of the device itself, bending / hiding of the device structure, etc., and there is room for improvement. Summary of the invention

[0006] The purpose of the present invention is to overcome the shortcomings of the existing technology that the borehole core images obtained have many blind spots and bad areas, and to provide a method and system for digital mapping of borehole cores based on mechanical rotation scanning.

[0007] In a first aspect, the present invention provides a drilling core digital mapping system based on mechanical rotation scanning, which at least includes: a detection device, a sensor device and a mapping device. The detection device at least includes a scanning lens that is lowered into the well and can rotate freely; the scanning lens is used to collect drilling core images. The scanning lens is equipped with a sensor device, and the sensor device is used for the scanning lens to perform directional positioning. The mapping device is electrically connected to the detection device, and performs a mapping operation on the real core posture according to the image collected by the scanning lens to generate a core environment simulation image.

[0008] Preferably, the mapping device is configured with a deep learning module; the deep learning module can identify the mapping results of the mapping device and confirm the bad areas in the mapping results; and the deep learning module can control the scanning lens to re-capture the image of the bad area, so that the mapping device can perform feature compensation on the image of the bad area to generate a complete and clear core environment simulation image.

[0009] According to a preferred embodiment, the deep learning module is equipped with a neural network algorithm model, and the neural network algorithm model performs neural network learning using the images captured by the scanning lens and the images generated by the imaging device as samples.

[0010] According to a preferred embodiment, the neural network algorithm model adopts RNN neural network. The calculation formula between input, output and change is:

[0011] yt=softmax(xt+ht*why+by)

[0012] Among them, yt is the output, why is the change, xt is the input, ht is the previous frame image, by is the next frame image, and the softmax function is the activation function of the RNN neural network; wherein the input is the parameter of the image conversion collected by the scanning lens, the output is the parameter learned and corrected by the neural network, and the change is the calculation rule of pooling or convolution in the neural network.

[0013] According to a preferred embodiment, after the deep learning module calculates the output quantity through the calculation formula between the input quantity, the output quantity and the change quantity, it comprehensively analyzes the output quantity of each image collected by the scanning lens and generates a set of discrete degrees, and then performs average difference processing on the set. And the deep learning module determines whether the output quantity is a bad area based on the discrete degree of each output quantity and the average difference. In the case where the difference between the discrete degree of the output quantity and the average discrete degree exceeds the average difference of the discrete degree, the deep learning module determines that the output quantity is a bad area.

[0014] According to a preferred embodiment, the detection device and the sensing device are arranged in a supporting manner and are topologically connected to the single-arranged mapping device.

[0015] According to a preferred embodiment, the deep learning module is also provided with a control unit for controlling the scanning lens to rescan the bad area when a bad area of ​​the image is identified, so that the imaging device can re-read the image of the bad area to compensate for the features.

[0016] According to a preferred embodiment, the specific manner in which the imaging device performs feature compensation includes at least:

[0017] The control unit controls the scanning lens to acquire a plurality of images of the defective area;

[0018] The neural network algorithm model carried by the deep learning module processes the plurality of images of the bad area to obtain a plurality of outputs corresponding to the bad area images, and generates a set of discrete degrees of the outputs of the bad area images;

[0019] The imaging device obtains an image in the set whose discreteness is closest to the average discreteness as feature compensation.

[0020] According to a preferred embodiment, the topological connection is a star topological structure consisting of a single imaging device and the matching detection device and sensor device, that is, the imaging device is the center of the galaxy, and the detection device and the sensor device are distributed as planetary points.

[0021] The present invention also provides a method for digital mapping of borehole cores based on mechanical rotation scanning. The method for digital mapping of borehole cores is implemented using the system for digital mapping of borehole cores based on mechanical rotation scanning provided by the present invention, and the method for digital mapping of borehole cores includes at least the following steps:

[0022] Step S1: The scanning lens moves and actively collects the borehole core image;

[0023] Step S2: the mapping device integrates and analyzes the image and simulates the core environment;

[0024] Step S3: The deep learning module controls the scanning lens to read the defective area and perform feature compensation;

[0025] Step S4: Circularly compensate for the poorly captured area and update the poor image.

[0026] Wherein, the step S3 also includes the deep learning module identifying the output result of the step S2 to determine the bad area in the simulated core environment.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] After initially acquiring images through a scanning lens, the present invention uses deep learning to control the scanning lens to stay in a poor acquisition area for additional acquisition, then generates compensatory features for the poor area, and finally obtains a borehole core image with high authenticity and clear environment through analysis, effectively solving the problem of poor acquisition in poor areas in the prior art and enhancing the integrity of core image acquisition and image quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a workflow diagram of a drilling core digital mapping system based on mechanical rotation scanning;

[0030] Figure 2 This is an imaging example diagram of an F-Theta scanning lens in an embodiment of a method for digital mapping of a borehole core based on mechanical rotation scanning;

[0031] Figure 3 The present invention is a sampling flow chart of an embodiment of a method for digital mapping of borehole cores based on mechanical rotation scanning. DETAILED DESCRIPTION

[0032] The present invention is further described in detail below in conjunction with test examples and specific implementation methods. However, this should not be understood as the scope of the above subject matter of the present invention being limited to the following embodiments, and all technologies realized based on the content of the present invention belong to the scope of the present invention.

[0033] Unless otherwise specified, in the description of the specific embodiments of the present invention, the terms indicating the orientation or position relationship such as "up", "down", "left", "right", "center", "inside", "outside", etc. are all expressions based on the orientation or position relationship shown in the drawings, or the orientation or position relationship when the invented product / equipment / device is usually used. These terms of orientation or position relationship are only for the convenience of describing the scheme of the present invention or simplifying the description in the specific embodiments, so as to facilitate the technicians to quickly understand the scheme, and do not indicate or imply that a specific device / component / element must have a specific orientation, or be constructed and operated in a specific position relationship, and therefore cannot be understood as a limitation on the present invention.

[0034] In addition, if the terms "horizontal", "vertical", "overhanging", "parallel" and the like appear, it does not mean that the corresponding devices / components / elements are required to be absolutely horizontal or vertical or overhanging or parallel, but may be slightly tilted or have deviations. For example, "horizontal" only means that its direction is more horizontal than "vertical", and does not mean that the structure must be completely horizontal, but may be slightly tilted. Alternatively, it can be simplified to mean that the corresponding devices / components / elements are set in directions such as "horizontal", "vertical", "overhanging", "parallel", etc., and can have an error / deviation of ±10% relative to the corresponding direction setting, more preferably an error / deviation within ±8%, more preferably an error / deviation within ±6%, more preferably an error / deviation within ±5%, and more preferably an error / deviation within ±4%. As long as the corresponding device / component / element is within the error / deviation range, it can still achieve its role in the scheme of the present invention.

[0035] In addition, the expressions “first”, “second”, “third”, etc. that appear in the terms are merely descriptions used to distinguish the same or similar components and should not be understood as emphasizing or implying the relative importance of specific components.

[0036] In addition, in the description of the embodiments of the present invention, "several", "plurality" and "a number" represent at least 2. It can be any number such as 2, 3, 4, 5, 6, 7, 8, 9, and even more than 9.

[0037] In addition, in the description of the technical solution of the present invention, unless otherwise clearly specified / defined / restricted, the terms "set", "install", "connect", "connected", "provided with", "laid", and "arranged" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection, and can be welding, riveting, bolting, threading, and other commonly used connection means in the field. This connection can be a mechanical connection, an electrical connection, or a communication connection; it can be a direct connection, or an indirect connection through an intermediate medium, and it can be the internal connection of two elements.

[0038] Example 1

[0039] The present embodiment provides a digital mapping system for borehole cores based on mechanical rotation scanning, which at least includes: a detection device, a sensor device and a mapping device. The detection device at least includes a scanning lens that is used to be lowered into the well and can rotate freely; the scanning lens is used to collect images of the borehole core. The scanning lens is configured with a sensor device. The sensor device can be installed inside the scanning lens and provide the scanning lens with directional positioning. The mapping device is electrically connected to the detection device, and performs a mapping operation on the real core posture according to the image collected by the scanning lens to generate a core environment simulation image. Preferably, the mapping device is configured with a deep learning module; the deep learning module can identify the mapping results of the mapping device and confirm the bad areas in the mapping results; and the deep learning module can control the scanning lens to re-collect images of the bad areas, so that the mapping device performs feature compensation on the images of the bad areas to generate a complete and clear core environment simulation image.

[0040] After initially acquiring an image through a scanning lens, this embodiment uses deep learning to control the scanning lens to stay in a poor acquisition area for additional acquisition, and then generates compensatory features for the poor area. Finally, after analysis, a borehole core image with high authenticity and clear environment is acquired, which effectively solves the problem of poor acquisition in poor areas in the prior art and enhances the integrity of core image acquisition and the quality of the image.

[0041] Example 2

[0042] This embodiment is a further improvement of Embodiment 1, and repeated contents are not repeated. In this embodiment, the deep learning module is equipped with a neural network algorithm model, and the neural network algorithm model uses the images collected by the scanning lens and the images generated by the imaging device as samples for neural network learning.

[0043] Preferably, the neural network algorithm model adopts RNN neural network. The calculation formula between input, output and change is:

[0044] yt=softmax(xt+ht*why+by)

[0045] Among them, yt is the output, why is the change, xt is the input, ht is the previous frame image, by is the next frame image, and the softmax function is the activation function of the RNN neural network; among them, the input is the parameter converted from the image collected by the scanning lens, the output is the parameter learned and corrected by the neural network, and the change is the calculation rule of pooling or convolution in the neural network.

[0046] Preferably, after calculating the output quantity through the calculation formula between the input quantity, the output quantity and the change quantity, the deep learning module comprehensively analyzes the output quantity of the images collected by each scanning lens and generates a set of discrete degrees, and then performs average difference processing on the set. And the deep learning module determines whether the output quantity is a bad area according to the discrete degree and the average difference of each output quantity. In the case where the difference between the discrete degree of the output quantity and the average discrete degree exceeds the average difference of the discrete degree, the deep learning module determines that the output quantity is a bad area.

[0047] Preferably, the detection device and the sensing device are arranged in a supporting manner and are topologically connected to a single arranged mapping device.

[0048] Preferably, the deep learning module is also provided with a control unit for controlling the scanning lens to rescan the bad area when a bad area of ​​the image is identified, so that the imaging device can re-read the image of the bad area to compensate for the features.

[0049] Preferably, the specific manner in which the imaging device performs feature compensation includes at least:

[0050] The control unit controls the scanning lens to obtain a plurality of images of the defective area;

[0051] The neural network algorithm model carried by the deep learning module processes several images of the bad area to obtain several outputs corresponding to the bad area images, and generates a set of discrete degrees of the output of the bad area images;

[0052] The imaging device obtains the image whose discreteness is closest to the average discreteness in the set as feature compensation.

[0053] Preferably, the topological connection is a star topological structure consisting of a single imaging device and matching detection devices and sensor devices, that is, the imaging device is the center of the galaxy, and the detection devices and sensor devices are distributed as planetary points.

[0054] Example 3

[0055] This embodiment is a further improvement of Embodiment 1 and Embodiment 2, and the repeated contents are not repeated here. This embodiment provides a drilling core digital mapping system based on mechanical rotation scanning.

[0056] The digital mapping system of the borehole core based on mechanical rotation scanning may include a detection device, a sensor device and a mapping device.

[0057] The detection equipment includes a scanning lens lowered into the well, and the scanning lens adopts a free-rotating structure. Preferably, the scanning lens can adopt an existing rotating structure similar to a spherical camera, which can rotate 360° or approximately 360°. Preferably, the scanning lens can also adopt a panoramic lens or a lens with a 360° adjustable shooting range. Preferably, the scanning lens is an F-Theta scanning lens. F-Theta lenses are commonly used in high-performance laser scanning systems to produce focal points, and the lateral displacement of these focal points depends linearly on the scanning angle, which is a required characteristic under normal circumstances, such as laser material processing applications. In VirtualLab Fusion, a scanning light source can be provided for such situations, and different field solvers are used in combination to conveniently evaluate the performance of such scanning lens systems. For details, please refer to Figure 2 , this figure is an example of the effect of the F-Theta scanning lens. It can be seen that the spot diameter sampled by the F-Theta scanning lens is accurate, because the F-Theta scanning lens has good steering control function. Preferably, the free rotation structure of the scanning lens not only borrows the motion structure of the spherical camera from the mechanical point of view, but also borrows the motion mode of the spherical camera from the control parameters, system and module to achieve remote / one-way / direct signal control mode. Therefore, the functional relationship between the sampling image and the sampling angle is accurate, which can effectively improve the position accuracy of the sampling of the drill core and reduce the interference caused by the position offset.

[0058] In this embodiment, the sensing device may be a position sensor, which is installed inside the scanning lens and is used by the scanning lens to perform direction positioning.

[0059] The sensing device is installed inside the scanning lens and provides the scanning lens with direction positioning. The sensing device includes a magnetic sensor. The magnetic sensor is a device that converts the change of magnetic properties of the sensitive element caused by external factors such as magnetic field, current, stress strain, temperature, light, etc. into an electrical signal, and detects the corresponding physical quantity in this way.

[0060] Magnetic sensors are divided into three categories:

[0061] Compass: The earth generates a magnetic field, and a sensor indicates direction based on the magnetic field on the earth's surface.

[0062] Current sensor: Current sensor is also a magnetic field sensor. Current sensor can be used in household appliances, smart grid, electric vehicles, wind power generation, etc.

[0063] Position sensor: If there is a position change between a magnet and a magnetic sensor relative to each other, if this position change is linear, it is a linear sensor; if it is rotational, it is a rotational sensor.

[0064] In this embodiment, the magnetic sensor is a position sensor. Specifically, a variable magnetization axis and a power supply are provided inside the magnetic sensor. The variable magnetization axis refers to the phenomenon that the magnetization direction of a material changes under the action of an external magnetic field. Usually, the magnetization axis refers to the direction in which the material is magnetized in different directions. In some specific materials, the direction of the external magnetic field will affect the magnetic properties of the material, causing the magnetization axis to change. This material is called a material with a variable magnetization axis. This property has important applications in fields such as magnetic recording and magnetic storage. Therefore, in this application, it is used to assist in controlling the activity of the scanning lens. The variable magnetization axis is electrically connected to the power supply, and the power supply is electrically connected to the imaging device.

[0065] The mapping device is electrically connected to the detection device and performs mapping operations on the detection device that truly displays the core posture. Specifically, the collected parameters are transmitted to the detection device through a magnetic sensor, and then uploaded to the mapping device by the detection device.

[0066] Preferably, the detection device and the sensing device are arranged in a matching manner and are topologically connected to a single mapping device. The arrangement of the detection device and the sensing device belongs to a one-to-many topological connection arrangement, that is, one set of detection device is used in conjunction with multiple sets of sensing devices. The detection device and the mapping device are a one-to-one topological connection arrangement.

[0067] Topological connection can be understood as a low-level connection method. Specifically, the network topology structure refers to the physical layout that uses transmission media to connect various devices such as computers. It refers to the geometric shape formed in the interconnection process. It can represent the network configuration of network servers and workstations and the connection between them. In other words, the topological connection here is used for network transmission between entities / units.

[0068] Preferably, the workflow of the drilling core digital mapping system is as follows: Figure 1 As shown, it may include:

[0069] Step S1: The scanning lens moves and actively collects the borehole core image;

[0070] Step S2: the mapping device integrates and analyzes the image and simulates the core environment;

[0071] Step S3: The deep learning module controls the scanning lens to read the defective area and perform feature compensation;

[0072] Step S4: Circularly compensate for the poorly captured area and update the poor image.

[0073] In step S1, the scanning lens actively collects the borehole core image in a manner indicating that the scanning lens itself is a particle and moves in six directions of up, down, left, right, front and back as the travel route of the scanning lens, and an avoidance sensor is provided on the outside of the scanning lens. When the scanning lens is obstructed during its movement, the avoidance sensor will automatically find the remaining five degrees of freedom directions to move, and then move in the movable direction after excluding the directions that cannot be moved in turn, and the lens in the scanning lens can rotate and shoot 360°, thereby simulating the image of the borehole core in space.

[0074] Preferably, the deep learning module is equipped with a neural network algorithm model, and the neural network algorithm model carried in the deep learning module performs neural network learning with various parameters captured by the scanning lens and pictures generated by the imaging device as samples, so that the deep learning module that has completed learning can identify the bad areas of the image output by the imaging device.

[0075] Preferably, the deep learning module is also provided with a control unit for controlling the F-Theta lens to rescan the bad area when a bad image area is identified, so that the imaging device can re-read the image of the bad area to compensate for the features, thereby generating a complete and clear core environment simulation image.

[0076] Preferably, the deep learning module can be set in the mapping device so that it can be exposed to more learning samples, which is beneficial to the accuracy of neural network learning.

[0077] The sampling flow chart is as follows Figure 3 As shown, the F-Theta lens scans and collects the drill core; the imaging device obtains the data scanned by the F-Theta lens and generates an image; the deep learning module recognizes the image generated by the imaging device to determine whether there is a bad area in the image; if there is no bad area in the image, the image is output; if there is a bad area in the image, the deep learning module controls the F-Theta lens to rescan the bad area; the above operations are repeated until there is no bad area in the image.

[0078] The specific way for the deep learning module to determine whether there is a bad area in the image is to determine whether the image is clear or blurry; if the image is clear, it means that the image is well imaged; if the image is blurry, it means that there is a bad area in the image.

[0079] This embodiment transmits the image through the electrical connection between the magnetic sensor and the imaging device on the one hand, and compensates for the poor feature imaging through the neural network on the other hand, and then generates the image of the core through comprehensive analysis.

[0080] Preferably, the drilling core digital mapping system further comprises a deep learning module. The deep learning module is arranged in the mapping device and performs feature compensation for poorly sampled areas.

[0081] Preferably, the deep learning module uses RNN neural network for deep learning. Since the RNN neural network captures a basic recurrent neuron (in this embodiment, it is the image captured by the scanning lens), then distinguishes the input amount (i.e., the parameters converted after image acquisition), output amount (i.e., the parameters learned and corrected by the neural network), and change amount (the calculation rules of pooling or convolution in the neural network) in the recurrent neuron, and performs specific neural network learning.

[0082] The calculation formula between input, output and change is:

[0083] yt=softmax(xt+ht*why+by)

[0084] Among them, yt is the output, why is the change, xt is the input, ht is the previous frame image, by is the next frame image, and the softmax function is the activation function of the RNN neural network, which is also the primary function for simulating the probability distribution of the neural network.

[0085] The calculation formula between input, output and change is regarded as a sub-formula n in the sum formula Sn. The change why is the average difference MD in the sum formula Sn.

[0086]

[0087] Among them, y n is one of the items in the calculation formula with original yt as output, why as change and xt as input, and n is a sub-formula of a specific item in the calculation formula with original yt as output, why as change and xt as input. After the output as the output image is calculated through the calculation formula between the input, output and change, the output of each image is comprehensively analyzed and a set of discrete degrees is generated. Then, the set is processed by average difference to obtain the image with the closest discrete degree as feature compensation, which can effectively avoid the problem of large image error caused by accidental value.

[0088] Preferably, in this embodiment, the topological connection is a star topology consisting of a single mapping device and supporting detection equipment and sensor equipment, that is, the mapping device is the center of the galaxy, and the detection equipment and sensor equipment are densely distributed as planetary points. In the star topology, each node in the network is connected to a central node (also known as a central transfer station, generally a hub or switch) in a point-to-point manner, and the central node transmits information to the destination node. From the comprehensive star topology, it can be seen that the central node implements a centralized communication control strategy, so the central node is quite complex and the burden is much heavier than each node. In a star network, any two nodes must be controlled by the central node to communicate, that is to say, the present application can use a star topology to realize the topological connection of multiple F-Theta scanning lenses and a single mapping device. This means can effectively shorten the acquisition time of the F-Theta scanning lens and improve the imaging rate of the drill core.

[0089] Preferably, the sensing device and the imaging device communicate based on the 1-Wire protocol. The 1-Wire technology has the advantages of saving IO resources, simple structure, low cost, and convenient bus expansion and maintenance. It has a variety of products such as temperature recorders, clocks, switches, memory, IO, etc., and is widely used in many fields such as temperature and humidity measurement, equipment identification and authentication, and storage. The 1-Wire protocol in this application has the advantages of high propagation rate and low propagation interference. The detection device and the imaging device communicate based on the IIC protocol, which is a serial communication interface with a bidirectional two-wire synchronous serial bus, usually consisting of two wires - SDA (serial data line) and SCL (serial clock line) and a pull-up resistor. They are used in projects that require many different components (such as sensors, pins, extensions, and drivers) to work together because they can connect up to 128 devices to the motherboard while maintaining a clear communication path. It is used to connect various low-speed devices together, such as microcontrollers, EEPROMs, A / D and D / A converters, etc., so the IIC protocol communication has reliable performance in the application.

[0090] Example 4

[0091] This embodiment provides a method for digital mapping of borehole cores based on mechanical rotation scanning. The method provided in this embodiment can be implemented using the digital mapping system for borehole cores based on mechanical rotation scanning provided in Embodiments 1, 2 and 3.

[0092] The method for digital mapping of drill cores based on mechanical rotation scanning provided in this embodiment may include the following steps:

[0093] Step S1: The scanning lens moves and actively collects the borehole core image;

[0094] Step S2: the mapping device integrates and analyzes the image and simulates the core environment;

[0095] Step S3: The deep learning module identifies the output result of step S2, determines the bad area in the simulated core environment, and the deep learning module can also control the scanning lens to read the bad area and perform feature compensation;

[0096] Step S4: Circularly compensate for the poorly captured area and update the poor image.

[0097] The working principle of this embodiment is: after initially acquiring an image through a scanning lens, deep learning is used to control the scanning lens to stay in a poor acquisition area for additional acquisition, and then compensatory features of the poor area are generated. Finally, after analysis, a borehole core image with high authenticity and clear environment is obtained, which effectively solves the problem of poor acquisition in poor areas in the prior art and enhances the integrity of image acquisition and the quality of the image.

[0098] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A digital mapping system for borehole cores based on mechanical rotation scanning, It is characterized in that At least: Detection equipment, sensing equipment and imaging equipment; The detection equipment at least includes a scanning lens that is lowered into the well and can rotate freely; the scanning lens is used to collect the drilling core image; The scanning lens is equipped with a sensor device, and the sensor device is used for the scanning lens to perform direction positioning; The mapping device is electrically connected to the detection device, and performs a mapping operation on the real core posture according to the image collected by the scanning lens to generate a core environment simulation image; Wherein, the mapping device is configured with a deep learning module; The deep learning module can identify the mapping results of the mapping device and confirm the bad areas in the mapping results; and the deep learning module can control the scanning lens to re-capture the image of the bad area, so that the mapping device can perform feature compensation on the image of the bad area to generate a complete and clear core environment simulation image.

2. The drilling core digital mapping system based on mechanical rotation scanning according to claim 1, It is characterized in that The deep learning module is equipped with a neural network algorithm model, and the neural network algorithm model uses the images captured by the scanning lens and the images generated by the imaging device as samples to perform neural network learning.

3. The drilling core digital mapping system based on mechanical rotation scanning according to claim 2, It is characterized in that The neural network algorithm model adopts RNN neural network. The calculation formula between input, output and change is: yt=softmax(xt+ht*why+by) Among them, yt is the output, why is the change, xt is the input, ht is the previous frame image, by is the next frame image, and the softmax function is the activation function of the RNN neural network; Among them, the input quantity is the parameter of the image conversion collected by the scanning lens, the output quantity is the parameter learned and corrected by the neural network, and the change quantity is the calculation rule of pooling or convolution in the neural network.

4. The drilling core digital mapping system based on mechanical rotation scanning according to claim 3, It is characterized in that After the deep learning module calculates the output quantity through the calculation formula between the input quantity, the output quantity and the change quantity, it comprehensively analyzes the output quantities of the images collected by each scanning lens and generates a set of discrete degrees, and then performs average difference processing on the set; And the deep learning module determines whether the output quantity is a bad area according to the discrete degree of each output quantity and the average difference; When the difference between the discrete degree of the output and the average discrete degree exceeds the average difference of the discrete degrees, the deep learning module determines that the output is a bad area.

5. The drilling core digital mapping system based on mechanical rotation scanning according to claim 4, It is characterized in that The detection device and the sensing device are arranged in a supporting manner and are topologically connected to the mapping device which is arranged individually.

6. The drilling core digital mapping system based on mechanical rotation scanning according to claim 5, It is characterized in that The deep learning module is also provided with a control unit for controlling the scanning lens to rescan the bad area when a bad area of ​​the image is identified, so that the imaging device can re-read the image of the bad area to compensate for the features.

7. The drilling core digital mapping system based on mechanical rotation scanning according to claim 6, It is characterized in that The specific method of feature compensation performed by the mapping device at least includes: The control unit controls the scanning lens to acquire a plurality of images of the defective area; The neural network algorithm model carried by the deep learning module processes the plurality of images of the bad area to obtain a plurality of outputs corresponding to the bad area images, and generates a set of discrete degrees of the outputs of the bad area images; The imaging device obtains an image in the set whose discreteness is closest to the average discreteness as feature compensation.

8. The drilling core digital mapping system based on mechanical rotation scanning according to claim 5, It is characterized in that The topological connection is a star topological structure consisting of a single mapping device and the matching detection device and sensor device.

9. A method for digital mapping of borehole cores based on mechanical rotation scanning. It is characterized in that The method for digital mapping of drill cores comprises at least the following steps: Step S1: The scanning lens moves and actively collects the borehole core image; Step S2: the mapping device integrates and analyzes the image and simulates the core environment; Step S3: The deep learning module controls the scanning lens to read the defective area and perform feature compensation; Step S4: cyclically compensating for the poorly captured area and updating the poor image; Wherein, the step S3 also includes the deep learning module identifying the output result of the step S2 to determine the bad area in the simulated core environment.

Images