Self-adaptive trajectory control method for rotary steerable drilling based on image processing

Abstract

The invention relates to the technical field of logging while drilling and intelligent steering control in petroleum and natural gas drilling engineering, in particular to a rotary steering drilling self-adaptive trajectory control method for image processing. Comprising the following steps: receiving an original logging-while-drilling image data frame of an underground transmission channel; generating a high-resolution prediction geological texture map; synchronously generating a feature uncertainty distribution map representing pixel probability distribution features in the high-resolution prediction geological texture map; calculating a matching degree between the high-resolution prediction geological texture map and a preset stress-induced artifact feature library; generating a sensor contamination marker signal; calling a safety baseline geologic model constructed before drilling; calculating a geological texture structure difference value relative to the safety baseline geological model; generating a geological illusion entropy; receiving a geological illusion entropy and a sensor pollution mark signal; the downhole steering tool is driven according to the activated mode. According to the method, it is ensured that the guiding instruction is always based on the real geological information, and drilling trajectory deviation caused by brain compensation imaginary textures is avoided.

Description

Technical Field

[0001] This invention relates to the field of logging-while-drilling and intelligent steerable control technology in oil and gas drilling engineering, specifically to an image-processing-based adaptive trajectory control method for rotary steerable drilling. Background Technology

[0002] Rotary steerable drilling systems are core equipment for oil and gas resource exploration and development. Their trajectory control accuracy directly determines the reservoir drilling rate and engineering safety. The system mainly relies on logging-while-drilling imaging data, downhole steering tools and decision algorithms to guide the drill bit to track the target reservoir by analyzing formation texture features in real time. Existing technologies are limited by the transmission bandwidth of downhole mud pulses or electromagnetic waves, resulting in low sampling rates, high noise, and blurred edges in raw drilling image data. Although deep learning super-resolution technology can improve image quality, generative models are prone to generating algorithmic illusions that violate physical principles in areas lacking data support. Furthermore, stress-induced artifacts caused by sensor crystal distortion under high temperature and pressure conditions are easily misjudged as real geological bedding. Traditional closed-loop control systems assume that the front-end sensing data is absolutely reliable and lack dynamic verification and circuit-breaking mechanisms for AI model uncertainties and hardware physical failures. This makes it difficult to distinguish between real geological changes and false signal interference, leading to the system executing incorrect guidance commands when there are model cognitive biases or sensor failures, causing trajectory deviations or even engineering accidents. Therefore, there is an urgent need for an adaptive trajectory control scheme that can effectively suppress the risks of sensing noise and model illusions.

[0003] The information disclosed in the background section above is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention discloses an image processing-based adaptive trajectory control method for rotary steerable drilling. Specifically, the technical solution of this invention includes: The raw logging-while-drilling image data frames are received from the downhole transmission channel through the low-quality sensing reconstruction unit. The original logging-while-drilling image data frames are processed by running a deep learning image super-resolution model to generate a high-resolution predicted geological texture map; Simultaneously generate a feature uncertainty distribution map that characterizes the pixel probability distribution features in the high-resolution predicted geological texture map; The matching degree between the high-resolution predicted geological texture map and the preset stress-induced artifact feature library is calculated by the artifact feature identification unit. When the matching degree exceeds a preset benchmark, a sensor contamination marker signal is generated; The geological illusion entropy calculation unit calls the safety baseline geological model constructed before drilling. Calculate the difference in geological texture structure between the high-resolution predicted geological texture map and the safe baseline geological model; Data from the feature uncertainty distribution map is extracted, and the geological texture structure difference values ​​are weighted to generate geological illusion entropy. The geological illusion entropy and the sensor contamination marker signal are received by the dual-modal arbitration control unit; If the geological illusion entropy is lower than a preset threshold and no contamination marker signal is received from the sensor, the intelligent navigation mode is activated, and the trajectory planning algorithm is given control authority. If the geological illusion entropy exceeds the preset cognitive collapse threshold, or if the sensor contamination marker signal is received, the mechanical baseline mode is activated, and the trajectory planning algorithm is deprived of its control authority. The downhole directional tool is driven by the execution and feedback correction unit according to the activated mode.

[0005] Preferably, the data in the feature uncertainty distribution map is extracted, and the difference values ​​of the geological texture structure are weighted to generate a geological illusion entropy, including: Perform modal alignment operation to project and rasterize the safe baseline geological model into a baseline trend field; The baseline trend field has the same spatial resolution as the high-resolution predicted geological texture map; Calculate the grayscale deviation or gradient deviation between the high-resolution predicted geological texture map and the baseline trend field pixel by pixel; For regions where the grayscale deviation or the gradient deviation exists, extract the corresponding feature uncertainty value of the region in the feature uncertainty distribution map; If the value of the uncertainty of the feature is higher than the preset high confidence threshold, the weight of the deviation value of that region in the total entropy calculation is increased exponentially; If the value of the uncertainty of the feature is lower than the preset low confidence threshold, the weight of the deviation value of that region in the total entropy calculation is suppressed; The weighted deviation values ​​of the entire image are aggregated to obtain the geological illusion entropy.

[0006] Preferably, the safety baseline geological model is constructed based on low-frequency seismic data or regional geological patterns; The safety baseline geological model only includes the basic stratigraphic trend of the average dip angle of the strata; The stress-induced artifact feature library contains regular stripe signal features generated by the sensor under pressure deformation; The regular stripe signal features have specific spatial frequency and directionality.

[0007] Preferably, activating the mechanical baseline mode and depriving the trajectory planning algorithm of control includes: Stop executing image feature-based tilting or azimuth-shifting commands; Lock the current wellbore trajectory tangent direction; Perform pure geometric steady-slope drilling operations to guide the drill bit through the interference area.

[0008] Preferably, the method further includes: During the operation of the mechanical baseline mode, the changing trend of the geological illusion entropy is monitored in real time; Once the geological illusion entropy is detected to have fallen back to a safe range and remained within a preset time window, it is determined that the data quality has been restored. Automatically unlock the intelligent navigation mode; Artificial intelligence algorithms are now allowed to intervene and fine-tune the trajectory.

[0009] Preferably, the downhole guidance tool is driven by an execution and feedback correction unit according to the activated mode, including: When the intelligent navigation mode command is received, the push plate or directional motor is driven to track the reservoir. When the instruction for the mechanical baseline mode is received, the control parameters are locked; Reduce the response sensitivity gain of the downhole guiding tool actuator; Filter out high-frequency oscillation commands caused by residual image noise.

[0010] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention introduces a geological illusion entropy calculation unit, innovatively utilizing low-frequency seismic data to construct a safe baseline geological model, and then comparing its topological structure differences with AI-generated super-resolution images. Combined with weighted calculations using feature uncertainty distribution maps, the system can quantitatively evaluate the credibility of the AI-reconstructed images. This mechanism effectively prevents deep learning models from generating false bedding or fracture features when processing low-quality logging data, ensuring that guidance commands are always based on real geological information and avoiding drilling trajectory deviations caused by algorithms artificially inventing textures.

[0011] 2. This invention addresses the problem of sensors in deep well drilling often deforming due to high pressure and outputting interference signals. The invention incorporates an artifact feature identification unit. By matching real-time images with a stress-induced artifact feature library containing regular stripe signals, the system can accurately identify and mark sensor contamination signals not caused by geological factors. Once such hardware noise is detected, sensor contamination marking is triggered, preventing the control system from misinterpreting mechanical deformation as formation boundaries or fractures. This eliminates the engineering risks of incorrect directional or azimuth adjustments caused by hardware failure data.

[0012] 3. This invention incorporates an automatic switching logic between intelligent navigation mode and mechanical baseline mode. When excessive geological illusion entropy or sensor contamination is detected, the dual-modal arbitration control unit immediately deprives the trajectory planning algorithm of its authority and forcibly switches to mechanical baseline mode. In this mode, by locking the current wellbore tangent direction, reducing the sensitivity of the actuator, and performing pure geometric stabilization, the system can blindly drill across interference zones even when the data environment deteriorates. This fault-oriented safety design ensures that the drill string can maintain a smooth and stable wellbore trajectory even when the AI ​​is blind or hallucinating.

[0013] 4. This invention significantly improves control stability through a unique weighted entropy calculation and hysteresis recovery strategy. On the one hand, by exponentially amplifying the deviation weights in high-confidence regions, the system becomes extremely sensitive to potential risks. On the other hand, when disabling the mechanical baseline mode, the geological illusion entropy is required to fall back to a safe range and remain within a preset time window. This sensitive triggering and prudent recovery strategy effectively filters out high-frequency oscillation commands caused by residual image noise, avoids frequent jumps between intelligent and mechanical control modes, and ensures the smoothness of the wellbore trajectory during long-section well operations. Attached Figure Description

[0014] The present invention will be further explained below with reference to the accompanying drawings and embodiments: Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0016] Example 1 Please see Figure 1 An image processing-based adaptive trajectory control method for rotary steered drilling includes: The raw logging-while-drilling image data frames are received from the downhole transmission channel through the low-quality sensing reconstruction unit. Run a deep learning image super-resolution model to process the original logging-while-drilling image data frames and generate high-resolution predicted geological texture maps; Simultaneously generate a feature uncertainty distribution map that characterizes the pixel probability distribution features in a high-resolution predicted geological texture map; The matching degree between the high-resolution predicted geological texture map and the preset stress-induced artifact feature library is calculated by the artifact feature discrimination unit. When the matching degree exceeds the preset benchmark, a sensor contamination marker signal is generated; the geological illusion entropy calculation unit calls the safety baseline geological model constructed before drilling. Calculate the difference in geological texture structure between the high-resolution predicted geological texture map and the safe baseline geological model; Data from the feature uncertainty distribution map is extracted, and the geological texture structure difference values ​​are weighted to generate the geological illusion entropy. The dual-modal arbitration control unit receives geological illusion entropy and sensor pollution marker signals; If the geological illusion entropy is lower than the preset threshold and no sensor pollution marker signal is received, the intelligent navigation mode is activated, and the trajectory planning algorithm is given control authority. If the geological illusion entropy exceeds the preset cognitive collapse threshold, or if a sensor contamination marker signal is received, the mechanical baseline mode is activated, and the trajectory planning algorithm is deprived of its control. The downhole directional tool is driven by the execution and feedback correction unit according to the activated mode; As the primary parsing hub of the data stream, the low-quality perception reconstruction unit's core task is to address the sparsity problem of raw data caused by the limited transmission of mud pulses or electromagnetic waves in downhole. This unit is configured to continuously ingest raw logging-while-drilling image data frames via a physical telemetry interface. Such data is typically limited by low sampling rates, resulting in high noise and blurred edges. To reconstruct formation details, this unit loads and executes a pre-trained deep learning image super-resolution model. This model employs a generative adversarial network architecture and reconstructs the raw low-quality data into a high-resolution predicted geological texture map with sub-millimeter texture features by learning the mapping relationship between a large amount of core scan data and logging responses. While outputting predicted textures, the model employs Monte Carlo dropout to generate uncertainty; that is, during the inference phase, the Dropout layer in the network remains active, and the model continuously executes the same frame of raw logging-while-drilling image data. The next random forward propagation obtains... Zhang has a prediction map with slight differences; calculate the value of each pixel. Gray-scale variance in Zhang's prediction chart After normalizing the variance matrix, a feature uncertainty distribution map is generated; The distribution map is essentially a two-dimensional scalar field aligned with the spatial resolution of the texture map. The numerical value of any coordinate point in the field quantifies the confidence level of the model in reconstructing the geological features at that location: high numerical regions indicate that the texture details at that location mainly originate from the statistical association and completion of the model, lacking direct support from the original sensor signals; low numerical regions indicate that the features at that location have a strong basis in the original data response. The artifact feature identification unit performs signal screening for non-geological interference. This unit uses a pattern matching algorithm to perform cross-correlation calculations between the frequency domain features of the high-resolution predicted geological texture map and a pre-set stress-induced artifact feature library. When the calculated matching degree index exceeds the preset benchmark, it indicates that the stripe features in the current image are highly consistent with the lattice distortion signal generated by the sensor crystal under high temperature and high pressure stress in terms of spatial frequency and directionality, rather than a natural sedimentary structure. At this time, the unit immediately triggers the sensor contamination label signal. The geological illusion entropy calculation unit initiates a quantitative assessment of AI cognitive risks. This unit retrieves a safe baseline geological model representing the physical bottom line of geological structures and calculates the difference in geological texture structure between the current predicted texture map and the baseline model in terms of bedding direction and continuity. To accurately distinguish between real geological abrupt changes and false algorithmic illusions, this unit introduces a feature uncertainty distribution map as a weighting operator to perform nonlinear weighted aggregation of the difference values, thereby outputting the geological illusion entropy. This parameter is a dynamically changing dimensionless index, whose mathematical essence is approximately the weighted relative entropy or Kullback-Leibler divergence between the predicted distribution and the physical baseline distribution. Its physical meaning lies in characterizing the degree of risk of the generative model deviating from the physical baseline and making excessive associations in the absence of data support. The decision-making terminal consists of a dual-modal arbitration control unit. This unit establishes a reverse coupling logic between perception confidence and control authority based on the safety fallback design principle. During real-time monitoring, if the geological illusion entropy remains below a preset threshold and there are no contamination markers, the system determines that the inference logic of the AI ​​model is convergent and reliable, and then activates the intelligent navigation mode, fully authorizing the trajectory planning algorithm to call the guidance tool to maximize the reservoir drilling rate. Once the geological illusion entropy exceeds the preset cognitive collapse threshold, or a sensor contamination marker signal is detected, the system determines that the current environment has exceeded the effective inference boundary of the AI ​​model and there is an extremely high risk of induced collision. It immediately forces a switch to the mechanical baseline mode. This operation will cut off the instruction output path of the optimization algorithm, forcing the system to execute a deterministic bottom-line drilling strategy. This embodiment elaborates on how the system establishes a dynamic balance between data-driven flexibility and engineering physical safety. Its core innovation lies in breaking through the limitation of the traditional closed-loop control system's assumption that the front-end sensing data is absolutely reliable. By introducing entropy evaluation and physical feature library dual verification, a circuit breaker protection mechanism is constructed to address the defects of generative AI illusions, effectively preventing major engineering accidents caused by drilling platforms deviating from the safety window due to the adoption of false geological images.

[0017] Example 2 Data from the feature uncertainty distribution map is extracted, and the differences in geological texture structure are weighted to generate geological illusion entropy, including: Perform modal alignment operation to project and rasterize the safe baseline geological model into a baseline trend field; The baseline trend field and the high-resolution predicted geological texture map have the same spatial resolution; the grayscale deviation or gradient deviation between the high-resolution predicted geological texture map and the baseline trend field is calculated pixel by pixel. For regions with grayscale deviation or gradient deviation, extract the corresponding feature uncertainty values ​​in the feature uncertainty distribution map for that region. If the feature uncertainty value is higher than the preset high confidence threshold, the weight of the deviation value in the total entropy calculation of that region is increased exponentially. If the feature uncertainty value is lower than the preset low confidence threshold, the weight of the deviation value in the total entropy calculation of that region is suppressed. By aggregating the weighted bias values ​​of the entire image, the geological illusion entropy is obtained; The geological illusion entropy calculation unit executes a rigorous quantization process to separate real geological signals from algorithmic noise using mathematical methods. This unit performs modal alignment, projecting a low-dimensional safety baseline geological model along the wellbore trajectory using spatial interpolation algorithms and rasterizing it into a baseline trend field with pixel-level correspondence to the high-resolution predicted geological texture map. Gaussian pyramid downsampling or low-pass filtering is then applied to the high-resolution predicted geological texture map to ensure its spatial resolution matches the baseline trend field, thus constructing a unified deviation calculation coordinate system. This step ensures that subsequent difference calculations are based on a unified spatial coordinate system and resolution. The unit performs pixel-by-pixel deviation calculation to obtain the numerical difference between the predicted texture map and the baseline trend field in terms of gray intensity or local gradient direction; the key processing logic lies in the subsequent weighted operation: for each pixel with significant deviation, the system indexes its corresponding value in the feature uncertainty distribution map. When the uncertainty value is higher than the preset high confidence threshold, the system determines that the deviation at that point is very likely to be a hallucination feature that is created out of thin air. To this end, the system uses an exponential function to weight and amplify the deviation value in that area, so that it occupies the dominant weight in the final entropy value, thereby significantly increasing the overall risk score. When the uncertainty value is lower than the preset low confidence threshold, the system determines that although the deviation at this point deviates from the baseline model, it is a real geological heterogeneity. At this time, the system uses a logarithmic decay or linear suppression function to reduce the weight of the deviation value to prevent real unexpected geological discoveries from triggering the circuit breaker mechanism. The system integrates or aggregates the weighted deviation values ​​of all pixels in the entire image to output the geological illusion entropy. This embodiment achieves refined quality inspection of AI-generated content by constructing this nonlinear weighted evaluation system based on reliability. It ensures that the system only alarms when the model generates non-physical real features, and remains silent when the model discovers real details not included in the baseline based on the data. This preserves the detection sensitivity of complex geology while ensuring safety.

[0018] Example 3 The safety baseline geological model is constructed based on low-frequency seismic data or regional geological patterns; The safe baseline geological model only includes the basic stratigraphic trend of the mean dip angle of the strata; The stress-induced artifact feature library contains regular stripe signal features generated by the compressive deformation of the sensor; Regular stripe signals have specific spatial frequencies and directions; The physical source and attributes of the system's decision-making criteria are clearly defined, which is the cornerstone of ensuring the system's objectivity. The safe baseline geological model refers to a coarse-grained model constructed using low-frequency surface seismic wave inversion data or historical geological patterns from regional exploration. The construction principle of this model is physical conservatism, retaining only low-frequency, large-scale basic stratigraphic trends such as average dip angle and main structural strike, while eliminating all high-frequency micro-texture details. This design aims to provide a physically absolutely credible background to highlight and verify whether the fine textures generated by AI conform to basic physical logic. The stress-induced artifact feature library is established based on the physical characteristic testing of sensor hardware and the analysis of historical failure data. The stress-induced artifacts included in the library specifically refer to the specific interference patterns output by logging-while-drilling instruments when the detection crystal undergoes elastic deformation or the circuit generates thermal noise under extreme conditions of high temperature and high pressure at the bottom of the well. These patterns appear as periodic signals with a fixed spatial frequency in the frequency domain and as regular stripes with a consistent direction in the spatial domain, showing significant statistical differences from the randomness and non-uniformity of sedimentary rock bedding in nature.

[0019] Example 4 Activate the mechanical baseline mode to deprive the trajectory planning algorithm of control, including: Stop executing image feature-based tilting or azimuth-shifting commands; Lock the current wellbore trajectory tangent direction; Perform pure geometric steady-slope drilling operations to guide the drill bit to drill through the interference area; The mechanical baseline mode is defined as an inertial navigation maintenance strategy under non-line-of-sight conditions. When the dual-modal arbitration control unit triggers this mode, the system executes a strict control handover procedure: immediately shielding all control signals originating from the image recognition module and ceasing to execute any dynamic adjustment commands that attempt to adjust the skew or azimuth based on the stratigraphic boundaries. The system locks the current wellbore trajectory tangent vector and switches to pure geometric stable drilling logic. In this state, the drilling control system degenerates into an attitude maintenance system based on gravity tool face and magnetic tool face. It no longer responds to changes in geological parameters, but strictly maintains the current well inclination angle and azimuth angle, guiding the drill bit through the signal interference area in a straight line or smooth arc geometric trajectory.

[0020] Example 5 The method also includes: During the operation of the mechanical baseline mode, the changing trend of geological illusion entropy is monitored in real time; Once the geological illusion entropy is detected to have fallen back to a safe range and remained within a preset time window, the data quality is deemed to have recovered. Automatically unlock the Smart Navigation Mode; Re-allowing artificial intelligence algorithms to intervene and fine-tune the trajectory; The system is configured with closed-loop recovery logic with hysteresis characteristics to solve the control stability problem under complex working conditions. Although the system has entered the mechanical baseline mode and locked the control, the background calculation process continues to monitor the real-time evolution of geological illusion entropy. As the drilling depth increases, when the drill bit physically moves away from the stress anomaly zone that causes sensor deformation, the artifact features in the imaging data will gradually fade, and the geological illusion entropy will show a downward trend. To prevent frequent mode switching caused by critical signal fluctuations, the system has set strict recovery criteria: only when the geological illusion entropy not only falls back to the safe range, but also maintains this state for a preset time window, will the system confirm that the data environment has substantially improved; after confirmation, the arbitration unit automatically releases the logic lockout on the intelligent navigation mode, allowing the artificial intelligence algorithm to reconnect to the control loop and make fine-tuning of the trajectory based on the restored clear geological images.

[0021] Example 6 The downhole directional tool is driven by an execution and feedback correction unit based on the activated mode, including: When a command for intelligent navigation mode is received, the push plate or directional motor is driven to track the reservoir. When a command for mechanical baseline mode is received, the control parameters are locked. Reduce the response sensitivity gain of the downhole guiding tool actuator; Filter out high-frequency oscillation commands caused by residual image noise; The execution and feedback correction unit is not only responsible for the physical execution of instructions, but also dynamically adjusts the response characteristics of the underlying servo system according to the currently activated control mode. When the system is in intelligent navigation mode, the unit is set to a high-gain state and drives the hydraulic push plate or directional motor to perform minute attitude adjustments with a millisecond-level response speed to achieve precise tracking of thin reservoirs. Once switched to mechanical baseline mode, in addition to locking key control parameters such as the tool face angle, this unit will also actively adjust the actuator control loop. Gain coefficient; this hardware-level desensitization operation aims to create a low-pass filter effect, directly filtering out high-frequency oscillation commands that may be caused by residual image noise or sensor tremors from the physical execution end.

[0022] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An image processing-based adaptive trajectory control method for rotary steered drilling, characterized in that, include: The raw logging-while-drilling image data frames are received from the downhole transmission channel through the low-quality sensing reconstruction unit. The original logging-while-drilling image data frames are processed by running a deep learning image super-resolution model to generate a high-resolution predicted geological texture map; Simultaneously generate a feature uncertainty distribution map that characterizes the pixel probability distribution features in the high-resolution predicted geological texture map; The matching degree between the high-resolution predicted geological texture map and the preset stress-induced artifact feature library is calculated by the artifact feature identification unit. When the matching degree exceeds a preset benchmark, a sensor contamination marker signal is generated; The geological illusion entropy calculation unit calls the safety baseline geological model constructed before drilling. Calculate the difference in geological texture structure between the high-resolution predicted geological texture map and the safe baseline geological model; Data from the feature uncertainty distribution map is extracted, and the geological texture structure difference values ​​are weighted to generate geological illusion entropy. The geological illusion entropy and the sensor contamination marker signal are received by the dual-modal arbitration control unit; If the geological illusion entropy is lower than a preset threshold and no contamination marker signal is received from the sensor, the intelligent navigation mode is activated, and the trajectory planning algorithm is given control authority. If the geological illusion entropy exceeds the preset cognitive collapse threshold, or if the sensor contamination marker signal is received, the mechanical baseline mode is activated, and the trajectory planning algorithm is deprived of its control authority. The downhole directional tool is driven by the execution and feedback correction unit according to the activated mode.

2. The image processing-based adaptive trajectory control method for rotary steered drilling according to claim 1, characterized in that, Data from the feature uncertainty distribution map is extracted, and the geological texture structure difference values ​​are weighted to generate a geological illusion entropy, including: Perform modal alignment operation to project and rasterize the safe baseline geological model into a baseline trend field; The baseline trend field has the same spatial resolution as the high-resolution predicted geological texture map; Calculate the grayscale deviation or gradient deviation between the high-resolution predicted geological texture map and the baseline trend field pixel by pixel; For regions where the grayscale deviation or the gradient deviation exists, extract the corresponding feature uncertainty value of the region in the feature uncertainty distribution map; If the value of the uncertainty of the feature is higher than the preset high confidence threshold, the weight of the deviation value of that region in the total entropy calculation is increased exponentially; If the value of the uncertainty of the feature is lower than the preset low confidence threshold, the weight of the deviation value of that region in the total entropy calculation is suppressed; The weighted deviation values ​​of the entire image are aggregated to obtain the geological illusion entropy.

3. The image processing-based adaptive trajectory control method for rotary steered drilling according to claim 1, characterized in that, The safety baseline geological model is constructed based on low-frequency seismic data or regional geological patterns. The safety baseline geological model only includes the basic stratigraphic trend of the average dip angle of the strata; The stress-induced artifact feature library contains regular stripe signal features generated by the sensor under pressure deformation; The regular stripe signal features have specific spatial frequency and directionality.

4. The image processing-based adaptive trajectory control method for rotary steered drilling according to claim 1, characterized in that, Activating the mechanical baseline mode and depriving the trajectory planning algorithm of control includes: Stop executing image feature-based tilting or azimuth-shifting commands; Lock the current wellbore trajectory tangent direction; Perform pure geometric steady-slope drilling operations to guide the drill bit through the interference area.

5. The image processing-based adaptive trajectory control method for rotary steered drilling according to claim 4, characterized in that, The method further includes: During the operation of the mechanical baseline mode, the changing trend of the geological illusion entropy is monitored in real time; Once the geological illusion entropy is detected to have fallen back to a safe range and remained within a preset time window, it is determined that the data quality has been restored. Automatically unlock the intelligent navigation mode; Artificial intelligence algorithms are now allowed to intervene and fine-tune the trajectory.

6. The image processing-based adaptive trajectory control method for rotary steered drilling according to claim 1, characterized in that, The downhole directional tool is driven by an execution and feedback correction unit based on the activated mode, including: When the intelligent navigation mode command is received, the push plate or directional motor is driven to track the reservoir. When the instruction for the mechanical baseline mode is received, the control parameters are locked; Reduce the response sensitivity gain of the downhole guiding tool actuator; Filter out high-frequency oscillation commands caused by residual image noise.

Images