Digital dial-based underground pipeline cctv camera view angle gimbaling precision control method and device

Abstract

The application relates to the technical field of underground pipeline CCTV detection control, and discloses a kind of underground pipeline CCTV camera view angle universal precision control method and equipment based on digital dial, which comprises the following steps: first, obtaining the real-time image inside the pipeline collected by the camera, extracting the image features through the pre-trained first deep learning model, and calculating the view angle deviation between the camera view angle and the target shooting area; the deviation is decomposed into horizontal and vertical rotation angle correction amount, and the corresponding digital dial assembly is controlled to drive the camera to rotate; the above steps are repeated for iterative control until the preset alignment condition is met; finally, the camera is controlled to perform focusing operation and shoot clear image. The application realizes the automatic and precise control of the camera view angle, reduces the dependence on manual operation, improves the detection efficiency and imaging quality, and is suitable for precise detection of underground pipeline disease points.

Description

Technical Field

[0001] This application relates to the field of underground pipeline CCTV detection and control technology, specifically a method and device for omnidirectional precision control of the viewing angle of an underground pipeline CCTV camera based on a digital dial. Background Technology

[0002] As a crucial component of urban infrastructure, underground pipelines directly impact the normal functioning of the city, making regular inspection and maintenance essential. CCTV (Closed-Circuit Television) inspection technology, due to its intuitive and efficient characteristics, has become the mainstream method for detecting underground pipeline defects. By capturing images of the pipeline's interior using cameras, it enables visual inspection of defects such as cracks, corrosion, and deformation.

[0003] However, in the existing CCTV inspection process for underground pipelines, there are still significant shortcomings in the camera angle adjustment. When operators discover suspicious defects in the images, they need to rely on their eyes to judge whether the camera rotation angle is correct and manually control the camera to aim at the defect. This method not only depends on the operator's experience and concentration, but also suffers from low angle control precision and poor aiming efficiency, which can easily lead to missed defects or blurry images.

[0004] Meanwhile, existing technologies lack automated and precise control mechanisms for camera perspectives. Even if some devices have basic auxiliary control functions, it is difficult to achieve real-time tracking and stable alignment of defect points. In addition, the complex internal environment of underground pipelines, with low lighting and narrow spaces, further increases the difficulty of manual adjustment, resulting in low inspection efficiency and high labor intensity, which cannot meet the needs of large-scale, high-precision pipeline inspection.

[0005] Therefore, how to achieve automated and precise control of the CCTV camera's field of view for underground pipelines, reduce reliance on manual labor, and improve the accuracy and efficiency of defect detection has become a pressing technical problem to be solved in the field of underground pipeline inspection. Summary of the Invention

[0006] The purpose of this application is to provide a method and device for omnidirectional precision control of the viewing angle of an underground pipeline CCTV camera based on a digital dial, so as to solve the problems mentioned in the background art.

[0007] According to the first aspect of this application, a method for omnidirectional precision control of the viewing angle of an underground pipeline CCTV camera based on a digital dial is provided, comprising the following steps: Acquire real-time images of the interior of underground pipelines captured by cameras; Based on the real-time images, image features are extracted using a pre-trained first deep learning model, and the angle deviation between the current camera view and the target shooting area is calculated. The aforementioned viewing angle deviation is decomposed into a lateral rotation angle correction and a longitudinal rotation angle correction. Based on the horizontal rotation angle correction amount, the horizontal rotation digital dial component 5 is controlled to drive the camera to perform horizontal rotation; Based on the longitudinal rotation angle correction amount, the longitudinal rotation digital dial component 6 is controlled to drive the camera to perform longitudinal rotation; After performing horizontal and vertical rotations, the aforementioned steps are repeated for iterative control until the camera's field of view meets the preset alignment conditions. When the alignment conditions are met, the camera is controlled to perform a focusing operation to capture a clear image.

[0008] Preferably, calculating the viewpoint deviation based on the real-time image using a deep learning model further includes: The real-time image is preprocessed, including at least one of scale normalization, noise suppression, and brightness correction; The preprocessed image is input into the pre-trained deep learning model to obtain the saliency distribution information of the target shooting area and the offset features relative to the image center; The viewpoint deviation is generated based on the offset feature.

[0009] Preferably, the preset alignment condition is: based on the output of the first deep learning model, it is determined that the saliency concentration degree and feature stability score of the target shooting area in the image reach a preset threshold.

[0010] Preferably, controlling the camera to perform a focusing operation includes: controlling the camera to acquire multiple frames of images at different focal lengths; inputting the multiple frames of images into a pre-trained second deep learning model to obtain a sharpness score of the target shooting area at each focal length; and selecting the focal length that optimizes the sharpness score as the final focusing parameter.

[0011] Preferably, before acquiring real-time images, the method further includes: controlling the hydraulic lifting assembly 4 to adjust the height of the camera in the vertical direction according to the pipeline conditions.

[0012] A second aspect of this application also provides a omnidirectional precision control device for the viewing angle of an underground pipeline CCTV camera based on a digital dial, used to perform the aforementioned method, characterized in that it includes: Walking chassis 1, used for moving within underground pipelines; Camera component 2 is used to capture real-time images of the interior of underground pipelines; A camera omnidirectional adjustment assembly 3, mounted on the walking chassis 1, includes: The hydraulic lifting assembly 4 is used to adjust the height of the camera assembly 2 in the vertical direction; The horizontal rotation digital dial assembly 5 is used to drive the camera assembly 2 to perform horizontal rotation; The vertical rotation digital dial assembly 6 is used to drive the camera assembly 2 to perform vertical rotation; The control processing unit is electrically connected to the camera assembly 2, the hydraulic lifting assembly 4, the horizontal rotating digital dial assembly 5, and the vertical rotating digital dial assembly 6. The control processing unit is configured as follows: Receive real-time images captured by the camera component 2; Based on the real-time images, the viewing angle deviation between the current camera viewpoint and the target shooting area is calculated using a pre-trained deep learning model. The aforementioned viewing angle deviation is decomposed into a lateral rotation angle correction and a longitudinal rotation angle correction. The horizontal rotating digital dial assembly 5 and the vertical rotating digital dial assembly 6 are controlled to perform corresponding rotations respectively; The above steps are repeated iteratively until the camera's field of view meets the preset alignment conditions; When the alignment conditions are met, the camera assembly 2 is controlled to perform a focusing operation.

[0013] Preferably, the horizontal rotating digital dial assembly 5 includes a horizontal rotating drive motor, a rotating transmission mechanism, and a digital angle acquisition module; The longitudinal rotating digital dial assembly 6 includes a longitudinal rotating drive motor, a rotating transmission mechanism, and a digital angle acquisition module; The digital angle acquisition module is used to acquire the rotation angle in real time and feed it back to the control processing unit.

[0014] Preferably, it also includes a supplementary lighting assembly, installed around the camera assembly 2, for providing auxiliary lighting.

[0015] Preferably, the control processing unit is an embedded controller or an industrial control computer, including a processor module, a storage module, an actuator control interface, and an image data interface.

[0016] Preferably, the walking chassis 1 is provided with a mounting base and a carrying handle for mounting the camera universal adjustment assembly 3.

[0017] This application's equipment achieves stable movement within pipelines via a mobile chassis 1. The hydraulic lifting component 4 and the horizontal and vertical rotating digital dial components 6 of the camera's omnidirectional adjustment assembly 3 constitute a precise omnidirectional adjustment structure. Combined with supplementary lighting components, a control processing unit, and data and power cables, it forms an integrated detection system. Relying on a convolutional neural network model incorporating an attention mechanism, it achieves automatic image feature extraction and viewpoint deviation calculation. Through iterative closed-loop control and adaptive focusing, it realizes intelligent and precise control of the camera's viewpoint. This significantly reduces reliance on manual operation, improves detection efficiency, reduces labor intensity, and simultaneously enhances the clarity and accuracy of images of defect points, providing reliable technical support for underground pipeline defect detection. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0019] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a schematic diagram of a omnidirectional precision control device for the viewing angle of an underground pipeline CCTV camera based on a digital dial, provided as an embodiment of this application.

[0020] Figure 2 This is a schematic flowchart illustrating a method for omnidirectional precision control of the viewing angle of an underground pipeline CCTV camera based on a digital dial, provided as an embodiment of this application.

[0021] In the diagram: 1. Walking chassis; 2. Camera assembly; 3. Camera universal adjustment assembly; 4. Hydraulic lifting assembly; 5. Horizontal rotating digital dial assembly; 6. Vertical rotating digital dial assembly. Detailed Implementation

[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0023] It should be noted that all user information (including but not limited to user device information, user personal information, object information corresponding to device usage data, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, device usage data, etc.) involved in all embodiments of this application are information and data authorized by the user or fully authorized by all parties.

[0024] The technical solution disclosed herein is applicable to the internal inspection of underground pipelines, especially to the need for accurate imaging of pipeline defects in different diameters and low-light environments. It typically relies on tracked mobile inspection equipment to move stably inside the pipeline, and the control processing unit maintains reliable electrical connection and signal transmission with each execution component. At the same time, the pre-trained deep learning model has been stored in the storage module of the control processing unit, which has the hardware conditions to process image data and generate control commands in real time.

[0025] The following detailed description, in conjunction with specific embodiments, illustrates the implementation process of the omnidirectional precision control method for the viewing angle of an underground pipeline CCTV camera based on a digital dial, as described in this application. It should be noted that this embodiment is merely for explaining this application and not for limiting the scope of protection of this application. Any conventional adjustments or substitutions made by those skilled in the art to the various steps without departing from the concept of this application should be included within the scope of protection of this application.

[0026] like Figure 1 The diagram shown is a schematic representation of a omnidirectional precision control device for the viewing angle of an underground pipeline CCTV camera based on a digital dial, provided in this embodiment of the application. The overall structure of the device and the details of each component are described below: The traveling chassis 1 can adopt a tracked or wheeled structure design. This structure provides stable and reliable support for the movement of the equipment within underground pipelines. The track or tire surface is made of non-slip and wear-resistant material, which can adapt to the rough and irregular pipe wall environment inside the pipeline, ensuring that the equipment can move smoothly within pipelines of different diameters. The frame of the traveling chassis 1 is made of high-strength lightweight alloy material, which effectively reduces the overall weight of the equipment while ensuring structural strength, facilitating the handling, lowering into the well, and recovery of the equipment.

[0027] In one embodiment, the mobile chassis 1 is also equipped with a mounting base and a carrying handle 2. The mounting base and the chassis frame are rigidly connected by welding or high-strength bolts. The connection parts are reinforced to withstand the weight of components such as the camera universal adjustment component 3 and the hydraulic lifting component 4, as well as vibrations during operation, providing solid mechanical support for the stable operation of subsequent components. The carrying handle 2 is symmetrically arranged on both sides of the chassis and adopts an anti-slip design. The connection strength between the handle and the chassis frame has been rigorously tested to ensure that it will not break or loosen during equipment transportation, making it convenient for operators to move equipment on the ground, send the equipment into underground pipelines, and retrieve the equipment after the operation is completed.

[0028] The chassis 1 integrates a drive motor and transmission system. The drive motor receives travel commands from the control processing unit and drives the tracks to rotate via the transmission system, enabling the equipment to move forward, backward, and turn along the pipeline axis. The power of the drive motor can be configured according to the overall weight of the equipment and the pipeline's resistance to ensure sufficient driving force. The transmission system uses gear or chain drive, offering high transmission efficiency and stability. It accurately transmits the motor's power to the tracks and also has a certain speed reduction and torque increase function, making the equipment's travel speed within the pipeline stable and controllable. For example, the travel speed can be adjusted to a suitable low-speed range for inspection operations to avoid missing defects due to excessive speed.

[0029] Camera Universal Adjustment Component 3. The camera universal adjustment component 3 is a hardware structure that enables precise adjustment of the camera's viewing angle. It is installed on the mounting base of the walking chassis 1. The whole adopts a modular design, which is convenient for assembly, disassembly and maintenance. It includes a hydraulic lifting component 4, a horizontal rotating digital dial component 5, a vertical rotating digital dial component 6 and a camera mounting bracket. The components work together through mechanical connections to form a complete universal adjustment system.

[0030] Hydraulic lifting assembly 4. Hydraulic lifting assembly 4 adopts a hydraulic rod structure. The hydraulic rod is selected based on mechanical calculations to ensure sufficient load-bearing capacity and telescopic stroke to meet the camera height adjustment requirements under different working conditions. The lower end of the hydraulic rod is fixed to the mounting base of the walking chassis 1 via a flange or special connector. This fixing method ensures that the hydraulic rod will not experience radial displacement or loosening during telescopic movement. The upper end is rigidly connected to the support frame of the universal adjustment assembly. The support frame is welded from square or round steel pipes, providing a stable structure and is used to install the horizontal rotating digital dial assembly 5 and subsequent related components.

[0031] The hydraulic lifting assembly 4 is equipped with a hydraulic drive unit and a displacement detection sensor. The hydraulic drive unit is electrically connected to the actuator control interface of the control processing unit, receives the extension and retraction control commands issued by the control processing unit, and drives the hydraulic rod to extend and retract by controlling the inflow and outflow of hydraulic oil. The displacement detection sensor is built into or externally mounted on the hydraulic rod, which can collect the extension and retraction displacement of the hydraulic rod in real time and convert the displacement signal into a digital signal to feed back to the control processing unit. The control processing unit compares the feedback displacement information with the preset target height, and adjusts the working state of the hydraulic drive unit through closed-loop control to ensure that the extension and retraction of the hydraulic rod accurately meets the requirements, thereby realizing high-precision lifting and lowering adjustment of the camera in the vertical direction.

[0032] Horizontal Rotation Digital Dial Assembly 5. The horizontal rotation digital dial assembly 5 is located at the output end of the hydraulic lifting assembly 4, between the support frame and the camera mounting bracket. Its overall structure is compact, occupying little space and adaptable to the internal installation layout of the equipment. This assembly includes a horizontal rotation drive motor, a rotation transmission mechanism, a digital angle acquisition module, a fixed base, and a rotatable output end. Precision assembly of each component ensures operational coordination and reliability.

[0033] The horizontal rotation drive motor is a high-precision stepper motor or servo motor. These motors feature high control precision, fast response, and stable operation, meeting the precise control requirements for the camera's horizontal rotation. The motor's output power and torque are matched to the weight of the camera and related components to ensure sufficient driving force to rotate the camera. The rotary transmission mechanism employs either gear reduction or worm gear transmission. The gear transmission mechanism uses precision-machined gears with smooth tooth surfaces, high meshing accuracy, and high transmission efficiency. The worm gear transmission has a self-locking function, maintaining the current rotation angle when the motor stops working, preventing angle deviation due to external forces. The two transmission methods can be selected based on the actual application scenario.

[0034] The digital angle acquisition module employs high-precision encoders such as photoelectric encoders and magnetic encoders. The encoder resolution can be configured to a high value, for example, outputting thousands or even tens of thousands of pulses per revolution, ensuring accurate acquisition of rotation angle information. It is installed on the output shaft or rotatable output end of the rotary transmission mechanism, detecting changes in the rotation angle in real time and transmitting the angle data as a digital signal to the control processing unit via signal lines. This provides real-time angle feedback to the control processing unit, enabling closed-loop control of the rotation angle. The fixed base is fixedly connected to the support frame of the hydraulic lifting assembly 4, and the rotatable output end is connected to the camera mounting bracket. Driven by the drive motor, the rotatable output end rotates continuously 360 degrees relative to the fixed base through the rotary transmission mechanism, meeting the adjustment needs of different lateral viewing angles.

[0035] Vertical Rotating Digital Dial Assembly 6. The vertical rotating digital dial assembly 6 is located between the camera mounting bracket and the camera body. Its structure is basically the same as that of the horizontal rotating digital dial assembly 5. It also includes a vertical rotation drive motor, a rotation transmission mechanism, a digital angle acquisition module, a fixed base, and a rotatable output end. The selection criteria for each component are the same as those for the horizontal rotating digital dial assembly 5 to ensure that the control accuracy and working stability of the two are consistent.

[0036] The core feature of this component is that its rotation axis is perpendicular to the rotation axis of the horizontal rotating digital dial component 5. The horizontal rotation axis is perpendicular to the pipeline axis, and the vertical rotation axis is parallel to the pipeline axis. This vertical layout allows the camera to achieve dual-degree-of-freedom omnidirectional rotation, that is, arbitrary angle adjustment in both the horizontal (horizontal) and vertical (vertical) directions, thereby covering the entire field of view inside the pipeline and ensuring accurate targeting of any defect point. Its working principle is the same as that of the horizontal rotating digital dial component 5. Under the control of the control processing unit, the vertical rotation drive motor drives the rotatable output end to rotate through the rotation transmission mechanism. The digital angle acquisition module collects the rotation angle in real time and feeds it back to the control processing unit, realizing precise control of the vertical rotation angle.

[0037] Camera assembly 2 and supplementary lighting assembly. Camera assembly 2 includes a camera body and a camera lens. The camera body is an industrial-grade high-definition camera with low-light shooting capabilities, enabling clear image data acquisition in low-light environments such as underground pipelines. Its resolution can be configured to high-definition or ultra-high-definition levels to ensure the capture of subtle features of defects. The camera body is fixedly mounted on the rotatable output end of the vertically rotating digital dial assembly 6. The mounting method uses bolts or special clamps to ensure a secure installation, preventing loosening or displacement during equipment movement and rotation, thus ensuring shooting stability.

[0038] The camera lens uses a zoom lens, with the focal length adjustment range selectable according to the shooting distance requirements of pipeline inspection. It can achieve close-up shooting of nearby defects and panoramic shooting of the internal environment of pipelines at a distance. The lens surface is equipped with an anti-fog and anti-dust coating, which can effectively prevent moisture and dust from the pipeline from adhering to the lens surface and affecting image quality. Camera component 2 has a built-in image sensor and data transmission module. The image sensor converts optical images into electrical signals, and the data transmission module transmits the electrical signals to the image data interface of the control processing unit through a data and power cable. The transmission process adopts a high-speed transmission protocol to ensure the real-time performance and integrity of the image data.

[0039] The supplementary lighting assembly is installed around the camera assembly 2, using a ring layout or a uniformly distributed multi-point layout to ensure that the supplementary lighting range completely covers the camera's field of view and avoids blind spots. The supplementary lights use LED chips, which are characterized by low energy consumption, high brightness, long lifespan, and fast response speed. Their brightness can be adjusted via a control processing unit, for example, automatically adjusting the supplementary light intensity based on the brightness of the image captured by the camera, or fixing the supplementary light brightness according to a preset lighting mode.

[0040] The fill light assembly features a waterproof and dustproof casing, enabling it to withstand the harsh environments of underground pipelines, including dampness and dust, and protecting the internal LED chips and circuitry from damage. The fill light is powered by the equipment's power supply system, transmitting power via data and power cables. Its operation is synchronized with camera assembly 2; that is, the fill light turns on synchronously when the camera starts shooting and turns off synchronously when the camera stops shooting, ensuring efficient energy utilization.

[0041] Control Processing Unit. The control processing unit is the control component of the equipment. It can be an embedded controller or an industrial control computer, possessing strong computing power, stable operating performance, and good scalability. It can meet the needs of complex tasks such as real-time image processing, control command generation, and coordinated control of various components. Its physical installation location can be selected to be set in a dedicated mounting cavity inside the chassis 1. The mounting cavity has shock absorption, dustproof, and heat dissipation functions. The shock absorption design can reduce the impact of vibration on the control processing unit during equipment movement. The dustproof design prevents dust from entering the interior and affecting the operation of electronic components. The heat dissipation design uses heat sinks, cooling fans, or natural heat dissipation to dissipate the heat generated by the control processing unit during operation in a timely manner, ensuring that its operating temperature is within the normal range.

[0042] The hardware structure of the control processing unit includes a processor module, a storage module, an actuator control interface, an image data interface, and a power management module. The processor module uses a high-performance microprocessor or microcontroller with multiple cores and high clock speed, enabling parallel processing of multiple tasks, such as simultaneous image preprocessing, deep learning model inference, and control command calculation. The storage module includes main memory and secondary storage. Main memory is used for temporary storage of data and program running status during processing, while secondary storage is used to store pre-trained deep learning model parameters, device configuration parameters, and acquired image data. Secondary storage can be a solid-state drive or a large-capacity flash memory, offering advantages such as fast read / write speeds and strong shock resistance. The actuator control interface uses multiple digital or analog output interfaces, which are electrically connected to the hydraulic drive unit, horizontal rotation drive motor, vertical rotation drive motor, and supplementary lighting assembly of the hydraulic lifting component 4, respectively, to output precise control commands. The interface has overcurrent and overvoltage protection functions to ensure safe operation of the equipment. The image data interface uses a high-speed image acquisition interface to receive real-time image data transmitted by the camera component 2, supporting real-time transmission of high-resolution, high-frame-rate images. The power management module converts the voltage provided by the equipment power supply system into the operating voltage required by each component of the control processing unit, ensuring stable power supply to each component.

[0043] The logical structure of the control processing unit includes an image preprocessing module, a deep learning feature extraction and discrimination module, and a control decision generation module. Each module implements its corresponding function through software programs, and the modules transmit and interact with each other through a data bus to ensure the smooth execution of the control logic.

[0044] Data and power supply cables. The data and power supply cables are integrated, simultaneously handling data transmission and power supply. The cables are made of specialized materials that are wear-resistant, bend-resistant, and have strong anti-interference capabilities. They can withstand repeated bending during equipment lifting and rotation, preventing cable damage due to bending. They also possess excellent electromagnetic shielding performance, reducing interference from the complex electromagnetic environment inside the pipeline on data transmission, ensuring stable transmission of image data, control commands, and other signals, as well as a reliable power supply.

[0045] The cable is laid along the hydraulic rod and rotating structure. Critical areas are protected using cable drag chains or protective sleeves to prevent damage from friction or compression with other components. Waterproof and dustproof connectors are used at the cable connections to each component. These connectors have a locking function to ensure a secure connection and prevent moisture and dust from entering the connectors and affecting electrical performance. The cable length is designed based on the equipment's maximum lifting height and maximum rotation angle to ensure sufficient slack even when the equipment reaches its maximum operating limits, preventing excessive stretching or pulling.

[0046] Power Supply System. The power supply system provides electrical support for the operation of the entire equipment, including the battery pack and power management module. The battery pack uses high-capacity, high-safety lithium batteries. Lithium batteries have the characteristics of high energy density, long cycle life, and stable charge and discharge performance, which can meet the power requirements of the equipment for long-term operation. For example, a single charge can support the equipment to work continuously for several hours. The battery pack is equipped with a battery management system to monitor parameters such as battery voltage, current, and temperature, and has overcharge, over-discharge, over-temperature, and short-circuit protection functions to ensure the safety of the battery pack and extend battery life.

[0047] The power management module connects to the battery pack, converting the battery pack's output voltage into the operating voltage required by each component. For example, it provides a low-voltage DC voltage to the control processing unit and the corresponding operating voltage to the drive motor and supplementary lighting components. It also features voltage stabilization to ensure stable operation of each component in environments with minimal voltage fluctuations. The power supply system can also be equipped with a charging interface for convenient charging of the battery pack after equipment operation. The charging interface has a design to prevent mis-insertion, ensuring safety during the charging process.

[0048] In one embodiment, during equipment operation, all components work collaboratively under the unified scheduling of the control and processing unit to form a complete detection and control process. First, the walking chassis 1 moves axially along the pipeline driven by the drive motor, while the supplementary lighting component is turned on to provide auxiliary lighting for the camera component 2. The camera component 2 collects image data inside the pipeline in real time and transmits it to the control and processing unit via data and power cables. After receiving the image data, the control and processing unit processes it through the image preprocessing module and then inputs it into a deep learning model for feature extraction and discrimination, generating a viewing angle deviation and decomposing it into lateral and longitudinal rotation angle corrections. Subsequently, the control and processing unit sends rotation control commands to the lateral rotation digital dial component 5 and the longitudinal rotation digital dial component 6 through the actuator control interface, driving the camera to adjust its viewing angle. At the same time, the hydraulic lifting component 4 adjusts the camera height according to the pipeline conditions. After adjustment, the camera collects images again, and the control and processing unit repeats the above process, using iterative closed-loop control to ensure that the camera is accurately aligned with the defect point. Finally, the control and processing unit controls the camera component 2 to perform a focusing operation to capture clear images of the defect point.

[0049] Throughout the entire process, the digital angle acquisition module, displacement detection sensor, and other components provide real-time feedback on their working status data. The control processing unit dynamically adjusts control commands based on the feedback data to ensure the collaborative working accuracy of each component and achieve automated and precise detection of pipeline defects.

[0050] In one embodiment, based on the pipeline conditions, the control processing unit first sends a height adjustment command to the hydraulic lifting assembly 4 via the actuator control interface. The pipeline conditions include information such as pipeline diameter and internal obstacle distribution, which can be pre-input to the control processing unit by the operator or collected in real-time by auxiliary sensors mounted on the equipment. The lower end of the hydraulic rod of the hydraulic lifting assembly 4 is fixed to the mounting base of the traveling chassis 1, and the upper end is connected to the support frame of the camera universal adjustment assembly 3. Its extension and retraction are calculated and determined by the control processing unit based on the operating parameters. For example, for larger diameter pipelines, the hydraulic rod can be extended to raise the camera height, placing the camera in a suitable initial shooting position. The displacement state of the hydraulic rod is fed back to the control processing unit in real-time via built-in sensors, ensuring precise height adjustment. This process provides a stable initial posture for subsequent angle rotation adjustment and shortens the distance between the camera and the pipe wall when approaching a defect point, laying the foundation for clear imaging.

[0051] In one embodiment, after the camera height is adjusted, camera assembly 2 begins to acquire real-time images of the interior of the underground pipeline. Camera assembly 2 is fixedly mounted at the output end of a vertically rotating digital dial, with the front camera lens facing inwards. Its frame rate can be configured to adapt to the detection requirements, ensuring the capture of dynamic scenes and static defects within the pipeline. Since the interior of the underground pipeline is a low-light environment, supplementary lighting assemblies surrounding camera assembly 2 are activated simultaneously. The brightness of these supplementary lighting assemblies can be adaptively adjusted by the control processing unit based on the brightness of the images acquired by the camera, or preset to a fixed brightness level, to provide uniform and sufficient auxiliary lighting and avoid problems such as blurriness and excessive noise in the images due to insufficient light. The real-time images acquired by the camera are transmitted to the image data interface of the control processing unit via a data and power cable. The cable is laid along the hydraulic rod and rotating structure, maintaining a reliable connection during the lifting and rotation of the equipment to ensure that image data is transmitted without loss or delay.

[0052] In one embodiment, such as Figure 2 The diagram shows a flowchart illustrating a method for omnidirectional precision control of the viewing angle of an underground pipeline CCTV camera based on a digital dial, as provided in this application embodiment. Specifically, it includes: S1, acquire real-time images of the interior of underground pipelines captured by the camera; S2, based on the real-time image, extract image features through a pre-trained first deep learning model and calculate the viewpoint deviation between the current camera viewpoint and the target shooting area; S3, decompose the viewing angle deviation into a horizontal rotation angle correction and a vertical rotation angle correction; S4, based on the horizontal rotation angle correction amount, control the horizontal rotation digital dial component 5 to drive the camera to perform horizontal rotation; S5, based on the longitudinal rotation angle correction amount, control the longitudinal rotation digital dial component 6 to drive the camera to perform longitudinal rotation; S6. After performing horizontal and vertical rotation, repeat the aforementioned steps to perform iterative control until the camera's field of view meets the preset alignment conditions. S7. When the alignment conditions are met, control the camera to perform a focusing operation to capture a clear image.

[0053] Specifically, in step S1, after the camera height is adjusted, camera assembly 2 begins to acquire real-time images of the interior of the underground pipeline. Camera assembly 2 is fixedly installed at the output end of the vertically rotating digital dial, with the front camera lens facing inwards towards the pipeline. Its frame rate can be configured to adapt to the detection requirements, ensuring that dynamic scenes and static defects inside the pipeline can be captured. Since the interior of the underground pipeline is a low-light environment, supplementary lighting assemblies surrounding camera assembly 2 are activated simultaneously. The brightness of these supplementary lighting assemblies can be adaptively adjusted by the control processing unit based on the brightness of the images acquired by the camera, or preset to a fixed brightness level, to provide uniform and sufficient auxiliary lighting and avoid problems such as blurriness and excessive noise in the images due to insufficient light. The real-time images acquired by the camera are transmitted to the image data interface of the control processing unit via a data and power cable. The cable is laid along the hydraulic rod and rotating structure, maintaining a reliable connection during the lifting and rotation of the equipment to ensure that image data is transmitted without loss or delay.

[0054] In one embodiment, for step S2, the viewpoint deviation is calculated. After receiving the real-time image, the control processing unit first processes the image using its internal image preprocessing module. This processing ensures accurate feature extraction by the subsequent deep learning model. The core steps are briefly described as follows: Scale normalization adjusts the image to a preset fixed size, ensuring that images acquired at different times have a uniform input specification, facilitating feature extraction by the deep learning model. For example, the image can be normalized to a preset pixel size range. Noise suppression uses common filtering algorithms to remove random noise and environmental interference from the image, while retaining effective features of the diseased area. Brightness correction adjusts the brightness values ​​of the image pixels according to the illumination intensity of the supplementary lighting and the actual brightness distribution of the image, making the overall brightness of the image uniform and avoiding local over-brightness or under-brightness that could affect feature recognition.

[0055] The first deep learning model used in this disclosure is a convolutional neural network with an attention mechanism. This model is the core of the viewpoint deviation calculation. It solves the problems of low efficiency and poor accuracy of traditional manual judgment of viewpoint deviation. The model automatically extracts image features and calculates the deviation, thereby realizing the automation and precision of viewpoint adjustment.

[0056] The overall model structure includes an input layer, a feature extraction backbone network, an attention mechanism module, a feature fusion layer, and an output layer. The input layer receives a pre-processed image of a fixed size. The feature extraction backbone network employs a multi-layered structure of alternating convolutional and pooling layers. Convolutional layers perform local feature convolution operations with the image using convolution kernels. Pooling layers downsample the convolutional feature maps, preserving key features and reducing computational cost. For example, the convolution kernel size can be configured as a common 3×3 or 5×5. Pooling layers use max pooling or average pooling. Through multi-layer stacking, low-level features such as edges and textures are progressively extracted from the image, up to high-level features such as the overall outline of the diseased area and local details. The attention mechanism module is embedded in the middle and top layers of the feature extraction backbone network to strengthen the weights of diseased area features and suppress... To mitigate background interference, the model focuses more intently on the target area by calculating the attention weight of each pixel in the feature map. Specifically, the attention weight is calculated based on the feature response intensity and spatial relationship of the pixels. For example, pixels in the diseased area have a higher feature response intensity than those in the background area, and therefore receive a greater attention weight. The feature fusion layer fuses feature maps from different levels and channels, integrating low-level detail features with high-level semantic features to improve the completeness of feature representation. The output layer maps the fused features to the target output through a fully connected layer, namely, the saliency distribution information of the diseased area, the offset features relative to the image center, and the feature stability evaluation results.

[0057] In one embodiment, the training process of the first deep learning model is completed on a computer device outside the control processing unit, and the model parameters after training are imported into the storage module of the control processing unit. The training dataset uses an image dataset containing various types of pipeline defects such as cracks, corrosion, and deformation. The images in the dataset cover scenes with different pipeline diameters, different lighting conditions, and different degrees of defect severity, ensuring that the model has good generalization ability. During training, the dataset is first labeled, including the location, boundaries, and type of defect areas. Then, the labeled images are input into the initial model, and the model parameters are iteratively updated using the backpropagation algorithm. The loss function adopts a multi-task loss function, comprehensively considering the saliency distribution prediction loss, offset feature calculation loss, and stability evaluation loss. For example, the loss function can be expressed as: in, This is the total loss value. To predict loss based on significance distribution, Calculate the loss for the offset features. For stability evaluation loss, These are weighting coefficients, which can be adjusted to a reasonable range based on training performance to balance the training priorities of various tasks. Through multiple rounds of iterative training, training stops when the model's loss value on the validation set stabilizes and reaches a preset threshold, resulting in a pre-trained deep learning model.

[0058] In one embodiment, after the preprocessed image is input into the pre-trained first deep learning model, it is passed through the input layer to the feature extraction backbone network. Multi-scale feature maps are extracted through operations in each convolutional layer and pooling layer. The attention mechanism module processes the intermediate and top-level feature maps, calculates the attention weights of each pixel, and adds them to the corresponding feature values ​​to enhance the features of the diseased area. The feature fusion layer concatenates or adds the weighted feature maps of different scales and channels element-wise to form a comprehensive feature map. The output layer performs a fully connected operation on the comprehensive feature map, outputting results in three dimensions: the saliency distribution information of the diseased area in the current image, represented as a saliency heatmap, where the value of each pixel in the heatmap represents the probability that the pixel belongs to the diseased area; the offset feature of the diseased area relative to the image center, expressed as an offset vector (…). , )express, This represents the horizontal offset of the center of the diseased area relative to the image center. The vertical offset is the value of the offset. The stability evaluation result of the disease area features in continuous image frames is represented by a stability score. The higher the score, the more stable the disease features are in continuous frames and the less affected by motion blur, noise and other interference.

[0059] The control processing unit generates the viewpoint deviation based on the offset features output by the deep learning model. First, a mapping relationship is established between image pixel coordinates and the camera's viewpoint angle. This mapping relationship is pre-calibrated based on camera intrinsic parameters such as focal length, pixel size, and installation position parameters. For example, the camera's focal length is... The pixel size is The pixel offset in the horizontal direction of the image Corresponding view offset angle satisfy The same applies to the vertical direction. Based on the above mapping relationship, the offset vector ( , This is converted into the corresponding viewing angle offset angle, which is the viewing angle deviation between the current camera view and the target shooting area. It reflects the size and direction of the angle that the camera needs to adjust in order to achieve precise alignment of the diseased area.

[0060] In one embodiment, for step S3, the control processing unit decomposes the generated viewing angle deviation into a horizontal rotation angle correction and a vertical rotation angle correction. This decomposition process is based on the rotation axis directional characteristics of the horizontal and vertical rotation digital dials. The rotation axis of the horizontal rotation digital dial is perpendicular to the pipeline axis and is used to adjust the camera's viewing angle in the horizontal direction. The rotation axis of the vertical rotation digital dial is parallel to the pipeline axis and is used to adjust the camera's viewing angle in the vertical direction. The two rotation axes are perpendicular to each other, forming a two-degree-of-freedom omnidirectional rotation structure. The viewing angle deviation can be regarded as a spatial angle vector. During decomposition, according to the spatial geometric relationship, this vector is projected onto the plane corresponding to the horizontal and vertical rotation axes. The angle values ​​obtained by projection are the corresponding horizontal and vertical rotation angle corrections. For example, let the viewing angle deviation be... Its projection angle on the horizontal rotation plane is This is the horizontal rotation angle correction, and its projection angle on the vertical rotation plane is... This refers to the longitudinal rotation angle correction, which satisfies the geometric relationship of spatial angle vector decomposition. This ensures that the decomposed angle correction accurately reflects the angle that the camera needs to adjust in both directions, providing precise instructions for subsequent rotation control.

[0061] In one embodiment, for step S4, lateral rotation control. Specifically, the control processing unit sends a lateral rotation control command to the lateral rotation digital dial assembly 5 through the actuator control interface. The command includes the lateral rotation angle correction amount and rotation direction information. The lateral rotation digital dial assembly 5 includes a lateral rotation drive motor, a rotary transmission mechanism, and a digital angle acquisition module. After receiving the control command, the lateral rotation drive motor drives the rotary transmission mechanism to move according to the angle information in the command. The rotary transmission mechanism uses gear transmission or belt transmission to transmit the rotational motion of the motor to the rotatable output end, driving the camera mounting bracket and the camera to rotate laterally. During the rotation, the digital angle acquisition module acquires the rotation angle information in real time. This module can use angle detection elements such as encoders, and its acquisition frequency matches the motor rotation speed to ensure that the rotation angle change can be captured in real time. The acquired angle information is fed back to the control processing unit in the form of a digital signal. The control processing unit compares the real-time rotation angle with the preset lateral rotation angle correction amount, and adjusts the motor drive signal through closed-loop control to ensure that the lateral rotation angle accurately reaches the target value. For example, when the real-time rotation angle is less than the target angle, the motor continues to rotate; when the real-time rotation angle reaches the target angle, the motor drive stops, and the lateral rotation adjustment is completed.

[0062] In one embodiment, for step S5, the longitudinal rotation control process is similar to the lateral rotation control. The control processing unit sends a longitudinal rotation control command to the longitudinal rotation digital dial assembly 6, which includes the longitudinal rotation angle correction amount and the rotation direction. The structure of the longitudinal rotation digital dial assembly 6 is basically the same as that of the lateral rotation digital dial assembly 5, including a longitudinal rotation drive motor, a rotation transmission mechanism, and a digital angle acquisition module. Its rotation axis is perpendicular to that of the lateral rotation digital dial assembly 5. The longitudinal rotation drive motor drives the rotation transmission mechanism according to the control command, causing the camera body to rotate longitudinally. The digital angle acquisition module acquires the longitudinal rotation angle in real time and feeds it back to the control processing unit. The control processing unit achieves precise control of the longitudinal rotation angle through closed-loop control, ensuring that the camera's viewing angle is adjusted correctly in the vertical direction. The lateral and longitudinal rotation actions can be executed synchronously or in a preset order, ultimately achieving precise adjustment of the camera's viewing angle towards the diseased area. This solves the problems of traditional manual control of rotation angle relying on human judgment, low accuracy, and poor efficiency. Through the closed-loop control of the digital dial assembly, the control accuracy of the rotation angle is improved, laying the foundation for precise alignment of the diseased area.

[0063] In one embodiment, for step S6, after the camera completes one horizontal and vertical rotation, the control processing unit again acquires a new real-time image captured by the camera through the image data interface. This image is the internal image of the pipeline after rotation adjustment, reflecting the shooting effect after the camera's viewing angle adjustment. The control processing unit repeatedly performs image preprocessing, deep learning feature extraction, viewing angle deviation calculation, and angle correction decomposition steps on the newly acquired image to generate new horizontal rotation angle corrections and vertical rotation angle corrections. It then sends rotation control commands to the horizontal rotation digital dial component 5 and the vertical rotation digital dial component 6 again to execute new rotation adjustments. This process is repeated cyclically, forming an iterative closed-loop control process. Each iteration is based on the image feedback after the previous rotation, fine-tuning the camera's viewing angle to gradually reduce the viewing angle deviation, so that the camera's viewing angle gradually approaches the optimal shooting direction of the diseased area. The number of iterations can be preset to a fixed number, or dynamically adjusted according to the magnitude of the viewing angle deviation. For example, when the viewing angle deviation is less than a preset small threshold, the iteration can be terminated early to balance control precision and control efficiency. This iterative closed-loop control mechanism ensures the continuity and accuracy of camera viewing angle adjustment, effectively compensates for possible errors in a single adjustment, and achieves stable tracking and alignment of the diseased area.

[0064] In each iteration, the control processing unit uses the saliency concentration and feature stability scores of the diseased area output by the deep learning model to determine whether the camera's viewpoint meets the preset alignment conditions. This determination process provides an objective and quantitative basis for accurate alignment of the diseased area, avoiding the subjectivity and uncertainty of manual judgment. The saliency concentration of the diseased area is calculated through the distribution characteristics of the saliency heatmap. For example, it calculates the percentage of pixels with saliency values ​​higher than a preset threshold in the heatmap, or it calculates the variance of the saliency values. The higher the percentage and the smaller the variance, the more concentrated the saliency, indicating that the diseased area is in a major position in the image and has not shown significant shift. The feature stability score directly adopts the stability evaluation results output by the model. This score comprehensively considers factors such as the consistency and sharpness of the diseased area features in consecutive image frames. The preset thresholds include a saliency concentration threshold and a feature stability score threshold. These thresholds can be adjusted according to actual detection needs and different disease types. For example, for crack-type diseases, a higher saliency concentration threshold and feature stability score threshold can be set to ensure accurate alignment of small cracks. When the control processing unit determines that the salience concentration of the diseased area reaches the preset salience concentration threshold and the feature stability score reaches the preset feature stability score threshold, it determines that the camera viewpoint meets the alignment conditions and stops the iterative closed-loop control process; if the preset threshold is not reached, it continues to execute the next iteration adjustment.

[0065] In one embodiment, for step S7, after the camera's viewing angle meets the alignment conditions, the control processing unit controls the camera assembly 2 to perform a focusing operation to obtain a clear image of the diseased area. This focusing process is based on the sharpness evaluation of the second deep learning model, achieving adaptive optimization of the focusing parameters and solving the problem of image blurring that may occur with traditional fixed focal length or manual focusing. The control processing unit first sends a focusing control command to the camera assembly 2, controlling the camera's focal length to change within a preset range in a certain step. The preset focal length range is determined according to the camera's optical parameters and the shooting distance requirements of pipeline detection. The step size can be configured to an appropriate value to ensure that it can fully cover the possible optimal focal length range. At each focal length setting, the camera acquires one frame of image. After all the images corresponding to the focal length settings have been acquired, multiple frames of images are sequentially input into the second deep learning model. The second deep learning model extracts features from the diseased area in each frame of image and calculates the sharpness score of the diseased area. This score is based on indicators such as edge sharpness and detail richness of the diseased area features. For example, by calculating the gradient value of edge pixels in the feature map, the larger the gradient value, the sharper the edge and the higher the sharpness score. The control and processing unit collects sharpness scores at each focal length, compares and selects the focal length parameter with the highest sharpness score, uses this focal length parameter as the final focusing parameter, sends a focusing command to the camera assembly 2, controls the camera to adjust to this focal length, completes the focusing operation, ensures that the captured images of the diseased area are clear and detailed, and provides high-quality data support for subsequent disease analysis and assessment.

[0066] The second deep learning model employs a lightweight convolutional neural network architecture, suitable for real-time inference scenarios on embedded devices. Its overall structure includes an input layer, a feature extraction layer, a detail enhancement layer, and an output layer. The input layer receives images preprocessed by scale normalization, noise suppression, and brightness correction. The feature extraction layer consists of 3-5 alternating convolutional and pooling layers. The convolutional layers use small 3×3 kernels to reduce computation while preserving detailed features of the affected area. The pooling layers use max pooling with a stride of 2 to achieve feature dimensionality reduction and key information preservation. The detail enhancement layer embeds an edge perception module, which enhances the expression of sharpness-related features through targeted extraction of edge features of the affected area. This module combines edge detection logic based on the Sobel operator with convolutional operations to highlight the response intensity of edge pixels in the image. The output layer is a single-neuron fully connected layer that maps the extracted detail features to a sharpness score in the 0-1 range, with the score positively correlated with the sharpness of the affected area.

[0067] The training process includes using an image set of underground pipelines with different focal lengths and different types of defects (cracks, corrosion, deformation, etc.). The dataset covers various sharpness states, including out-of-focus, slightly out-of-focus, sharp, and super sharp, with a balanced number of image samples in each state. Before training, the images are labeled with their corresponding sharpness levels (quantized into a 0-1 range). The labeled images are then divided into training and validation sets in a 7:3 ratio and input into the initial model for training. The training process uses the mean squared error loss function and the Adam optimizer. The model parameters are iteratively updated using the backpropagation algorithm. During the iteration, the loss value of the validation set is monitored in real time. When the loss value stabilizes after several rounds and reaches a preset threshold, training stops, resulting in a trained second deep learning model. The model parameters are then imported into the storage module of the control processing unit for sharpness score calculation during the focusing process.

[0068] The disclosed equipment achieves stable movement within pipelines via a tracked chassis 1. The hydraulic lifting assembly 4 and the horizontal and vertical rotating digital dial assembly 6 of the camera's omnidirectional adjustment component 3 constitute a precise omnidirectional adjustment structure. Combined with supplementary lighting components, a control processing unit, and data and power cables, it forms an integrated detection system. Relying on a convolutional neural network model incorporating an attention mechanism, it achieves automatic image feature extraction and viewpoint deviation calculation. Through iterative closed-loop control and adaptive focusing, it realizes intelligent and precise control of the camera's viewpoint. This significantly reduces reliance on manual operation, improves detection efficiency, reduces labor intensity, and simultaneously enhances the clarity and accuracy of images of defect points, providing reliable technical support for underground pipeline defect detection.

[0069] It should be noted that although the operations of the method of this application are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the steps depicted in the flowchart can be performed in a different order. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0070] The above description represents the preferred embodiments of the present invention. It should be noted that, for those skilled in the art, various improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for precise omnidirectional control of the viewing angle of an underground pipeline CCTV camera based on a digital dial, characterized in that, Includes the following steps: Acquire real-time images of the interior of underground pipelines captured by cameras; Based on the real-time images, image features are extracted using a pre-trained first deep learning model, and the angle deviation between the current camera view and the target shooting area is calculated. The aforementioned viewing angle deviation is decomposed into a lateral rotation angle correction and a longitudinal rotation angle correction. Based on the horizontal rotation angle correction amount, the horizontal rotation digital dial component (5) is controlled to drive the camera to perform horizontal rotation; Based on the longitudinal rotation angle correction amount, the longitudinal rotation digital dial component (6) is controlled to drive the camera to perform longitudinal rotation; After performing horizontal and vertical rotations, the aforementioned steps are repeated for iterative control until the camera's field of view meets the preset alignment conditions. When the alignment conditions are met, the camera is controlled to perform a focusing operation to capture a clear image.

2. The method for omnidirectional precision control of the viewing angle of an underground pipeline CCTV camera based on a digital dial, as described in claim 1, is characterized in that... The calculation of the viewpoint deviation based on the real-time image using a deep learning model further includes: The real-time image is preprocessed, including at least one of scale normalization, noise suppression, and brightness correction; The preprocessed image is input into the pre-trained deep learning model to obtain the saliency distribution information of the target shooting area and the offset features relative to the image center; The viewpoint deviation is generated based on the offset feature.

3. The method for omnidirectional precision control of the viewing angle of an underground pipeline CCTV camera based on a digital dial, as described in claim 2, is characterized in that... The preset alignment condition is: based on the output of the first deep learning model, it is determined that the salience concentration of the target shooting area in the image and the feature stability score reach a preset threshold.

4. The method for precise omnidirectional control of the viewing angle of an underground pipeline CCTV camera based on a digital dial, as described in claim 1, is characterized in that... Controlling the camera to perform focusing operations includes: controlling the camera to acquire multiple frames of images at different focal lengths; inputting the multiple frames of images into a pre-trained second deep learning model to obtain the sharpness score of the target shooting area at each focal length; and selecting the focal length that optimizes the sharpness score as the final focusing parameter.

5. The method for precise omnidirectional control of the viewing angle of an underground pipeline CCTV camera based on a digital dial, as described in claim 1, is characterized in that... Before acquiring real-time images, the method also includes: controlling the hydraulic lifting assembly (4) to adjust the height of the camera in the vertical direction according to the pipeline conditions.

6. A precision control device for the field of view of an underground pipeline CCTV camera based on a digital dial, used to execute the method as described in any one of claims 1 to 5, characterized in that, include: A mobile chassis (1) is used for moving within underground pipelines; Camera component (2) is used to collect real-time images of the interior of underground pipelines; A camera omnidirectional adjustment assembly (3), mounted on the walking chassis (1), includes: A hydraulic lifting assembly (4) is used to adjust the height of the camera assembly (2) in the vertical direction; A horizontal rotation digital dial assembly (5) is used to drive the camera assembly (2) to perform horizontal rotation; A vertically rotating digital dial assembly (6) is used to drive the camera assembly (2) to perform vertical rotation; The control processing unit is electrically connected to the camera assembly (2), the hydraulic lifting assembly (4), the horizontal rotating digital dial assembly (5), and the vertical rotating digital dial assembly (6), and the control processing unit is configured as follows: Receive real-time images captured by the camera component (2); Based on the real-time images, the viewing angle deviation between the current camera viewpoint and the target shooting area is calculated using a pre-trained deep learning model. The aforementioned viewing angle deviation is decomposed into a lateral rotation angle correction and a longitudinal rotation angle correction. The horizontal rotating digital dial assembly (5) and the vertical rotating digital dial assembly (6) are controlled to perform corresponding rotations respectively; The above steps are repeated iteratively until the camera's field of view meets the preset alignment conditions; When the alignment conditions are met, the camera assembly (2) is controlled to perform a focusing operation.

7. The device according to claim 6, characterized in that, The horizontal rotation digital dial assembly (5) includes a horizontal rotation drive motor, a rotation transmission mechanism, and a digital angle acquisition module; The longitudinal rotating digital dial assembly (6) includes a longitudinal rotating drive motor, a rotating transmission mechanism, and a digital angle acquisition module; The digital angle acquisition module is used to acquire the rotation angle in real time and feed it back to the control processing unit.

8. The device according to claim 6, characterized in that, It also includes a fill light assembly, which is installed around the camera assembly (2) to provide auxiliary lighting.

9. The device according to claim 6, characterized in that, The control processing unit is an embedded controller or an industrial control computer, including a processor module, a storage module, an actuator control interface, and an image data interface.

10. The device according to claim 6, characterized in that, The chassis (1) is provided with a mounting base and a carrying handle for mounting the camera universal adjustment assembly (3).

Images