A magnetic navigation endoscopic detection method and device for small diameter pipes

Abstract

This invention discloses a magnetically controlled navigation pipe endoscopy inspection method and device for small-diameter pipes. The inspection method includes: placing a biomimetic multi-segment flexible camera module inside the pipe to be tested; acquiring an initial image of the inner wall of the pipe through the camera, identifying the current pipe material through a processing terminal, and adjusting the excitation output parameters of the magnetic controller; when the biomimetic multi-segment flexible camera module travels to a branch node of the pipe to be tested, acquiring a branch opening image to determine the target branch direction; adjusting the magnetic field polarity of the magnetic controller according to the target branch direction; during the travel, performing radial distortion remapping and tangential distortion remapping based on the refractive index parameters of the liquid medium through the processing terminal; mapping the corrected annular image into a distortion-free two-dimensional planar unfolded image reflecting the geometric morphology of the pipe inner wall; inputting the distortion-free two-dimensional planar unfolded image into a defect identification model built into the processing terminal to identify whether a preset defect type exists on the inner wall of the pipe to be tested.

Description

Technical Field

[0001] This invention relates to the field of small-diameter pipe inspection technology, specifically to a magnetically controlled navigation endoscopic inspection method and device for small-diameter pipes. Background Technology

[0002] Residential water supply pipes often use small-diameter designs, made of materials including plastic pipes such as PPR and PE, as well as metal pipes such as stainless steel and copper. These pipes have complex installation paths, with numerous bends and branches, and are often concealed within walls, floors, or ceilings. Over time, these pipes are prone to problems such as micro-cracks in the pipe walls and leaks at joints due to material aging, loosening of joint seals, or release of residual stress from construction. Because the pipes are concealed, initial leaks are difficult to detect. By the time obvious symptoms such as damp walls and bulging floors appear, the leak has already persisted for a considerable period, resulting not only in water waste and structural damage to the building, but also potential disputes with neighbors.

[0003] Existing testing methods for small-diameter water supply pipelines are mainly divided into two categories: non-invasive testing and invasive testing.

[0004] Non-invasive leak detection includes technologies such as acoustic leak detection, thermal imaging detection, tracer gas detection, and ground-penetrating radar (GPR). Acoustic leak detection is severely affected by noise interference in the home environment, and plastic pipes have weak sound transmission capabilities and rapid signal attenuation, making it difficult to accurately locate leaks. Thermal imaging detection only responds to large-scale leaks and is insensitive to early-stage defects such as micro-cracks and loose joints. Furthermore, it cannot penetrate walls to capture temperature changes due to the thickness of the decorative layer. Tracer gas detection requires draining water from the pipe, making it complex and time-consuming, and unsuitable for pipes already in use. Ground-penetrating radar has limited penetration capabilities in non-metallic pipes and is prone to signal reflection interference in metallic pipes; it can only detect the pipe's direction and cannot identify defects in the inner wall of the pipe.

[0005] Invasive inspection methods include wheeled pipe inspection robots, pressure testing, intelligent sensor monitoring, and traditional pipe endoscopy. Wheeled robots, limited by size, cannot enter pipes with diameters of 15mm to 25mm. Pressure testing can only determine the presence of leaks, not pinpoint their location. Intelligent sensors require drilling into the pipe wall for installation, disrupting pipe integrity and being difficult to implement in small-diameter pipes. Traditional pipe endoscopy offers the advantage of close-range image acquisition of the pipe's inner wall; however, industrial endoscopy equipment is often a rigid, monolithic structure, unable to maneuver flexibly at branch points such as tees and crosses in small-diameter household pipes; its wiring interfaces lack sealing designs, making operation in pressurized, full-water environments impossible; and defect identification relies on manual image reading, resulting in low efficiency and a high rate of missed diagnoses. Summary of the Invention

[0006] To overcome the shortcomings of the prior art, the present invention provides a magnetically controlled navigation endoscopic inspection method and device for small-diameter pipes, so as to solve the problems in the prior art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: A magnetically controlled navigation pipe endoscope inspection method for small-diameter pipes is applied to a magnetically controlled navigation pipe endoscope inspection device. The inspection device includes a biomimetic multi-segment flexible camera module composed of multiple spliced ​​segments connected in series by flexible connectors, a magnetic response component sleeved on the outer periphery of the spliced ​​segments, a magnetic controller set on the outside of the pipe to be tested, and a processing terminal. The detection method includes: The biomimetic multi-segment flexible camera module is placed inside the pipe to be tested, and the initial image of the inner wall of the pipe to be tested is obtained through the camera. The processing terminal identifies the current pipe material type based on the initial image and adjusts the excitation output parameters of the magnetron based on the identification result, so that the effective magnetic field strength acting on the magnetic response component after penetrating the pipe wall is maintained within a preset working range. The bionic multi-segment flexible camera module is propelled forward by a rigid wire connected to it. When it reaches a branch node of the pipe under test, the camera captures an image of the branch opening. The processing terminal determines the target branch direction based on the image of the branch opening and controls the magnetron to adjust the polarity of the magnetic field to guide the bionic multi-segment flexible camera module to bend toward the target branch pipe. During the process, the processing terminal performs coordinate mapping transformation on each pixel in the original annular image captured by the camera based on the refractive index parameter of the liquid medium in the pipe under test, and maps the corrected annular image into a distortion-free two-dimensional planar unfolded image that reflects the geometric shape of the inner wall of the pipe. The processing terminal inputs the distortion-free two-dimensional planar unfolded image into the defect identification model. When a defect is identified, it performs spatial coordinate conversion based on the current travel mileage data associated with the hard conductor's delivery length and the guide trajectory record of the magnetic controller to determine the physical location of the defect in the pipeline under test.

[0008] In one embodiment, the magnetic response component is sleeved on the outer periphery of at least one intermediate splicing segment in the biomimetic multi-segment flexible camera module, excluding the first splicing segment in which the camera is integrated, and a high magnetic permeability material layer is encapsulated inside the flexible connector between the magnetic response component and the first splicing segment. During the magnetic field polarity switching process, the high permeability material layer confines the magnetic field lines to the region where the magnetic response component is located, thereby suppressing the leakage of the magnetic field to the region where the camera is located.

[0009] In one embodiment, the outer peripheral surface of the magnetic response component is provided with a plurality of elastic limiting beads. When the biomimetic multi-segment flexible camera module travels in pipe segments with different inner diameters, the elastic limiting beads support the first splicing segment integrating the camera at the center position of the pipe under test by elastic contact with the inner wall of the pipe under test, so that the optical axis of the camera is basically coincident with the center line of the pipe under test. The contact friction between the elastic limiting bead and the inner wall of the pipe under test is configured to be less than the maximum magnetic traction force that the magnetron can apply through the magnetic response component.

[0010] In one embodiment, the processing terminal identifies the current pipe material type based on the initial image and adjusts the excitation output parameters, including: The processing terminal calls a pre-trained material classification model to identify the current pipe material type based on the texture features and reflective properties of the pipe's inner wall. The processing terminal retrieves and identifies the excitation compensation coefficient corresponding to the pipe material type from the preset mapping relationship between material and excitation compensation parameters. The processing terminal adjusts the base excitation current of the magnetron according to the excitation compensation coefficient.

[0011] In one embodiment, it further includes: When the biomimetic multi-segment flexible camera module enters the bend or branch node area of ​​the pipeline under test, the processing terminal monitors the magnetic field coupling state between the magnetron and the magnetic response component in real time. When the magnetic field coupling strength is detected to drop to a preset threshold, the processing terminal controls the magnetic controller to output an instantaneously enhanced auxiliary excitation pulse. The intensity of the auxiliary excitation pulse is higher than the steady-state excitation strength, so that the magnetic response component generates an instantaneous traction torque to overcome static friction resistance. After determining that the bionic multi-segment flexible camera module has completed the turning action based on the changes in the image captured by the camera, the processing terminal controls the magnetic controller to restore the excitation intensity to a steady-state level.

[0012] In one embodiment, the processing terminal performs a coordinate mapping transformation on each pixel in the original annular image captured by the camera based on the refractive index parameter of the liquid medium inside the pipe under test, including: Obtain the refractive index parameters of the liquid medium inside the pipe under test; Based on the refractive index parameter, the equivalent camera intrinsic parameter matrix of the camera in the underwater environment is recalibrated, wherein the equivalent focal length parameter is scaled proportionally according to the ratio of the refractive index of the medium to the refractive index of air. Using the recalibrated camera intrinsic parameter matrix, a coordinate mapping transformation is performed on each pixel in the original ring image to compensate for the refraction and deflection of light at the interface between the liquid medium and the camera optical element.

[0013] In one embodiment, before inputting the distortion-free two-dimensional planar unfolded image into the disease identification model, the method further includes: The processing terminal performs inter-frame difference analysis on multiple consecutive frames of distortion-free two-dimensional planar unfolded images; The processing terminal identifies candidate regions whose pixel positions remain relatively fixed across multiple frames and whose pixel change patterns differ from the background pixel displacement patterns based on the inter-frame difference results. The processing terminal performs weighted enhancement on the image features of the candidate region before inputting them into the disease identification model.

[0014] In one embodiment, determining the physical location of the defect in the pipe to be tested includes: Acquire the mileage signal associated with the length of the hard conductor pull-out; Based on the mileage signal, the processing terminal performs spatial registration and stitching on multiple consecutive frames of distortion-free two-dimensional planar unfolded images to generate a continuous panoramic detection image of the inner wall of the pipe under test. The processing terminal overlays the identified disease type and location information as visual markers onto the continuous panoramic detection image and triggers a location marker indication on the outside of the pipeline under test.

[0015] In one embodiment, the detection device further includes a quick-sealing connector installed at the inlet of the pipe to be tested, the quick-sealing connector having a flexible sealing body and an elastic compensating body inside; The detection method further includes: During the extraction of the rigid wire, the main sealing interface is formed by the sliding contact between the inner surface of the flexible seal and the outer peripheral surface of the rigid wire. The elastic compensator continuously applies a radially inward preload force to the flexible seal; When the flexible seal experiences material loss due to relative movement with the rigid wire, the elastic compensator releases elastic potential energy to push the flexible seal to contract inward, thereby compensating for the fit gap caused by material loss.

[0016] A magnetically controlled navigation endoscopic inspection device for small-diameter pipes, using the magnetically controlled navigation endoscopic inspection method for small-diameter pipes as described in any one of the above claims, characterized in that it includes: The biomimetic multi-segment flexible camera module is composed of multiple splicing segments connected in series by flexible connectors, with the camera and LED integrated in the first splicing segment. A magnetic response component is sleeved on the outer periphery of at least one splicing segment other than the first splicing segment, and is physically separated from the camera; the magnetic response component is embedded with at least one strong guiding magnet. A rigid wire, the proximal end of which is fixedly connected to the tail end of the biomimetic multi-segment flexible camera module; A quick-sealing wire connector is fitted around the outer circumference of the rigid wire and is used for installation at the inlet of the pipe to be tested. The magnetron is installed outside the pipe to be tested; The processing terminal is located outside the pipe to be tested and is communicatively connected to the camera.

[0017] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention's biomimetic multi-segment flexible camera module utilizes a combination of segmented splicing sections and flexible connectors to provide basic bending capability through bends with small curvature radii; the external magnetic control generation component provides selective lateral traction force by acting on the magnetic response component, and the two work together to achieve directional steering at multi-branch nodes in small-diameter pipes, solving the problem that traditional endoscopic equipment cannot selectively enter target branches in small-diameter pipes.

[0018] By processing the terminal to identify the pipe material and automatically adjusting the excitation output of the magnetic control generator according to the preset mapping relationship between the material and the excitation compensation parameters, the effective magnetic field strength after penetrating the pipe wall of different materials remains stable, avoiding the problem of guide failure caused by magnetic field attenuation in metal pipes and ensuring versatility in both plastic and metal pipes.

[0019] By introducing intrinsic parameter recalibration and pixel coordinate remapping based on the refractive index of the liquid medium in the image preprocessing step, image distortion caused by light refraction at the interface between water and optical elements is eliminated; the cylindrical unfolding module maps the annular fisheye image into a planar unfolded image, enabling the subsequent defect identification model to process images that reflect the true geometric shape of the pipe inner wall, significantly improving the identification accuracy.

[0020] By performing inter-frame difference analysis on multiple consecutive images, candidate regions with fixed positions and pixel change patterns that differ from the background motion pattern are identified. These candidate regions are then weighted to enhance their features before being input into the disease identification model. This effectively improves the detection confidence of static small targets such as microcracks against a dynamic moving background.

[0021] The dynamic sealing unit adopts a combination structure of flexible sealing body and elastic compensator. When the flexible sealing body is worn due to repeated pulling of the wire, the elastic compensator pushes the flexible sealing body to contract inward by releasing the pre-tightening force, automatically compensating for the wear gap, ensuring continuous sealing of the liquid medium inside the pipeline throughout the entire testing process, and preventing water leakage. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the workflow of a magnetically controlled navigation endoscopic inspection method for small-diameter pipes provided in an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the application status of a magnetically controlled navigation endoscopic inspection device for small-diameter pipes according to an embodiment of the present invention. Figure 3 This is a schematic diagram of the overall structure of a magnetically controlled navigation endoscopic detection device for small-diameter pipes provided in an embodiment of the present invention. Figure 4 This is a schematic diagram of a biomimetic multi-segment flexible camera module structure for a magnetically controlled navigation pipe endoscope detection device for small-diameter pipes, provided in an embodiment of the present invention. Figure 5 This is a schematic diagram of the magnetic response component structure of a magnetically controlled navigation pipe endoscope detection device for small-diameter pipes provided in an embodiment of the present invention. Figure 6 A schematic diagram of a quick-sealing wire connector structure for a magnetically controlled navigation pipe endoscope detection device for small-diameter pipes provided in an embodiment of the present invention; Figure 7 This is a schematic diagram of the axial structure of the quick-sealing wire connector of a magnetically controlled navigation pipe endoscope for small-diameter pipes, provided in an embodiment of the present invention.

[0023] In the figure: 1. Bionic multi-segment flexible camera module; 1-1. Camera; 1-2. LED bead; 2. Hard wire; 3. Ring-shaped substrate; 3-1. Guide magnet; 3-2. Limiting bead; 4. Quick-sealing guide joint; 4-1. Flexible sealing body; 5. Pipe under test; 6. Magnetizer. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] To enable those skilled in the art to fully understand the present invention, the following is combined with Figures 2 to 7 The device is described in detail.

[0026] See Figures 2 to 7An embodiment of the present invention provides a magnetically controlled navigation endoscopic inspection device for small-diameter pipes, comprising a biomimetic multi-segment flexible camera module 1 composed of multiple spliced ​​segments connected in series by flexible connectors, a magnetic response component sleeved on the outer periphery of the spliced ​​segments and separated from the camera inside the first spliced ​​segment, a magnetic controller 6 disposed outside the pipe 5 to be tested, and a processing terminal, wherein the biomimetic multi-segment flexible camera module 1 is connected to a rigid wire 2.

[0027] The biomimetic multi-segment flexible camera module 1 consists of multiple splicing segments connected in series via flexible connectors. In this embodiment, there are five splicing segments, all made of aluminum alloy and cylindrical in shape, with an outer diameter of 8mm and an axial length of 10mm per segment. The front end of the first splicing segment has a rounded chamfer to reduce frictional resistance when traveling within the pipe 5 under test. The first splicing segment integrates an image acquisition unit, including a 4K resolution miniature fisheye camera 1-1 and multiple LED fill light beads 1-2 evenly distributed around the camera. The optical axis of the miniature fisheye camera 1-1 coincides with the axis of the first splicing segment. The second to fifth splicing segments have basically the same structure as the first splicing segment, but do not have an image acquisition unit inside.

[0028] The various splicing segments are connected in series via flexible connectors. These flexible connectors are made of thermoplastic polyurethane (TPU) elastomer material with a hardness of 85A, and their ends are respectively embedded and fixed in the end holes of adjacent splicing segments. In its natural state, the flexible connector is straight, and when subjected to lateral external force, it can produce a directional deflection of at least a preset angle; after the external force is removed, it returns to its straight shape due to its own elastic restoring force. This structure solves the technical problem of traditional rigid probes easily getting stuck when passing through bends with small curvature radii, achieving the technical effect of enabling the camera module to smoothly travel along complex pipeline networks.

[0029] The magnetic response component is a separate part, installed on the outer periphery of the second splicing segment of the biomimetic multi-segment flexible camera module 1 using a rear-mounting method. The magnetic response component includes an annular substrate 3, made of food-grade silicone material, whose inner diameter is interference-fitted with the outer diameter of the second splicing segment. Multiple guiding strong magnets 3-1, made of neodymium iron boron permanent magnet material, are uniformly embedded in the inner sidewall of the annular substrate 3, with the magnetic pole direction of each guiding strong magnet 3-1 arranged radially.

[0030] It should be noted that the magnetic response component and the first splicing section supporting the camera 1-1 are physically separated. A magnetic shielding layer composed of a permalloy thin film of a predetermined thickness (e.g., 0.2 mm) is encapsulated inside this flexible connector. This avoids electromagnetic interference that might occur from external changing magnetic fields on the electronic shutter and color reproduction of the camera's CMOS image sensor, achieving the technical effect of confining magnetic field lines to the actuator area and ensuring image acquisition quality.

[0031] Furthermore, the outer circumferential surface of the annular substrate 3 is provided with multiple limiting beads 3-2. The limiting beads 3-2 are semi-circular protrusions made of silicone material, evenly distributed circumferentially. In the free state, the outer diameter of the limiting beads is slightly larger than the inner diameter of the pipe 5 under test; after entering the pipe 5, the limiting beads 3-2 undergo elastic deformation due to compression by the pipe wall. This solves the technical problem of the camera's optical axis easily deviating from the center of the pipe in non-horizontal or variable-diameter pipes, resulting in a limited field of view, and achieves the technical effect of elastically supporting the first-end splicing section at the center of the pipe, ensuring that the acquired image is a symmetrical panoramic view. Simultaneously, the contact friction between the limiting beads 3-2 and the pipe wall is designed to be less than the maximum magnetic traction force that the magnetic controller 6 can apply through the magnetic response component, thus ensuring that the turning action is unimpeded.

[0032] The near end of the rigid conductor 2 is fixedly connected to the tail splicing section of the biomimetic multi-segment flexible camera module 1 via a threaded connection structure. The rigid conductor 2 is composed of an inner multi-core data transmission cable, a middle power supply conductor, an outer rigid insulating sheath, and an outermost wear-resistant nylon braided layer. This solves the problem that traditional flexible cables cannot provide sufficient pushing force in a full water pipe, achieving the effect of transmitting both electrical energy and data, as well as transmitting axial thrust to drive the module forward.

[0033] The quick-seal wire connector 4 is installed at the inlet of the pipe 5 to be tested during the testing process. Internally, it contains a flexible sealing body 4-1 made of food-grade silicone material and an elastic compensator (such as a V-shaped stainless steel spring energy storage ring) on ​​its outer periphery. This solves the technical problem of seal failure and leakage of pressurized liquid inside the pipe due to wear during repeated wire feeding and extraction. It achieves the effect of automatically releasing pre-tightening force through the elastic compensator to compensate for wear gaps and realize a dynamic seal throughout the entire process.

[0034] The processing terminal is located outside the pipe 5 under test, and it has a built-in material classification model, underwater distortion correction module, cylindrical surface unfolding module, temporal feature enhancement module and defect identification model.

[0035] To enable those skilled in the art to better understand how to apply the above-mentioned device in actual testing operations and achieve the technical effects, the testing method of the present invention will be described below in conjunction with specific application scenarios.

[0036] This embodiment takes the leakage detection of a 20mm inner diameter PPR plastic water supply pipe in a residential building as an example to specifically illustrate the implementation steps of the detection method of the present invention, the technical problems it solves, and the technical effects it achieves.

[0037] Example 1: Inspection of PPR Plastic Water Supply Pipes S100. Place the biomimetic multi-segment flexible camera module into the pipe to be tested, and obtain the initial image of the inner wall of the pipe through the camera. In this embodiment, the operator closes the main pipeline valve and inserts the biomimetic multi-segment flexible camera module 1 into the pipeline 5 to be tested through a pre-reserved interface, such as an angle valve interface. A quick-sealing wire connector 4 is installed at the pipeline inlet, and the locking nut is tightened to establish an initial seal between the flexible sealing body within the sealing tongue 4-1 and the outer periphery of the rigid wire 2. This solves the problem of traditional endoscope wire interfaces lacking a sealing design, making them unsuitable for use in pressurized water supply pipelines.

[0038] S200: The processing terminal identifies the current pipe material type based on the initial image and adjusts the excitation output parameters of the magnetic controller based on the identification result to compensate for magnetic field attenuation caused by differences in pipe material or wall thickness, so that the effective magnetic field strength acting on the magnetic response component after penetrating the pipe wall is maintained within the preset working range; specifically: S210. The processing terminal calls the pre-trained material classification model to identify the current pipe material type based on the texture features and reflective properties of the inner wall of the pipe. S220: The processing terminal retrieves and identifies the excitation compensation coefficient corresponding to the pipe material type from the preset mapping relationship between material and excitation compensation parameters. S230, The processing terminal adjusts the base excitation current of the magnetic controller according to the excitation compensation coefficient; In this implementation, power is supplied to camera 1-1 and LED beads 1-2 via hard wire 2, and the camera begins to acquire initial images of the inner wall of the pipe. The processing terminal calls a pre-trained material classification model, and based on the matte texture and milky white color characteristics of the inner wall of the pipe, identifies the current pipe material as PPR.

[0039] Subsequently, the processing terminal retrieves the excitation compensation coefficient corresponding to the PPR material and 2.8mm wall thickness from its internally preset "material-excitation compensation parameter mapping table". Based on this coefficient, the processing terminal generates a current adjustment command and sends it to the magnetic controller 6, setting the base excitation current of the magnetic controller 6 to 0.7A.

[0040] It achieves adaptive adjustment of excitation intensity, ensuring that the effective magnetic field strength acting on the magnetic response component after penetrating the pipe wall remains within a stable preset working range regardless of changes in pipe material. This guarantees the consistency and reliability of the guiding force, making one set of devices applicable to both plastic and metal pipes.

[0041] To avoid the varying degrees of magnetic field attenuation due to different plastic / metal materials and wall thicknesses. If a fixed excitation output is used, excessive magnetic field attenuation in metal pipes may lead to guidance failure, while in plastic pipes it may result in energy waste or an excessively strong magnetic field.

[0042] S300: The biomimetic multi-segment flexible camera module is propelled forward by a rigid wire connected to it. When it reaches a branch node of the pipe under test, the camera captures an image of the branch opening. The processing terminal determines the target branch direction based on the image and controls the magnetron to adjust the magnetic field polarity. This causes the magnetic response component to generate a deflection torque on its fitted splicing segment under the action of the changing magnetic field. Through the directional deflection of the flexible connector, the local axis of the biomimetic multi-segment flexible camera module bends towards the target branch pipe. Specifically: The outer peripheral surface of the magnetic response component is provided with multiple elastic limiting beads. When the biomimetic multi-segment flexible camera module travels in pipe segments with different inner diameters, the elastic limiting beads support the first splicing segment with integrated camera at the center position of the pipe through elastic contact with the inner wall of the pipe under test, so that the optical axis of the camera is basically coincident with the center line of the pipe under test. The contact friction between the elastic limiting bead and the inner wall of the pipe under test is configured to be less than the maximum magnetic traction force that the magnetron can apply through the magnetic response component.

[0043] In this embodiment, the operator holds the magnetic controller 6 close to the outer wall of the pipe and pushes the rigid wire 2 to make the biomimetic multi-segment flexible camera module 1 move at a constant speed along the axial direction of the pipe. During this process, the elastic limiting beads 3-2 on the outer periphery of the magnetic response component are compressed by the pipe wall, generating elastic support force to keep the first splicing segment carrying the camera 1-1 always supported in the center position of the pipe. This avoids the biomimetic multi-segment flexible camera module 1 from drooping due to gravity in non-vertical or slightly undulating pipes, causing the camera to deviate to one side of the pipe wall, resulting in asymmetrical images and blind spots. It ensures that the optical axis of the camera is basically aligned with the center line of the pipe, and the original image acquired is a standard centrally symmetrical annular fisheye image, providing a high-quality image source for subsequent accurate distortion correction and cylindrical unfolding.

[0044] In one embodiment, the magnetic response component is sleeved on the outer periphery of at least one intermediate splicing segment in the biomimetic multi-segment flexible camera module, excluding the first splicing segment that integrates the camera, and a high magnetic permeability material layer is encapsulated inside the flexible connector between the magnetic response component and the first splicing segment. During the magnetic field polarity switching process, the high permeability material layer confines the magnetic field lines to the region where the magnetic response component is located, thereby suppressing the leakage of the magnetic field to the region where the camera is located.

[0045] In one embodiment, it further includes: S310. When the biomimetic multi-segment flexible camera module enters the bend or branch node area of ​​the pipeline under test, the processing terminal monitors the magnetic field coupling state between the magnetron and the magnetic response component in real time. S320. When the magnetic field coupling strength is detected to drop to a preset threshold, the processing terminal controls the magnetic controller to output an instantaneously enhanced auxiliary excitation pulse. The intensity of the auxiliary excitation pulse is higher than the steady-state excitation strength, so that the magnetic response component generates an instantaneous traction torque to overcome static friction resistance. S330: After determining that the bionic multi-segment flexible camera module has completed the turning action based on the changes in the image captured by the camera, the processing terminal controls the magnetic controller to restore the excitation intensity to a steady-state level.

[0046] In this example, when the biomimetic multi-segment flexible camera module 1 moves to the T-junction branch node between the kitchen and the bathroom, camera 1-1 captures an image of the branch. The processing terminal analyzes the image, or the operator judges based on the displayed screen, to determine that the target branch direction is the bathroom direction.

[0047] The processing terminal then controls the magnetic controller 6 to switch the magnetic field polarity of the electromagnet unit. For example, when a turn to the right branch is required, the magnetic controller 6 generates a magnetic field distribution that repels the guide strong magnet 3-1. This repulsive force acts on the magnetic response component on the second splicing segment, generating a lateral deflection torque.

[0048] The deflection torque is transmitted through the flexible connector, forcing the local axis of the bionic multi-segment flexible camera module 1 to bend toward the bathroom branch pipe, thereby guiding the entire module smoothly into the target branch and preventing it from mistakenly entering the kitchen branch.

[0049] During this turning process, the processing terminal monitors the magnetic field coupling status in real time. Due to the bending of the flexible connector and friction of the pipe wall, the travel resistance increases sharply, which may cause a temporary decrease in the magnetic field coupling strength. When this value is detected to be lower than a preset threshold, the processing terminal controls the magnetic controller 6 to output a momentarily enhanced auxiliary excitation pulse, such as 1.1A, to generate a momentary strong traction torque to help the module overcome static friction resistance. Once image analysis confirms that the module has completed the turning, the excitation strength automatically recovers to the steady-state 0.7A.

[0050] This invention achieves precise directional and selective entry at multi-branch nodes in small-diameter pipes through external non-contact magnetic navigation. A transient magnetization-assisted strategy ensures steering reliability under conditions of sudden increases in resistance, preventing jamming.

[0051] S400. During the process, the processing terminal performs coordinate mapping transformation on each pixel in the original annular image captured by the camera based on the refractive index parameters of the liquid medium inside the pipe under test, in order to compensate for the deformation caused by light refraction, and maps the corrected annular image into a distortion-free two-dimensional planar unfolded image reflecting the geometric morphology of the inner wall of the pipe; specifically: S410. Obtain the refractive index parameters of the liquid medium inside the pipe to be tested; S420. Based on the refractive index parameter, the equivalent camera intrinsic parameter matrix of the camera in the underwater environment is recalibrated, wherein the equivalent focal length parameter is scaled proportionally according to the ratio of the refractive index of the medium to the refractive index of air. S430: Using the recalibrated camera intrinsic parameter matrix, perform coordinate mapping transformation on each pixel in the original circular image to compensate for the refraction and deflection of light at the interface between the liquid medium and the camera optical element. In this embodiment, during travel, camera 1-1 captures a raw, circular fisheye image inside a water-filled pipe. Due to refraction of light as it enters the camera's air-medium lens from the water medium, this image suffers from severe radial and tangential distortion.

[0052] The processing terminal first obtains a preset water refractive index parameter, such as 1.33. Then, its built-in underwater distortion correction module recalibrates the equivalent camera intrinsic parameter matrix based on this refractive index parameter. Specifically, the equivalent focal length parameter is scaled according to the ratio of "water refractive index / air refractive index". Using the recalibrated intrinsic parameter matrix, a coordinate mapping transformation is performed on each pixel in the original image, thereby accurately compensating for the geometric deformation caused by the refraction of light at the medium interface.

[0053] Next, the cylindrical surface unfolding module maps the distortion-corrected annular image, based on the cylindrical surface projection model established by the pipe's inner diameter parameters, into a distortion-free two-dimensional planar unfolded image that reflects the true geometric shape of the pipe's inner wall.

[0054] The underwater environment causes non-linear refractive distortion in images captured by fisheye lenses, resulting in severely distorted pipe wall images when directly unfolded. Furthermore, the circular fisheye view does not conform to human visual perception and cannot be directly input into AI disease recognition models trained on conventional views.

[0055] This invention restores the true geometric proportions of the pipe's inner wall by eliminating the influence of liquid medium refraction on image quality. It transforms complex panoramic images into intuitive two-dimensional plans, enabling subsequent high-precision identification based on images that reflect the true shape, significantly improving recognition accuracy.

[0056] The S500 processing terminal inputs the distortion-free two-dimensional planar unfolded image into the defect identification model. When a defect is identified, it performs spatial coordinate conversion based on the current travel distance data associated with the hard conductor's delivery length and the guide trajectory record of the magnetic controller to determine the physical location of the defect in the pipeline under test. Specifically: S510, Acquire the mileage signal associated with the hard conductor pull-out length; The S520 and processing terminal, based on mileage signals, spatially register and stitch together multiple consecutive frames of distortion-free two-dimensional planar unfolded images to generate a continuous panoramic detection image of the inner wall of the pipe under test. The S530 and processing terminal will overlay the identified disease types and location information as visual markers onto the continuous panoramic detection image, and trigger a location marker indication on the outside of the pipeline under test.

[0057] In this embodiment, when a defect is detected, the processing terminal initiates a positioning program. This program performs a dual-verification spatial coordinate conversion based on the following two data points: Mileage data: Mileage signals recorded by a wire feed encoder located at the pipe inlet, precisely correlated with the length of the hard wire 2 being drawn, are used to determine the axial distance of the defect.

[0058] Guide trajectory recording: The direction of each turn and the distance traveled in each branch are recorded by the magneto controller 6, which is used to determine the specific branch pipeline where the defect is located.

[0059] Based on the above information, the processing terminal accurately calculates the three-dimensional spatial location of the defect relative to the pipeline inlet.

[0060] Simultaneously, the processing terminal will detect all distortion-free two-dimensional planar unfolded images generated throughout the process, perform spatial registration and seamless stitching based on mileage signals, and generate a complete continuous panoramic inspection image of the pipe's inner wall. Finally, the identified "loose interface" defect type identifier and its precise location will be superimposed on this panoramic image as a prominent visual marker, and an external laser pointer will be triggered to project a visible light spot at the corresponding location on the bathroom wall.

[0061] This invention eliminates the cumulative error of a single positioning method by using dual verification of mileage data and guide trajectory, achieving centimeter-level precise positioning of the physical location of defects. By generating continuous panoramic inspection images and overlaying defect markers, it provides maintenance personnel with an intuitive and comprehensive pipeline "health report," guiding precise excavation and avoiding blind, large-scale destructive construction.

[0062] In one embodiment, before inputting the distortion-free two-dimensional planar unfolded image into the disease identification model, the method further includes: The processing terminal performs inter-frame difference analysis on multiple consecutive frames of distortion-free two-dimensional planar unfolded images; Based on the inter-frame difference results, the processing terminal identifies candidate regions where the pixel position remains relatively fixed across multiple frames and the pixel change pattern differs from the background pixel displacement pattern. After the processing terminal performs weighted enhancement on the image features of the candidate regions, it is then input into the disease identification model.

[0063] It should be noted that in the normal background area of ​​the pipe wall, the pixel change pattern exhibits a regular flow caused by the movement of the camera. However, static defects such as microcracks and tiny pinholes have relatively fixed pixel positions across multiple frames, and their texture change patterns are independent of the background flow pattern. Through inter-frame differencing, these candidate regions with "inconsistent motion" can be accurately identified. Subsequently, the algorithm assigns a higher weight coefficient to the image feature response values ​​of these candidate regions than to the background regions.

[0064] After feature enhancement, the entire image frame is input into a defect identification model using an improved YOLOv8 network structure. When the model detects a "loose connection" feature at a bend in a branch pipe in the bathroom, an alarm is immediately triggered.

[0065] In dynamically captured video streams, small static targets such as microcracks with a width of only 0.1 mm are easily obscured by rapidly changing background textures, leading to a high false negative rate in traditional single-frame image recognition methods. This invention enhances the spatial dimension of static target features by introducing inter-frame difference information in the temporal dimension, significantly improving the detection confidence and recall rate of minute static defects in dynamic backgrounds.

[0066] In one embodiment, the detection device further includes a quick-sealing connector installed at the inlet of the pipe to be tested, the quick-sealing connector having a flexible sealing body 4-1 and an elastic compensator inside; The detection methods also include: During the extraction of the rigid conductor, the main sealing interface is formed by the sliding contact between the inner surface of the flexible seal and the outer peripheral surface of the rigid conductor. The flexible seal is continuously subjected to radial inward preload by the elastic compensator; When the flexible seal experiences material loss due to relative motion with the rigid wire, the elastic compensator releases elastic potential energy to push the flexible seal to contract inward, thereby compensating for the fit gap caused by material loss.

[0067] In this embodiment, after the detection is completed, the excitation output of the magnetron 6 is maintained, the polarity of the magnetic field is adjusted to generate a reverse traction force, and in conjunction with the operator's pulling force on the rigid wire 2, the bionic multi-segment flexible camera module 1 is guided to exit along the original path.

[0068] Throughout the testing and withdrawal process, the rigid wire 2 is repeatedly pumped in and out. The flexible seal inside the quick-sealing wire connector 4 experiences minor material loss due to friction. At this time, the elastic compensator located on its outer side releases its pre-stored elastic potential energy, continuously pushing the flexible seal inward to ensure it always fits tightly against the outer circumference of the rigid wire 2.

[0069] This avoids the inevitable seal failure and leakage of pressurized water from the pipe caused by repeated movement of the conductor during inspection operations that last for tens of minutes. It achieves dynamic adaptive compensation of the sealing structure, ensuring continuous and reliable sealing of the liquid medium inside the pipeline throughout the entire inspection cycle, eliminating water leakage and ensuring a clean and safe work site.

[0070] Example 2: Inspection of Stainless Steel Water Supply Pipelines This embodiment is basically the same as Embodiment 1, except that the pipe to be tested is a stainless steel pipe with an inner diameter of 25mm. The following only describes the different technical features and their resulting technical effects.

[0071] In step S200, the processing terminal identifies the pipe material as stainless steel based on the initial frame image. Since the stainless steel pipe wall has a significant shielding and attenuation effect on the magnetic field, the processing terminal retrieves the corresponding excitation compensation coefficient from the mapping relationship and sends a command to the magnetic controller 6 to automatically adjust its base excitation current to 1.2A.

[0072] In this embodiment, although the stainless steel pipe wall caused severe magnetic field attenuation, through intelligent compensation, the effective magnetic field strength acting on the magnetic response component after penetrating the pipe wall was successfully maintained at a level basically equivalent to that under the PPR pipe working condition in Embodiment 1. Therefore, during the subsequent turning process, the magnetic traction force remained sufficient, the guiding action was smooth and reliable, and there was no guiding failure or turning overshoot caused by magnetic field attenuation.

[0073] Example 3: Microcrack Detection in PE Plastic Water Supply Pipes This embodiment is basically the same as Embodiment 1, except that the pipe to be tested is a PE plastic pipe with an inner diameter of 20mm, and there is a suspicion of extremely small initial leakage cracks.

[0074] Because the inner wall of PE pipes has a relatively smooth texture, microcracks are extremely inconspicuous in a single frame image. The processing terminal successfully identified a candidate region with a fixed location across multiple frames, but whose pixel variation pattern subtly differed from the background flow pattern, by performing inter-frame differing analysis on three consecutive frames. After the features of this candidate region were enhanced by the algorithm, they were input into the defect identification model, which ultimately identified with high confidence the presence of an axial microcrack with a width of only 0.1 mm at that location.

[0075] In traditional methods without this step, such microcracks are highly likely to be missed as background noise. This invention addresses the industry pain point of small, initial leaks being difficult to detect visually by introducing time-dimensional information.

[0076] The processing terminal stitches the entire image into a continuous panoramic inspection image and marks the precise location of the microcrack on it. Based on this panoramic report, maintenance personnel can accurately locate and repair the defect by performing only minimal local disassembly and assembly.

[0077] Not only is the location accurate, but it also provides a global perspective for maintenance decisions, avoiding the large-scale destructive excavation that traditional methods have to carry out due to the inability to locate accurately, and greatly reducing maintenance costs and the waste of social resources.

[0078] A magnetically controlled navigation endoscopic inspection device for small-diameter pipes includes: The biomimetic multi-segment flexible camera module 1 is composed of multiple splicing segments connected in series by flexible connectors. The first splicing segment of the biomimetic multi-segment flexible camera module integrates an image acquisition unit, which includes a camera 1-1 and LED beads 1-2, and is used to acquire images of the inner wall of the pipe in the liquid medium environment inside the pipe to be tested. A magnetic response component is fitted around the periphery of at least one splicing segment in the biomimetic multi-segment flexible camera module, excluding the first splicing segment, and is physically separated from the camera of the image acquisition unit; at least one guiding strong magnet 3-1 is embedded in the magnetic response component; The rigid wire 2 is fixedly connected at its proximal end to the tail end of the bionic multi-segment flexible camera module 1. The rigid wire 2 is used to supply power to the image acquisition unit, transmit image data of the inner wall of the pipe, and provide pushing force for the bionic multi-segment flexible camera module 1 to move in the straight pipe section. The quick-seal wire connector 4 is used to be installed at the inlet of the pipe 5 to be tested during the testing process. The quick-seal wire connector 4 is sleeved on the outer periphery of the rigid wire 2 and is configured to allow the rigid wire 2 to be pumped axially while maintaining the seal of the liquid medium inside the pipe 5 to be tested. Magnetizer 6 is set outside the pipe 5 to be tested. Magnetizer 6 is configured to generate a variable magnetic field. After penetrating the pipe wall, the variable magnetic field acts on the guide strong magnet 3-1 of the magnetic response component to generate a magnetic traction force to directionally turn the biomimetic multi-segment flexible camera module 1. The processing terminal is located outside the pipe 5 to be tested and is connected to the image acquisition unit via a hard wire 2. The processing terminal has an underwater distortion correction module, a cylindrical surface unfolding module and a defect identification model built in, and is configured to perform distortion correction, cylindrical surface unfolding and defect identification processing on the original images acquired by the image acquisition unit in sequence, so as to identify defects on the inner wall of the pipe and determine the location of the defects.

[0079] The foregoing description of specific exemplary embodiments of the present invention is for illustrative and explanatory purposes. These descriptions are not intended to limit the invention to the precise forms disclosed, and it is obvious that many changes and variations can be made based on the above teachings. Although embodiments of the invention have been shown and described, these specific embodiments are merely explanations of the invention and are not intended to limit it. The specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. The purpose of selecting and describing exemplary embodiments is to explain the specific principles of the invention and its practical application, so that those skilled in the art, after reading this specification, can make modifications, substitutions, variations, and various choices and changes to the embodiments as needed without departing from the principles and spirit of the invention, provided that such modifications, substitutions, variations, and choices and changes are within the scope of the claims of the invention and are protected by patent law.

Claims

1. A magnetically controlled navigation pipe endoscopic inspection method for small-diameter pipes, applied to a magnetically controlled navigation pipe endoscopic inspection device, characterized in that, The detection device includes a biomimetic multi-segment flexible camera module consisting of multiple spliced ​​segments connected in series by flexible connectors, a magnetic response component sleeved on the outer periphery of the spliced ​​segments, a magnetic controller set on the outside of the pipe to be tested, and a processing terminal. The detection method includes: The biomimetic multi-segment flexible camera module is placed inside the pipe to be tested, and the initial image of the inner wall of the pipe to be tested is obtained through the camera. The processing terminal identifies the current pipe material type based on the initial image and adjusts the excitation output parameters of the magnetron based on the identification result, so that the effective magnetic field strength acting on the magnetic response component after penetrating the pipe wall is maintained within a preset working range. The bionic multi-segment flexible camera module is propelled forward by a rigid wire connected to it. When it reaches a branch node of the pipe under test, the camera captures an image of the branch opening. The processing terminal determines the target branch direction based on the image of the branch opening and controls the magnetron to adjust the polarity of the magnetic field to guide the bionic multi-segment flexible camera module to bend toward the target branch pipe. During the process, the processing terminal performs coordinate mapping transformation on each pixel in the original annular image captured by the camera based on the refractive index parameter of the liquid medium in the pipe under test, and maps the corrected annular image into a distortion-free two-dimensional planar unfolded image that reflects the geometric shape of the inner wall of the pipe. The processing terminal inputs the distortion-free two-dimensional planar unfolded image into the defect identification model. When a defect is identified, it performs spatial coordinate conversion based on the current travel mileage data associated with the hard conductor's delivery length and the guide trajectory record of the magnetic controller to determine the physical location of the defect in the pipeline under test.

2. The magnetically controlled navigation endoscopic inspection method for small-diameter pipes according to claim 1, characterized in that, The magnetic response component is sleeved on the outer periphery of at least one intermediate splicing segment in the biomimetic multi-segment flexible camera module, excluding the first splicing segment that integrates the camera, and a high magnetic permeability material layer is encapsulated inside the flexible connector between the magnetic response component and the first splicing segment. During the magnetic field polarity switching process, the high permeability material layer confines the magnetic field lines to the region where the magnetic response component is located, thereby suppressing the leakage of the magnetic field to the region where the camera is located.

3. The magnetically controlled navigation endoscopic inspection method for small-diameter pipes according to claim 2, characterized in that, The outer peripheral surface of the magnetic response component is provided with multiple elastic limiting beads. When the biomimetic multi-segment flexible camera module travels in pipe segments with different inner diameters, the elastic limiting beads support the first splicing segment integrating the camera at the center position of the pipe under test by elastic contact with the inner wall of the pipe under test, so that the optical axis of the camera is basically coincident with the center line of the pipe under test. The contact friction between the elastic limiting bead and the inner wall of the pipe under test is configured to be less than the maximum magnetic traction force that the magnetron can apply through the magnetic response component.

4. The magnetically controlled navigation endoscopic inspection method for small-diameter pipes according to claim 1, characterized in that, The processing terminal identifies the current pipe material type based on the initial image and adjusts the excitation output parameters, including: The processing terminal calls a pre-trained material classification model to identify the current pipe material type based on the texture features and reflective properties of the pipe's inner wall. The processing terminal retrieves and identifies the excitation compensation coefficient corresponding to the pipe material type from the preset mapping relationship between material and excitation compensation parameters. The processing terminal adjusts the base excitation current of the magnetron according to the excitation compensation coefficient.

5. The magnetically controlled navigation endoscopic inspection method for small-diameter pipes according to claim 4, characterized in that, Also includes: When the biomimetic multi-segment flexible camera module enters the bend or branch node area of ​​the pipeline under test, the processing terminal monitors the magnetic field coupling state between the magnetron and the magnetic response component in real time. When the magnetic field coupling strength is detected to drop to a preset threshold, the processing terminal controls the magnetic controller to output an instantaneously enhanced auxiliary excitation pulse. The intensity of the auxiliary excitation pulse is higher than the steady-state excitation strength, so that the magnetic response component generates an instantaneous traction torque to overcome static friction resistance. After determining that the bionic multi-segment flexible camera module has completed the turning action based on the changes in the image captured by the camera, the processing terminal controls the magnetic controller to restore the excitation intensity to a steady-state level.

6. The magnetically controlled navigation endoscopic inspection method for small-diameter pipes according to claim 1, characterized in that, The processing terminal performs coordinate mapping transformation on each pixel in the original annular image captured by the camera based on the refractive index parameter of the liquid medium inside the pipe under test, including: Obtain the refractive index parameters of the liquid medium inside the pipe under test; Based on the refractive index parameter, the equivalent camera intrinsic parameter matrix of the camera in the underwater environment is recalibrated, wherein the equivalent focal length parameter is scaled proportionally according to the ratio of the refractive index of the medium to the refractive index of air. Using the recalibrated camera intrinsic parameter matrix, a coordinate mapping transformation is performed on each pixel in the original ring image to compensate for the refraction and deflection of light at the interface between the liquid medium and the camera optical element.

7. The magnetically controlled navigation endoscopic inspection method for small-diameter pipes according to claim 1, characterized in that, Before inputting the distortion-free two-dimensional planar unfolded image into the disease identification model, the method further includes: The processing terminal performs inter-frame difference analysis on multiple consecutive frames of distortion-free two-dimensional planar unfolded images; The processing terminal identifies candidate regions whose pixel positions remain relatively fixed across multiple frames and whose pixel change patterns differ from the background pixel displacement patterns based on the inter-frame difference results. The processing terminal performs weighted enhancement on the image features of the candidate region before inputting them into the disease identification model.

8. The magnetically controlled navigation endoscopic inspection method for small-diameter pipes according to claim 1, characterized in that, Determining the physical location of the defect in the pipe to be tested includes: Acquire the mileage signal associated with the length of the hard conductor pull-out; Based on the mileage signal, the processing terminal performs spatial registration and stitching on multiple consecutive frames of distortion-free two-dimensional planar unfolded images to generate a continuous panoramic detection image of the inner wall of the pipe under test. The processing terminal overlays the identified disease type and location information as visual markers onto the continuous panoramic detection image and triggers a location marker indication on the outside of the pipeline under test.

9. The magnetically controlled navigation endoscopic inspection method for small-diameter pipes according to claim 1, characterized in that, The detection device also includes a quick-sealing connector installed at the inlet of the pipe to be tested, and the quick-sealing connector is provided with a flexible sealing body and an elastic compensating body inside. The detection method further includes: During the extraction of the rigid wire, the main sealing interface is formed by the sliding contact between the inner surface of the flexible seal and the outer peripheral surface of the rigid wire. The elastic compensator continuously applies a radially inward preload force to the flexible seal; When the flexible seal experiences material loss due to relative movement with the rigid wire, the elastic compensator releases elastic potential energy to push the flexible seal to contract inward, thereby compensating for the fit gap caused by material loss.

10. A magnetically controlled navigation endoscopic inspection device for small-diameter pipes, using the magnetically controlled navigation endoscopic inspection method for small-diameter pipes as described in any one of claims 1-9, characterized in that, include: The biomimetic multi-segment flexible camera module is composed of multiple splicing segments connected in series by flexible connectors, with the camera and LED integrated in the first splicing segment. A magnetic response component is sleeved on the outer periphery of at least one splicing segment other than the first splicing segment, and is physically separated from the camera; the magnetic response component is embedded with at least one strong guiding magnet. A rigid wire, the proximal end of which is fixedly connected to the tail end of the biomimetic multi-segment flexible camera module; A quick-sealing wire connector is fitted around the outer circumference of the rigid wire and is used for installation at the inlet of the pipe to be tested. The magnetron is installed outside the pipe to be tested; The processing terminal is located outside the pipe to be tested and is communicatively connected to the camera.

Images