CN-121811205-B - Multi-mode data fusion driven drainage pipeline monitoring equipment advancing method

CN121811205BCN 121811205 BCN121811205 BCN 121811205BCN-121811205-B

Abstract

The invention relates to the technical field of drainage pipeline detection and multi-mode data fusion, and discloses a drainage pipeline monitoring equipment advancing method driven by multi-mode data fusion, which comprises the following steps of acquiring original acquisition data through the drainage pipeline monitoring equipment; the drainage pipeline monitoring device integrates a vision acquisition unit, an infrared thermal imaging unit, a self-adaptive pipe diameter ground penetrating radar unit and a travelling mechanism, performs fusion processing on raw acquired data to construct a standardized multi-mode data set, performs fusion feature extraction on the standardized multi-mode data set, outputs disease environment fusion data of the drainage pipeline, wherein the disease environment fusion data comprise disease types, disease position coordinates, disease confidence and pipeline pipe diameter real-time measuring and calculating values, controls the travelling of the drainage pipeline monitoring device through the disease environment fusion data, and controls travelling path optimization based on the disease types during travelling. According to the technical scheme, the monitoring equipment can be adapted to drainage pipelines with different diameters and different diseases.

Inventors

ZHANG HAICHAO
ZHANG RAN
YANG MENG
FANG HONGYUAN
WANG NIANNIAN
HUANG SHICHENG
ZHANG QING
LUO CHAO
WEN PENG
WANG HANTAO
PING YANG

Assignees

中国电建集团贵阳勘测设计研究院有限公司

Dates

Publication Date: 20260512
Application Date: 20260310

Claims (7)

1. The multi-mode data fusion driven drainage pipeline monitoring equipment advancing method is characterized by comprising the following steps of: The method comprises the steps that original acquisition data are acquired through drainage pipeline monitoring equipment, the drainage pipeline monitoring equipment integrates a visual acquisition unit, an infrared thermal imaging unit, a self-adaptive pipe diameter ground penetrating radar unit and a travelling mechanism, the self-adaptive pipe diameter ground penetrating radar unit comprises a ground penetrating radar sensor and a self-adaptive mechanical arm, the self-adaptive mechanical arm adopts a multi-section telescopic structure and is provided with a laser distance sensor, a ground penetrating radar sensor, a PID controller and a torque adjusting module, the original acquisition data comprise drainage pipeline visual information, pipeline inner wall temperature distribution data, pipeline inner radar signals, position coordinates of the drainage pipeline monitoring equipment and mechanical arm gesture data, the drainage pipeline visual information is used for reflecting inner wall textures and disease forms of the drainage pipeline, the pipeline inner wall temperature distribution data are used for highlighting temperature differences between disease areas and normal areas, and the pipeline inner radar signals are used for reflecting three-dimensional structure data of pipeline structure layer thickness and inner defects; the method comprises the steps of carrying out fusion processing on the original acquired data to construct a standardized multi-mode data set, wherein the construction of the standardized multi-mode data set comprises the steps of carrying out space alignment on a standardized panoramic image sequence, an infrared thermal imaging temperature value graph and a two-dimensional structure characteristic graph through time stamp synchronization and space coordinate mapping, adding position coordinates of drainage pipeline monitoring equipment, and unifying data formats to generate the standardized multi-mode data set; The method comprises the steps of carrying out fusion feature extraction on a standardized multi-mode dataset to output disease environment fusion data of a drainage pipeline, wherein the disease environment fusion data comprise disease types, disease position coordinates, disease confidence and pipeline pipe diameter real-time measuring and calculating values, carrying out feature coding on the standardized panoramic image sequence by adopting an improved ResNet-50 network to output a semantic feature map for capturing outlines and details of pipeline surface diseases, carrying out processing on the infrared thermal imaging temperature value map by using a 3-layer convolutional neural network to output a temperature anomaly feature map and highlight temperature anomaly region features corresponding to the diseases, carrying out feature extraction on the two-dimensional structural feature map by using a CNN+ full-connection layer to obtain a structural feature map, reflecting the thickness of a pipeline structure and the structural feature of the defect, constructing a dynamic weighting fusion module driven by an attention mechanism, inputting the semantic feature map, the temperature anomaly feature map and the structural feature map to generate a fusion feature map with uniform dimensions, decoding the fusion feature map, and outputting the drainage pipeline disease environment fusion data; Controlling the running of the drainage pipeline monitoring equipment through the disease environment fusion data, and controlling the running path optimization based on the disease type during running, wherein the running path optimization comprises the steps of controlling the detection mode of the mechanical arm and controlling the running path of the drainage pipeline monitoring equipment; The self-adaptive control of the disease of the travelling system is realized by combining disease distribution characteristics and pipeline environment by a travelling mechanism of the drainage pipeline monitoring equipment, and dynamically adjusting speed, track and travelling parameters; The detection mode of the mechanical arm is an implementation mode combining the disease type and the disease confidence, the real-time measurement and calculation value of the pipe diameter is taken as a reference, the mechanical arm is driven by a PID control algorithm to adjust the unfolding angle and the length in real time, and the detection mode is adapted to different pipe diameter environments, and comprises a mechanical arm disease self-adaptive adjustment mode and sensor parameter disease self-adaptive optimization.
2. The multi-modality data fusion driven drain line monitoring facility travel method of claim 1, wherein the constructing a standardized multi-modality data set comprises the steps of: performing panoramic video preprocessing on the visual information of the drainage pipeline to generate a standardized panoramic image sequence; Preprocessing the temperature distribution data of the inner wall of the pipeline by using infrared thermal imaging data to form an infrared thermal imaging temperature value diagram; and preprocessing ground penetrating radar data of the radar signals in the pipeline, and converting the radar signals into a two-dimensional structural feature map.
3. A multi-modal data fusion driven drain line monitoring apparatus travel method as defined in claim 1 wherein, The sensor parameter disease self-adaptive optimization means that the sensor adjusts working parameters according to disease types, and data acquisition accuracy and effectiveness are maximized; The self-adaptive adjustment mode of the mechanical arm disease comprises a fine detection mode, a planar scanning mode, a focusing tracking mode and a safe avoiding mode.
4. A multi-modal data fusion driven drain line monitoring apparatus travel method as claimed in claim 3 wherein, Adjusting the sampling frequency of a laser distance sensor and the adjusting precision of a PID controller, ensuring that a radar sensor is attached to the pipe wall along a disease area, and collecting crack depth and extension track data; the planar scanning mode comprises the steps of controlling a mechanical arm to transversely move a ground penetrating radar sensor in a designated step length, performing grid scanning on a disease area, synchronously recording the structural thickness changes at different positions, and generating a disease degree distribution map; The focusing tracking mode comprises the steps of obtaining position coordinates of a disease area center point, driving a radar sensor to be close to the disease area center point, and enhancing signal strength to identify aperture and trend of the center point; The safe avoidance mode comprises the steps of shortening the telescopic length of the mechanical arm, adjusting the angle of the radar sensor, obliquely collecting three-dimensional structure data of a disease area, and avoiding signal distortion caused by direct opposite to the disease area.
5. A multi-modal data fusion driven drain line monitoring apparatus travel method as claimed in claim 3 wherein said controlling travel path optimization based on disease type comprises: If the disease confidence reaches a specified threshold, switching the self-adaptive adjustment mode of the mechanical arm disease to a fine detection mode; the sensor parameter disease self-adaptive optimization comprises the steps of adjusting the frame rate of a panoramic camera, improving the image resolution, enhancing the detail of cracks, enhancing the contrast of edges of the cracks, focusing a thermal infrared imager temperature measuring interval, improving the temperature resolution, capturing fine temperature differences at the cracks, only keeping temperature data of the periphery of the cracks, and shortening the temperature sampling interval; If the disease confidence reaches a specified threshold value, the self-adaptive adjusting mode of the mechanical arm disease is switched to a planar scanning mode; the sensor parameter disease self-adaptive optimization comprises the steps of adjusting the center frequency of a ground penetrating radar to match different corrosion layer thicknesses, starting an area for constant temperature monitoring by using a thermal infrared imager, recording the temperature average value of the corrosion area with a specified preset frequency, calculating the temperature difference value of the corrosion area and a normal area, and generating a corrosion degree thermodynamic diagram; The self-adaptive control of the travelling system disease comprises that if leakage points are dense, a drainage pipeline monitoring device pauses travelling and rotates in situ, and the relative positions of the leakage points and a detection area are positioned by combining the high-definition industrial camera with a ground penetrating radar sensor to generate a 'leakage point-area' correlation map; When the disease type is structural damage defect, the self-adaptive regulation mode of the mechanical arm disease is switched to a safe avoidance mode, the self-adaptive optimization of sensor parameter disease comprises the steps of improving the transmitting power of a radar sensor to ensure that deep structural data are acquired by penetrating through a deposit at a damaged position, and the self-adaptive control of the disease of the travelling system comprises the steps of controlling drainage pipeline monitoring equipment to bypass, adjusting the action direction and avoiding equipment from sinking into a gap at the damaged position.
6. The multi-modal data fusion driven drainage pipeline monitoring device travel method of claim 1, wherein the travel path optimization is achieved through adaptive control, including mechanical arm adaptive adjustment control, device travel path and speed control and sensor parameter adaptive control.
7. The multi-mode data fusion-driven drainage pipeline monitoring equipment advancing method according to claim 6 is characterized in that the mechanical arm self-adaptive adjustment control comprises the steps of taking a real-time measuring and calculating value of the pipeline pipe diameter as a core control target, and constructing a PID closed-loop control system by combining real-time distance data fed back by a laser distance sensor; after the laser distance sensor detects the distance, judging whether the distance is within a specified range, if so, keeping the current gesture of the mechanical arm, otherwise, rapidly calculating the telescopic adjustment quantity by the equipment control system, driving the mechanical arm to perform telescopic action, adjusting the unfolding angle and length, and triggering the laser distance sensor to detect the distance again; The equipment travel path and speed control comprises the steps of generating an optimal travel track based on disease position coordinates, identifying a dense disease area, reducing travel speed, triggering bypass planning if large-scale structural defects are identified, and maintaining high-efficiency travel speed in a normal area without obvious diseases; the sensor parameter self-adaptive control comprises the steps of constructing a multi-mode data feedback closed loop, transmitting disease environment fusion data and original acquisition data to an equipment control system in real time to generate a control instruction, and dynamically adjusting the working parameters of the sensor.

Description

Multi-mode data fusion driven drainage pipeline monitoring equipment advancing method Technical Field The invention relates to the technical field of drainage pipeline detection and multi-mode data fusion, in particular to a drainage pipeline monitoring equipment advancing method driven by multi-mode data fusion. Background The drainage pipeline is a core component of urban infrastructure, and the health state of the drainage pipeline is directly related to urban flood control and drainage and ecological environment safety. In the long-term service process, the pipeline is easily influenced by factors such as geological sedimentation, water flow scouring, chemical corrosion and the like, and diseases such as cracks, corrosion, leakage, structural damage and the like are generated, and if the pipeline cannot be timely detected and repaired, serious consequences such as road collapse, water pollution and the like can be caused. The existing drainage pipeline detection equipment mainly has the following defects that firstly, data acquisition dimension is single, most equipment only depends on single video or radar data, visual state, temperature abnormality and structural defect inside a pipeline are difficult to comprehensively describe, secondly, pipe diameter suitability is poor, a traditional ground penetrating radar sensor is fixedly installed, detection accuracy is unstable due to the fact that detection distances cannot be adjusted according to pipelines with different pipe diameters, thirdly, data fusion degree is low, multi-source data are processed by adopting a simple splicing mode, complementary advantages of all mode data cannot be fully exerted, detection blind areas and misjudgment are easy to occur, thirdly, panoramic detection capability is insufficient, common video acquisition has visual field limitation, and dead angle coverage of the inner wall of the pipeline is difficult to realize. In recent years, the development of multi-modal data fusion and adaptive mechanical design techniques provides a new path for solving the above problems. In the prior art, part of pipeline detection equipment starts to try multi-mode data integration, but lacks a special fusion algorithm aiming at panoramic videos, infrared thermal imaging and ground penetrating radars, meanwhile, the self-adaptive mechanical arm design is applied to industrial assembly scenes, is not combined with the depth of the pipeline detection equipment, cannot meet the detection requirements of the ground penetrating radars in pipelines with different pipe diameters, and the advancing strategy of drainage pipeline monitoring equipment in the pipelines does not adjust different diseases in the pipelines, so that the disease degree is easily increased, and secondary damage of the pipelines is caused. Therefore, a technical scheme is needed, multi-mode data fusion is applied to drainage pipeline monitoring, the advancing strategy of drainage pipeline monitoring equipment is adjusted for different diseases in pipelines, and the detection requirements of pipelines with different pipe diameters are met, so that the method is suitable for omnibearing detection and disease diagnosis of urban drainage pipelines. Disclosure of Invention In order to achieve the above purpose, the present application provides a multi-mode data fusion driven drainage pipeline monitoring device traveling method, comprising the following steps: The method comprises the steps of acquiring original acquisition data through drainage pipeline monitoring equipment, wherein the drainage pipeline monitoring equipment integrates a visual acquisition unit 401, an infrared thermal imaging unit 403, a self-adaptive pipe diameter ground penetrating radar unit 402 and a travelling mechanism 404, the self-adaptive pipe diameter ground penetrating radar unit 402 comprises a ground penetrating radar sensor and a self-adaptive mechanical arm, the self-adaptive mechanical arm adopts a multi-section telescopic structure and is provided with a laser distance sensor, a ground penetrating radar sensor, a PID controller and a torque adjusting module, and the original acquisition data comprises drainage pipeline visual information, pipeline inner wall temperature distribution data, a pipeline internal radar signal, position coordinates of the drainage pipeline monitoring equipment and mechanical arm posture data; carrying out fusion processing on the original acquired data to construct a standardized multi-mode data set; extracting fusion characteristics of the standardized multi-mode data set, and outputting disease environment fusion data of the drainage pipeline, wherein the disease environment fusion data comprises disease types, disease position coordinates, disease confidence and pipeline pipe diameter real-time measuring and calculating values; the method comprises the steps of controlling the drainage pipeline monitoring equipment to run through disease environment fusion data, con