CN-121998483-A - Multi-mode sensing concrete construction quality detection system and method

CN121998483ACN 121998483 ACN121998483 ACN 121998483ACN-121998483-A

Abstract

The invention provides a multi-mode sensing concrete construction quality detection system and method, which comprises an integrated sensing module, a time synchronization module, a data processing module and a concrete quality diagnosis module, wherein the integrated sensing module comprises a plurality of sensors with fixed rigidity and calibrated spatial relations, the data of an acoustic sensor and an infrared thermal imager are synchronously collected, the time synchronization module is connected with the integrated sensing module, when a robot moves to a detection point, the time synchronization module sends out a synchronous pulse signal and simultaneously triggers the sensors in the integrated sensing module to collect data, the data processing module generates a multi-attribute fusion three-dimensional point cloud model from the data collected by all the sensors in the integrated sensing module, the concrete quality diagnosis module maps the multi-mode data to a three-dimensional space, extracts multi-channel characteristic parameters of the same spatial position and inputs the multi-channel characteristic parameters to a pre-trained cross-joint diagnosis depth learning model to obtain the defect type, severity and confidence of the position.

Inventors

CHEN YUANHONG
FANG TINGCHEN
FENG YU
HE HONGYU
GUO YUANHAO

Assignees

上海建工集团股份有限公司

Dates

Publication Date: 20260508
Application Date: 20251231

Claims (7)

1. A multi-modal perceived concrete construction quality detection system, comprising: The integrated sensing module comprises a high-definition camera, a multispectral camera, a laser scanner, a thermal infrared imager and an impact echo sensor, wherein the high-definition camera, the multispectral camera, the laser scanner, the thermal infrared imager and the impact echo sensor are rigidly fixed and have calibrated spatial relations, the laser scanner and the high-definition camera are used as cores for arrangement, an automatic flick excitation device and an acoustic sensor are arranged in the impact echo sensor, the flick excitation device is triggered and linked with a main synchronous signal, the transient thermal response is triggered by tapping, the defect contrast is enhanced, and the data of the acoustic sensor and the thermal infrared imager are synchronously acquired; the time synchronization module is connected with the integrated sensing module, and when the robot moves to a detection point, the time synchronization module sends out a synchronization pulse signal and simultaneously triggers all sensors in the integrated sensing module to acquire data; the data processing module is connected with the integrated sensing module, performs unified processing on data acquired by all sensors in the integrated sensing module, takes laser point clouds as unified space references, and maps multi-source 2D image data textures to the references to generate a multi-attribute fusion three-dimensional point cloud model; The concrete quality diagnosis module is connected with the data processing module and is used for acquiring space-time synchronous multi-mode data, mapping the multi-mode data to a three-dimensional space, extracting multi-channel characteristic parameters, multi-spectrum characteristics and surface temperature of the same space position, inputting the multi-channel characteristics to a pre-trained cross-mode joint diagnosis deep learning model, and obtaining the defect type, severity and confidence of the position.
2. The multi-modal perceived concrete construction quality inspection system of claim 1 wherein all of the sensors in the integrated sensor module are rigidly affixed to a precisely calibrated common support such that the optical centers of all of the sensors are aligned and have a fixed relative positional and attitude relationship.
3. The multi-mode sensing concrete construction quality detection system according to claim 1, wherein the lens axis of the thermal infrared imager is parallel to the antenna of the millimeter wave radar, so that the detection depth and the detection area can be referred to each other, and the detection depth and the detection area are used for distinguishing the superficial temperature anomaly from the deep structure anomaly.
4. The multi-modal perceived concrete construction quality detection system of claim 1, wherein the multi-channel characteristics of the same spatial location include RGB texture, multi-spectral water content index, temperature, three-dimensional curvature, and millimeter wave echo intensity underneath.
5. A method for detecting the construction quality of the multi-modal perceived concrete, characterized in that the multi-modal perceived concrete construction quality detection system according to any one of claims 1 to 4 is provided, the method comprising the steps of: Step S1, training a cross-mode joint diagnosis deep learning model; Step S2, combining the specific wave band ratio calculated by the multispectral camera with the slight temperature difference captured by the thermal infrared imager, and judging that the model is early water stain or seepage with high confidence when high water index and low temperature abnormality occur in a certain area at the same time; S3, if a surface crack is found, analyzing acoustic features and thermal image features of an area right below the surface crack by taking a unified coordinate as a reference, if the surface crack is found, the crack, a low-frequency peak of an impact echo spectrum and abnormal diffusion of a hot spot diagram after knocking occur, judging that the surface crack is caused by structural hollowness, and if the surface crack is only found, performing secondary plastering treatment on the surface shrinkage joint; step S4, providing the depth and the outline of the steel bar through a millimeter wave radar, reversely deducing the propagation path of millimeter radar wave by combining an accurate surface model provided by a three-dimensional laser point cloud, and improving the positioning accuracy of the inner cavity boundary and the steel bar size through a wave velocity tomography algorithm; And S5, inputting the multichannel characteristics into a pre-trained cross-mode joint diagnosis deep learning model to obtain the defect type, severity and confidence of the position.
6. The method of claim 5, wherein the concrete quality diagnostic module evaluates quality trends using a bayesian network dynamic quality evaluation and prediction algorithm, and specifically comprises: Firstly, collecting environmental temperature and humidity in real time, fusing defect probability output by a diagnosis model and age of a component; and secondly, constructing a Bayesian network, taking the environmental conditions and the current defect state as input nodes, and evaluating future quality evolution risks, so as to provide prospective enclosure decision support.
7. The method according to claim 5, wherein the data processing method comprises the steps of: firstly, inputting collected data, and synchronously collecting RGB images, multispectral images, infrared thermal images, depth maps, millimeter wave Lei Dadian clouds and laser point clouds by taking unified coordinates of a physical space as a reference; Establishing a unified reference, and taking the high-precision laser point cloud as a unified three-dimensional space coordinate system; Thirdly, registering data, namely accurately texture mapping data parameters carried by a two-dimensional image to a three-dimensional point cloud model based on camera calibration parameters, so that each three-dimensional point not only has coordinates, but also has color, multispectral characteristics and surface temperature attributes; And fourthly, generating a model, and generating a multi-attribute fused three-dimensional point cloud model, wherein the apparent, internal and geometric information of a certain position can be inquired at will on the model.

Description

Multi-mode sensing concrete construction quality detection system and method Technical Field The invention relates to the technical field of civil engineering, in particular to a multi-mode sensing concrete construction quality detection system and method. Background Concrete is a main component of a building structure, and the construction quality of the concrete directly influences the engineering quality and even the service life of the structure. At present, concrete quality detection mostly relies on the manual work to adopt single-point equipment (such as guiding rule, resiliometer, crack observation appearance) to detect, and detection efficiency is low, unable global coverage, and the testing result receives the subjective influence of inspector, unable global coverage, is difficult to discover the defect that inner structure exists. With the development of technology, some digitization technologies are applied to concrete quality detection, such as an image acquisition instrument, a laser scanner, an ultrasonic radar and the like, and at present, data formats, space-time standards and the like of different detection devices are not uniform, only single item detection can be realized, and correlation analysis is difficult. Therefore, how to provide a system and a method for detecting the construction quality of multi-mode sensing concrete is a technical problem that needs to be solved by those skilled in the art. Disclosure of Invention In order to solve the problems in the prior art, the invention provides a multi-mode sensing concrete construction quality detection system and method, which solve the technical problems in the current concrete detection process by synchronous triggering of multiple sensors, data fusion processing and artificial intelligent diagnosis and realize integrated, synchronized and accurate intelligent detection of the apparent mass, the internal mass, the geometric dimension and the environmental state of the concrete. The technical scheme of the multi-mode sensing concrete construction quality detection system and method is as follows: A multi-modal perceived concrete construction quality detection system, comprising: The integrated sensing module comprises a high-definition camera, a multispectral camera, a laser scanner, a thermal infrared imager and an impact echo sensor, wherein the high-definition camera, the multispectral camera, the laser scanner, the thermal infrared imager and the impact echo sensor are rigidly fixed and have calibrated spatial relations, the laser scanner and the high-definition camera are used as cores for arrangement, an automatic flick excitation device and an acoustic sensor are arranged in the impact echo sensor, the flick excitation device is triggered and linked with a main synchronous signal, the transient thermal response is triggered by tapping, the defect contrast is enhanced, and the data of the acoustic sensor and the thermal infrared imager are synchronously acquired; the time synchronization module is connected with the integrated sensing module, and when the robot moves to a detection point, the time synchronization module sends out a synchronization pulse signal and simultaneously triggers all sensors in the integrated sensing module to acquire data; the data processing module is connected with the integrated sensing module, performs unified processing on data acquired by all sensors in the integrated sensing module, takes laser point clouds as unified space references, and maps multi-source 2D image data textures to the references to generate a multi-attribute fusion three-dimensional point cloud model; The concrete quality diagnosis module is connected with the data processing module and is used for acquiring space-time synchronous multi-mode data, mapping the multi-mode data to a three-dimensional space, extracting multi-channel characteristic parameters, multi-spectrum characteristics and surface temperature of the same space position, inputting the multi-channel characteristics to a pre-trained cross-mode joint diagnosis deep learning model, and obtaining the defect type, severity and confidence of the position. Further, all the high-definition cameras, the multispectral cameras, the laser scanners, the thermal infrared imagers and the impact echo sensors in the integrated sensing module are rigidly fixed on a common bracket which is precisely calibrated, so that the optical centers of all the sensors are aligned and have fixed relative position and posture relation. Further, the lens axis of the thermal infrared imager is parallel to the antenna of the millimeter wave radar, so that the detection depth and the detection area of the thermal infrared imager can be referred to each other, and the thermal infrared imager is used for distinguishing the superficial temperature anomaly from the deep structure anomaly. Further, the multi-channel features of the same spatial location include RGB texture, multi-spectral water content