Engineering surveying and mapping pavement flatness detection device and method

Abstract

The invention discloses an engineering surveying and mapping road surface flatness detection device and method, and the device comprises a measurement vehicle which is internally provided with a vehicle-mounted industrial control computer. A laser array module is fixedly mounted at the front end of a chassis of the measuring vehicle, a multi-axis inertia compensation unit is mounted at the gravity center of the chassis of the measuring vehicle, a vibration vision fusion sensor group is mounted on the measuring vehicle, and a space-time alignment center module is integrated in the vehicle-mounted industrial personal computer; a multi-modal fusion processing module is also integrated in the on-board industrial personal computer; the on-board industrial personal computer is provided with a block chain evidence storage interface, and the block chain evidence storage interface communicates with the cloud server through the on-board 5G module; and the block chain evidence storage interface generates a unique hash value for each detection road section. The invention relates to the technical field of engineering surveying and mapping. According to the invention, the problems of low efficiency, limited precision, insufficient data processing and analysis capability and lack of a long-term data management and tracing mechanism in the traditional road surface flatness detection technology are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of engineering surveying and mapping, and in particular to a device and method for detecting road surface flatness in engineering surveying and mapping. Background Art

[0002] In the field of road engineering, road surface smoothness is a key indicator for measuring road quality and performance. It is not only directly related to driving comfort and safety, but also closely related to vehicle wear, fuel consumption and road service life. With the continuous growth of traffic flow and the continuous increase in vehicle speed, the demand for accurate and efficient detection of road surface smoothness has become more urgent. However, traditional road surface smoothness detection technology has many limitations:

[0003] Low detection efficiency: When using tools such as three-meter rulers and continuous roughness meters for detection, manual operation is required and the speed is slow. When faced with long-distance road detection tasks, it consumes a lot of manpower, material resources and time costs, and cannot meet the needs of large-area and high-frequency road detection under the rapid development of modern transportation.

[0004] Limited detection accuracy: This type of method is easily affected by human and environmental factors. For example, differences in manual measurement techniques and different weather and lighting conditions may lead to deviations in the judgment of measurement results, making it difficult to accurately obtain microscopic unevenness information on the road surface and unable to meet the requirements for refined road surface detection.

[0005] Insufficient data processing and analysis capabilities: Most of the data obtained by traditional testing methods rely on manual recording and simple calculation and analysis, which makes it difficult to achieve the fusion processing of multi-source data. Faced with massive amounts of test data, it is impossible to timely and comprehensively mine the road quality problems reflected behind the data, and it is also difficult to establish an effective road condition assessment system, which affects the scientific formulation of road maintenance and repair decisions.

[0006] Lack of long-term data management and traceability mechanism: The traditional storage method of inspection data is relatively scattered and irregular, and data security and integrity are difficult to guarantee. During the long-term use of roads, it is impossible to conveniently trace historical inspection data, which is not conducive to quality control and management of the entire life cycle of roads.

[0007] Under the current trend of smart transportation and digital construction, the existing road surface roughness detection technology can no longer meet the new development needs. There is an urgent need for an engineering surveying and mapping road surface roughness detection device and method to achieve efficient and high-precision road surface roughness detection, while also having strong data processing, analysis and long-term management capabilities to provide reliable technical support for the planning, construction and maintenance of road projects. Summary of the invention

[0008] In order to solve the problems of low efficiency, limited accuracy, insufficient data processing and analysis capabilities, and lack of long-term data management and traceability mechanism in traditional road surface smoothness detection technology, the purpose of the present invention is to provide an engineering surveying and mapping road surface smoothness detection device and method.

[0009] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: an engineering surveying and mapping road surface flatness detection device, comprising a measuring vehicle, in which an on-board industrial computer is installed for comprehensive data processing; a laser array module is fixedly installed at the front end of the chassis of the measuring vehicle, and the laser array module is fixedly installed at the front end of the chassis of the measuring vehicle, and comprises three laser emitting units distributed in a triangle;

[0010] A multi-axis inertia compensation unit is installed at the center of gravity of the chassis of the measuring vehicle;

[0011] The measuring vehicle is equipped with a vibration and vision fusion sensor group, which includes a piezoelectric vibration sensor arranged along the diagonal of the measuring vehicle and a polarization vision module installed on the roof;

[0012] The on-board industrial computer is integrated with a time-space alignment hub module, which is connected to the laser array module, multi-axis inertial compensation unit, piezoelectric vibration sensor and polarization vision module through the PCIE expansion slot.

[0013] A multi-modal fusion processing module is also integrated in the industrial computer. The PCIE expansion slot of the multi-modal fusion processing module is connected to the time-space alignment center to receive multi-source data aligned by the time-space alignment center.

[0014] The onboard industrial computer is equipped with a blockchain evidence storage interface, which communicates with the cloud server through the onboard 5G module;

[0015] The spatiotemporal alignment hub module performs three levels of spatiotemporal calibration: hardware clock synchronization, spatial calibration, and dynamic time warping;

[0016] The multimodal fusion processing module fuses laser curvature, vibration energy spectrum and visual crack features through a gated attention mechanism;

[0017] The blockchain evidence storage interface generates a unique hash value for each detection section.

[0018] Preferably, the laser array module is installed at the front bumper of the measuring vehicle, and the three laser emitting units are distributed in an equilateral triangle with a side length of 80 cm; the laser wavelengths emitted by the three laser emitting units are 1550nm, 1310nm and 980nm respectively; the emission frequencies of the three laser emitting units are 10kHz, 15kHz and 20kHz respectively.

[0019] Preferably, the multi-axis inertial compensation unit is embedded in the center of gravity of the chassis of the measurement vehicle and is integrated with a six-degree-of-freedom MEMS sensor for real-time output of the vehicle posture quaternion.

[0020] Preferably, the piezoelectric vibration sensor is installed on the four-wheel suspension bearing seat, with its axial direction consistent with the suspension movement direction, and the detection frequency band is 0.1-500Hz; an electric lifting bracket is installed on the top of the measuring vehicle, and the polarization vision module is installed on the telescopic end of the electric lifting bracket. The polarization vision module is equipped with a four-way rotating polarization filter.

[0021] Preferably, when the space-time alignment central module performs hardware clock synchronization, the IEEE1588 time protocol is used to align the clocks of each sensor; the space-time alignment central module performs spatial calibration to solve the transformation matrix from the laser point cloud to the visual coordinate system; dynamic time warping aligns non-uniformly sampled data streams by improving the DTW algorithm.

[0022] Preferably, the hardware implementation of the improved DTW algorithm includes a coarse alignment unit, a refinement unit and a post-processing unit. The coarse alignment unit downsamples to 1 / 4 resolution and calculates the initial path; the refinement unit searches for the optimal path within ±5 points in the neighborhood of the initial path; and the post-processing unit generates a continuous alignment sequence through cubic spline interpolation.

[0023] Preferably, the multimodal fusion processing module includes a laser feature encoder, a vibration feature encoder and a visual feature encoder; the laser feature encoder extracts curvature features; the vibration feature encoder generates an energy spectrum; and the visual feature encoder is used for visual feature generation.

[0024] A method for detecting road surface flatness in engineering surveying and mapping comprises the following steps:

[0025] S1, multi-source data collection:

[0026] a1, the frequency modulation signal is emitted by the laser array module, and the original elevation is calculated by the echo phase difference. The road elevation point cloud is , the road elevation is measured by triangulation :

[0027] ,in is the distance value measured by each laser unit;

[0028] a2, multi-axis inertia compensation unit calculates vehicle roll angle in real time , Pitch angle , and generate laser matrix dynamic compensation parameters:

[0029] ;

[0030] Output the vehicle pose quaternion at the same time ;

[0031] a3, piezoelectric vibration sensor collects acceleration data in the frequency band of 0.1-500Hz;

[0032] a4, polarized vision module capture Four-way polarization image ;

[0033] S2, the spatiotemporal alignment central module performs spatiotemporal alignment processing:

[0034] b1, hardware clock synchronization: align the clocks of each sensor;

[0035] b2, space calibration: unify the coordinate system through the Lie group SE (3) transformation matrix:

[0036] ,in ;

[0037] b3, dynamic time warping, for non-uniformly sampled data, the dynamic time warping DTW algorithm is improved, and the cumulative distance matrix is: , path search constraint window width , where T is the sequence length;

[0038] S3, multimodal fusion processing module performs multimodal feature fusion:

[0039] c1, laser feature extraction, local curvature ;

[0040] c2, vibration feature coding, generating 16 sub-band energy spectrum through wavelet packet decomposition;

[0041] c3, visual feature generation, DeepLabv3+ network outputs pixel-level crack probability map;

[0042] c4, gated attention fusion, weights each modality feature according to the weight:

[0043] Weight ;in is the laser mode eigenvector, is the vibration mode eigenvector, is the visual modality feature vector, is the weight matrix, is the attention score vector;

[0044] S4, flatness determination and decision output:

[0045] d1, calculate the flatness index , where L is the length of the detection section, m is the half-width of the window, and N is the number of sampling points;

[0046] d2, calculation of local defect index ;

[0047] d3, defect classification:

[0048] Level 1 defect: IRI>8 and LDI>0.8;

[0049] Second level defect: 6≦IRI≦8, or 0.6>LDI≧0.8;

[0050] Level 3 defect: IRI<6, and LDI≦0.6;

[0051] d4, generate blockchain evidence hash ;

[0052] d5, output maintenance decision.

[0053] Preferably, the maintenance decision includes a primary defect decision, a secondary defect decision and a tertiary defect decision.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] 1. The present invention adopts multi-sensor collaborative work. The laser array module obtains accurate elevation data through lasers of different wavelengths and frequencies. The multi-axis inertial compensation unit compensates for the influence of vehicle posture on measurement in real time. The piezoelectric vibration sensor and polarization vision module collect data from different dimensions. After spatiotemporal alignment and multimodal fusion processing, the road surface flatness detection accuracy is significantly improved, and tiny road surface defects can be accurately identified.

[0056] 2. In the present invention, the time-space alignment central module performs three-level time-space calibration, uses the time protocol to synchronize the clock, transforms the matrix to unify the coordinate system, improves the DTW algorithm to process non-uniformly sampled data, ensures accurate alignment of multi-source data, and lays the foundation for subsequent fusion analysis; the multimodal fusion processing module uses a gated attention mechanism to fuse laser, vibration and visual features, fully taps the data value, and comprehensively reflects the road conditions.

[0057] 3. The present invention provides a clear quantitative basis for pavement maintenance by calculating the flatness index and the local defect index and classifying defects according to set standards; the generated maintenance decisions are highly targeted and can reasonably arrange maintenance resources according to different defect levels, thereby improving maintenance efficiency and economy.

[0058] 4. In the present invention, the blockchain evidence storage interface generates a unique hash value for each detected road section, and uploads it to the cloud server in combination with the on-board 5G module to ensure that the detection data cannot be tampered with and is traceable, thereby improving data security and credibility, facilitating subsequent inquiries and quality traceability, and providing strong support for long-term road management. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0060] Figure 1 It is a schematic diagram of the road surface smoothness detection device system for engineering surveying and mapping;

[0061] Figure 2 It is a schematic diagram of the process of the engineering surveying and mapping road surface flatness detection method of the present invention. DETAILED DESCRIPTION

[0062] The following is a description of the implementation of the present invention by means of specific embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification.

[0063] See also Figure 1 to Figure 2 . It should be noted that the structures, proportions, sizes, etc. illustrated in the drawings in this specification are only used to match the contents disclosed in the specification for people familiar with this technology to understand and read, and are not used to limit the limiting conditions for the implementation of the present invention, so they have no substantial technical significance. Any modification of the structure, change in the proportional relationship, or adjustment of the size should still fall within the scope of the technical content disclosed by the present invention without affecting the effects and purposes that can be achieved by the present invention. At the same time, the terms such as "upper", "lower", "left", "right", "middle" and "one" quoted in this specification are only for the convenience of description, and are not used to limit the scope of the implementation of the present invention. Changes or adjustments in their relative relationships should also be regarded as the scope of the implementation of the present invention without substantially changing the technical content.

[0064] Embodiment 1, a road surface smoothness detection device for engineering surveying and mapping, includes a survey vehicle, and a laser array module is installed at the front bumper of the survey vehicle. The module includes three laser emitting units distributed in an equilateral triangle, and the side length is set to 80cm. The laser wavelengths emitted by the three laser emitting units are 1550nm, 1310nm and 980nm, respectively, and the emission frequencies are 10kHz, 15kHz and 20kHz, respectively. This design enables the laser array module to obtain rich road surface information based on lasers of different wavelengths and frequencies when emitting frequency modulated signals, accurately calculate the original elevation through the echo phase difference, and provide basic data for subsequent road surface smoothness analysis. At the same time, the dual wavelengths of 1550nm and 980nm eliminate temperature drift errors and adapt to extreme environments of -20℃ to 60℃.

[0065] The multi-axis inertial compensation unit is embedded in the center of gravity of the chassis of the measurement vehicle. The unit integrates a six-degree-of-freedom MEMS sensor and can output the vehicle's position quaternion in real time. During the driving process of the vehicle, the vehicle roll angle and pitch angle are calculated in real time, and the laser matrix dynamic compensation parameters are generated to effectively compensate for the error caused by the change of vehicle posture to the laser measurement elevation data, thereby improving the accuracy of the measurement.

[0066] The vibration vision fusion sensor group consists of a piezoelectric vibration sensor and a polarization vision module. The piezoelectric vibration sensor is installed on the four-wheel suspension bearing seat, with its axial direction consistent with the direction of suspension movement, and the detection frequency band is 0.1-500Hz. Such an installation position and detection frequency band setting can accurately collect vibration acceleration data caused by uneven road surface during vehicle driving. An electric lifting bracket is installed on the top of the measuring vehicle, and the polarization vision module is installed on the telescopic end of the electric lifting bracket and is equipped with a four-way rotating polarization filter. The electric lifting bracket can adjust the height of the polarization vision module according to actual detection needs. The four-way rotating polarization filter can capture four-way polarization images of 0°, 45°, 90°, and 135° to obtain more comprehensive road surface visual information, including defect features such as cracks.

[0067] The four-way polarizing filter eliminates rain, fog and reflection interference, greatly improving image contrast.

[0068] The on-board industrial computer installed in the measuring vehicle is the data processing core of the entire device. The on-board industrial computer is integrated with a space-time alignment hub module, which is connected to the laser array module, multi-axis inertial compensation unit, piezoelectric vibration sensor and polarization vision module through the PCIE expansion slot. At the same time, the on-board industrial computer is also integrated with a multimodal fusion processing module, which is connected to the space-time alignment hub through the PCIE expansion slot to receive multi-source data aligned by the space-time alignment hub. In addition, a blockchain evidence storage interface is set on the on-board industrial computer, which communicates with the cloud server through the on-board 5G module to realize blockchain evidence storage and remote transmission of detection data.

[0069] The spatiotemporal alignment central module performs three-level spatiotemporal calibration. In terms of hardware clock synchronization, the IEEE1588 time protocol is used to align the clocks of each sensor to ensure the time consistency of data collected by different sensors. Spatial calibration is achieved by solving the transformation matrix from the laser point cloud to the visual coordinate system. Specifically, the coordinate system is unified through the Lie group SE (3) transformation matrix to achieve a unified expression of data from different sensors in space. Dynamic time warping uses an improved DTW algorithm to align non-uniformly sampled data streams. Its hardware implementation includes a coarse alignment unit, a refinement unit, and a post-processing unit. The coarse alignment unit downsamples the data to 1 / 4 resolution and calculates the initial path; the refinement unit searches for the optimal path within ±5 points in the neighborhood of the initial path; the post-processing unit generates a continuous alignment sequence through cubic spline interpolation, which effectively solves the alignment problem of non-uniformly sampled data and provides an accurate data basis for subsequent multimodal fusion processing.

[0070] The multimodal fusion processing module includes a laser feature encoder, a vibration feature encoder, and a visual feature encoder. The laser feature encoder extracts local curvature features, calculates local curvature through a formula, and obtains the microscopic geometric features of the road surface. The vibration feature encoder generates a 16-subband energy spectrum through wavelet packet decomposition, analyzes the vibration energy distribution in different frequency bands, and reflects the impact of road surface flatness on vehicle vibration. The visual feature encoder uses the DeepLabv3+ network to output a pixel-level crack probability map to identify defects such as road cracks. Finally, the laser curvature, vibration energy spectrum, and visual crack features are fused through a gated attention mechanism, and the features of each modality are weighted according to the weight to improve the comprehensive judgment ability of the road condition.

[0071] The blockchain evidence storage interface generates a unique hash value for each test section. During the test process, each time a section of road is tested, the test data of that section is hashed to generate a unique hash value, which is then uploaded to the cloud server through the on-board 5G module for blockchain evidence storage. This method ensures that the test data cannot be tampered with and is traceable, providing a reliable data basis for subsequent road maintenance decisions and quality assessments.

[0072] Embodiment 2, a method for detecting the flatness of a road surface in engineering surveying and mapping, comprising the following steps:

[0073] S1, multi-source data collection:

[0074] a1, the frequency modulation signal is emitted by the laser array module, and the original elevation is calculated by the echo phase difference. The road elevation point cloud is , the road elevation is measured by triangulation :

[0075] ,in is the distance value measured by each laser unit;

[0076] a2, multi-axis inertia compensation unit calculates vehicle roll angle in real time , Pitch angle , and generate laser matrix dynamic compensation parameters:

[0077] ;

[0078] Output the vehicle pose quaternion at the same time ;

[0079] a3, piezoelectric vibration sensor collects acceleration data in the frequency band of 0.1-500Hz;

[0080] a4, polarized vision module capture Four-way polarization image ;

[0081] S2, the spatiotemporal alignment central module performs spatiotemporal alignment processing:

[0082] b1, hardware clock synchronization: align the clocks of each sensor;

[0083] b2, space calibration: unify the coordinate system through the Lie group SE (3) transformation matrix:

[0084] ,in ;

[0085] b3, dynamic time warping, for non-uniformly sampled data, the dynamic time warping DTW algorithm is improved, and the cumulative distance matrix is: , path search constraint window width , where T is the sequence length;

[0086] S3, multimodal fusion processing module performs multimodal feature fusion:

[0087] c1, laser feature extraction, local curvature ;

[0088] c2, vibration feature coding, generating 16 sub-band energy spectrum through wavelet packet decomposition;

[0089] c3, visual feature generation, DeepLabv3+ network outputs pixel-level crack probability map;

[0090] c4, gated attention fusion, weights each modality feature according to the weight:

[0091] Weight ;in is the laser mode eigenvector, is the vibration mode eigenvector, is the visual modality feature vector, is the weight matrix, is the attention score vector;

[0092] S4, flatness determination and decision output:

[0093] d1, calculate the flatness index , where L is the length of the detection section, m is the half-width of the window, and N is the number of sampling points;

[0094] d2, calculation of local defect index ;

[0095] d3, defect classification:

[0096] Level 1 defect: IRI>8 and LDI>0.8;

[0097] Second level defect: 6≦IRI≦8, or 0.6>LDI≧0.8;

[0098] Level 3 defect: IRI<6, and LDI≦0.6;

[0099] d4, generate blockchain evidence hash ;

[0100] d5, output maintenance decision.

[0101] Maintenance decisions include primary defect decisions, secondary defect decisions and tertiary defect decisions.

[0102] Level 1 defect decision: trigger drone re-inspection and generate a 3D defect model.

[0103] Secondary defect decision: automatically plan the milling path and output the repair volume.

[0104] Level 3 defect decision: Mark as annual maintenance area and generate material budget list.

[0105] Example 3, Level 1 defect scenario: When inspecting a highway, a detection vehicle finds severe rutting and deep cracks on a certain road section, and collects data:

[0106] Laser elevation sequence: [5.2, 7.8, 9.1, 6.3, 12.4, 8.9, 4.7, 10.2], maximum elevation difference ΔD_max = 12.4 - 4.7 = 7.7 mm.

[0107] Vibration signal: Impact energy spectrum peak: 82g·s>50g·s threshold; Main frequency band: 80-120Hz, corresponding to the rutting resonance frequency;

[0108] Visual image: crack pixel ratio: 22%>15% threshold; texture anisotropy index A_index=0.85>0.7 threshold;

[0109] but ;

[0110] ;

[0111] Then IRI=9.2>8.0, LDI=5.835>0.8, and the defect level is a first-level defect.

[0112] First level defect decision:

[0113] Trigger the drone re-inspection and generate a three-dimensional disease model: rutting depth 9.1mm, crack length 3.2m.

[0114] Automatically push emergency maintenance work orders, close road sections and start hot regeneration repairs.

[0115] The blockchain evidence is marked as an "emergency event" and the maintenance unit must sign for it within the specified time.

[0116] Example 4, secondary defect scenario: The detection of urban secondary roads found that there were moderate rutting and surface cracks in some sections of the road, and the data was collected:

[0117] Laser elevation sequence: [3.5, 4.2, 5.8, 6.1, 7.3, 5.9, 4.7, 6.5], maximum elevation difference ΔD_max = 7.3 - 3.5 = 3.8 mm.

[0118] Vibration signal: Impact energy spectrum peak: 42g·s, between the 20-50g·s threshold range; Main frequency band: 30-60Hz, corresponding to the rutting resonance frequency;

[0119] Visual image: crack pixel ratio: 9%, between 5% and 15% threshold range; texture anisotropy index A_index = 0.62;

[0120] but ;

[0121] ;

[0122] Then IRI=6.7∈[6.0,8.0], LDI=4.196∈[0.6,0.8], and the defect level is a secondary defect.

[0123] Secondary defect decision: automatically plan the milling path and output the repair volume.

[0124] Blockchain evidence storage:

[0125] The hash mark is a yellow alert and the maintenance unit needs to handle it within the specified time.

[0126] Smart contracts automatically generate electronic work orders and assign them to the nearest maintenance fleet.

[0127] Example 5, Level 3 defect scenario: Routine inspection of urban main roads found uniform wear on the road surface and no structural damage. Data collection:

[0128] Laser elevation sequence: [2.1, 1.9, 2.3, 2.0, 1.8, 2.2, 2.1, 1.9], maximum elevation difference ΔD_max = 2.3 - 1.8 = 0.5mm.

[0129] Vibration signal: Impact energy spectrum peak: 18g·s<20g·s threshold; Main frequency band: 5-15Hz, normal driving vibration range;

[0130] Visual image: crack pixel ratio: 0.8%<1% threshold; texture anisotropy index A_index=0.35;

[0131] but ;

[0132] ;

[0133] Then IRI=4.3<6.0, LDI=2.63<0.6, and the defect level is level three.

[0134] Level 3 defect decision: Mark as annual maintenance area, include it in the annual micro-surfacing plan, and generate a material budget list.

[0135] Blockchain evidence storage:

[0136] The test results are recorded in the blockchain evidence, triggering an automatic re-inspection reminder at a specified time.

[0137] Example 6, technical logic comparison table of Example 3, Example 4 and Example 5:

[0138] Determination Dimension Level 1 defect Secondary defects Level 3 Defect IRI Range >8.0 m / km 6.0-8.0 m / km <6.0 m / km LDI Threshold >0.8 0.6-0.8 ≤0.6 Vibration energy peak >50g·s 20-50g·s <20g·s Crack area ratio >15% 5%-15% <5% Response time ≤2 hours ≤15 days Annual Plan Restoration Technology Thermal regeneration + structural layer repair Surface milling + resurfacing Micro surfacing prevention Blockchain token color red yellow blue

[0139] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A road surface flatness detection device for engineering surveying and mapping, comprising a measuring vehicle, wherein the measuring vehicle is equipped with an on-board industrial computer for comprehensive data processing; characterized in that: A laser array module is fixedly installed at the front end of the chassis of the measuring vehicle. The laser array module is fixedly installed at the front end of the chassis of the measuring vehicle and includes three laser emitting units distributed in a triangle; A multi-axis inertia compensation unit is installed at the center of gravity of the chassis of the measuring vehicle; The measuring vehicle is equipped with a vibration and vision fusion sensor group, which includes a piezoelectric vibration sensor arranged along the diagonal of the measuring vehicle and a polarization vision module installed on the roof; The vehicle-mounted industrial computer is integrated with a time-space alignment central module, and the time-space alignment central module is respectively connected to the laser array module, the multi-axis inertial compensation unit, the piezoelectric vibration sensor and the polarization vision module through the PCIE expansion slot; The onboard industrial computer is also integrated with a multimodal fusion processing module, and the PCIE expansion slot of the multimodal fusion processing module is connected to the spatiotemporal alignment hub, and is used to receive multi-source data aligned by the spatiotemporal alignment hub; The onboard industrial computer is provided with a blockchain evidence storage interface, and the blockchain evidence storage interface communicates with the cloud server through the onboard 5G module; The spatiotemporal alignment hub module performs three levels of spatiotemporal calibration: hardware clock synchronization, spatial calibration, and dynamic time warping; The multimodal fusion processing module fuses laser curvature, vibration energy spectrum and visual crack features through a gated attention mechanism; The blockchain evidence storage interface generates a unique hash value for each detected road section.

2. The engineering surveying and mapping road surface flatness detection device according to claim 1, characterized in that: The laser array module is installed at the front bumper of the measuring vehicle. The three laser emitting units are distributed in an equilateral triangle with a side length of 80 cm. The laser wavelengths emitted by the three laser emitting units are 1550nm, 1310nm and 980nm respectively. The emission frequencies of the three laser emitting units are 10kHz, 15kHz and 20kHz respectively.

3. The engineering surveying and mapping road surface flatness detection device according to claim 1, characterized in that: The multi-axis inertial compensation unit is embedded in the center of gravity of the chassis of the measurement vehicle and is integrated with a six-degree-of-freedom MEMS sensor for real-time output of the vehicle posture quaternion.

4. The engineering surveying and mapping road surface flatness detection device according to claim 1, characterized in that: The piezoelectric vibration sensor is installed on the four-wheel suspension bearing seat, its axial direction is consistent with the suspension movement direction, and the detection frequency band is 0.1-500Hz; an electric lifting bracket is installed on the top of the measuring vehicle, and the polarization vision module is installed on the telescopic end of the electric lifting bracket. The polarization vision module is equipped with a four-way rotating polarization filter.

5. The engineering surveying and mapping road surface flatness detection device according to claim 1, characterized in that: When the space-time alignment central module performs hardware clock synchronization, the IEEE1588 time protocol is used to align the clocks of each sensor; the space-time alignment central module performs spatial calibration to solve the transformation matrix from the laser point cloud to the visual coordinate system; the dynamic time warping aligns non-uniform sampling data streams by improving the DTW algorithm.

6. The device for detecting the flatness of a road surface for engineering surveying and mapping according to claim 5, characterized in that: The hardware implementation of the improved DTW algorithm includes a coarse alignment unit, a refinement unit and a post-processing unit. The coarse alignment unit downsamples to 1 / 4 resolution and calculates the initial path; the refinement unit searches for the optimal path within ±5 points in the neighborhood of the initial path; and the post-processing unit generates a continuous alignment sequence through cubic spline interpolation.

7. The engineering surveying and mapping road surface flatness detection device according to claim 1, characterized in that: The multimodal fusion processing module includes a laser feature encoder, a vibration feature encoder and a visual feature encoder; the laser feature encoder extracts curvature features; the vibration feature encoder generates a band energy spectrum; and the visual feature encoder is used for visual feature generation.

8. A method for detecting the flatness of a road surface in engineering surveying and mapping, characterized in that: The engineering surveying and mapping road surface flatness detection device according to any one of claims 1 to 7 comprises the following steps: S1, multi-source data collection: a1, the frequency modulation signal is emitted by the laser array module, and the original elevation is calculated by the echo phase difference. The road elevation point cloud is , the road elevation is measured by triangulation : ,in is the distance value measured by each laser unit; a2, multi-axis inertia compensation unit calculates vehicle roll angle in real time , Pitch angle , and generate laser matrix dynamic compensation parameters: ； Output the vehicle pose quaternion at the same time ; a3, piezoelectric vibration sensor collects acceleration data in the frequency band of 0.1-500Hz; a4, polarized vision module capture Four-way polarization image ; S2, the spatiotemporal alignment central module performs spatiotemporal alignment processing: b1, hardware clock synchronization: align the clocks of each sensor; b2, space calibration: unify the coordinate system through the Lie group SE (3) transformation matrix: ,in ; b3, dynamic time warping, for non-uniformly sampled data, the dynamic time warping DTW algorithm is improved, and the cumulative distance matrix is: , path search constraint window width , where T is the sequence length; S3, multimodal fusion processing module performs multimodal feature fusion: c1, laser feature extraction, local curvature ; c2, vibration feature coding, generating 16 sub-band energy spectrum through wavelet packet decomposition; c3, visual feature generation, DeepLabv3+ network outputs pixel-level crack probability map; c4, gated attention fusion, weights each modality feature according to the weight: Weight ;in is the laser mode eigenvector, is the vibration mode eigenvector, is the visual modality feature vector, is the weight matrix, is the attention score vector; S4, flatness determination and decision output: d1, calculate the flatness index , where L is the length of the detection section, m is the half-width of the window, and N is the number of sampling points; d2, calculation of local defect index ; d3, defect classification: Level 1 defect: IRI>8 and LDI>0.8; Second level defect: 6≦IRI≦8, or 0.6>LDI≧0.8; Level 3 defect: IRI<6, and LDI≦0.6; d4, generate blockchain evidence hash ; d5, output maintenance decision.

9. The method for detecting the smoothness of a road surface in engineering surveying and mapping according to claim 8, characterized in that: The maintenance decision includes a primary defect decision, a secondary defect decision and a tertiary defect decision.

Images