Roadbed settlement detection method and system based on image data processing

Abstract

The invention provides a roadbed settlement detection method and system based on image data processing, and relates to the field of data processing.The method comprises the steps that settlement related information, at least including geological information, roadbed filler information and operation load information, of a roadbed to be detected is obtained; determining a settlement risk value of the to-be-detected roadbed in the current detection period according to the settlement related information of the to-be-detected roadbed and the meteorological information of the current detection period; determining a laser detection scheme of the current detection period according to the settlement risk value of the to-be-detected roadbed in the current detection period; collecting laser detection information of the roadbed to be detected according to the laser detection scheme of the current detection period; judging whether to perform image acquisition according to the laser detection information; if it is judged that image collection is carried out, the real-time image of the to-be-detected roadbed is collected, and the settlement detection result of the to-be-detected roadbed is generated based on the real-time image of the to-be-detected roadbed. The roadbed settlement detection method has the advantages that the roadbed settlement detection efficiency is improved, and the roadbed settlement detection cost is reduced.

Description

Technical Field

[0001] The invention relates to the field of data processing, and in particular to a roadbed settlement detection method and system based on image data processing. Background Art

[0002] Roadbed settlement refers to the phenomenon of ground subsidence caused by the consolidation or compaction of the roadbed soil layer during road or railway construction. It usually occurs within a period of time after the roadbed construction, and may also gradually intensify during long-term use. Roadbed settlement can be manifested as uniform settlement and uneven settlement according to different settlement forms. Among them, the damage caused by uneven settlement is more prominent. Once it occurs, especially after the highway is opened to traffic, it will cause varying degrees of damage to the pavement structure, directly affecting the service life of the entire highway.

[0003] In the prior art, leveling is a commonly used method for observing roadbed settlement. Its principle is to use a precision level to measure the settlement of the roadbed at different times. Before the roadbed is filled, it is necessary to set observation piles on the leveling points on both sides of the roadbed and record the initial elevation. As the roadbed is filled, the observation piles are measured regularly to obtain the settlement data of the roadbed. The advantages of the leveling method are high accuracy and good stability, and it is suitable for various terrains and construction conditions. However, the observation requires a lot of time and manpower investment, and the cost is high. The layered settlement meter measurement method is a method specifically used to observe the layered settlement of the roadbed. Before the roadbed is filled, it is necessary to set observation points in the roadbed soil layers at different depths and record the initial elevation. As the roadbed is filled, the observation points are measured regularly using the layered settlement meter to obtain the settlement data of each soil layer. The advantage of the layered settlement meter measurement method is that it can accurately observe the layered settlement of the roadbed, which is suitable for large-scale projects and complex terrains. However, due to the high equipment cost, the cost is relatively high.

[0004] Therefore, it is necessary to provide a roadbed settlement detection method and system based on image data processing to improve the efficiency of roadbed settlement detection and reduce the cost of roadbed settlement detection. Summary of the invention

[0005] The present invention provides a roadbed settlement detection method based on image data processing, comprising: obtaining settlement-related information of a roadbed to be detected, wherein the settlement-related information at least includes geological information, roadbed filling material information and operating load information; obtaining meteorological information of a current detection period; determining a settlement risk value of the roadbed to be detected in the current detection period according to the settlement-related information of the roadbed to be detected and the meteorological information of the current detection period; determining a laser detection scheme of the current detection period according to the settlement risk value of the roadbed to be detected in the current detection period; collecting laser detection information of the roadbed to be detected according to the laser detection scheme of the current detection period; judging whether to perform image acquisition according to the laser detection information; if it is determined to perform image acquisition, collecting a real-time image of the roadbed to be detected, and generating a settlement detection result of the roadbed to be detected based on the real-time image of the roadbed to be detected.

[0006] Furthermore, meteorological information of the current detection period is obtained, including: obtaining geological information, roadbed filling information, operating load information, meteorological information and settlement information of multiple sample roadbeds; determining key meteorological factors corresponding to the roadbed to be detected based on the geological information, roadbed filling information and settlement information of multiple sample roadbeds and the geological information and roadbed filling information of the roadbed to be detected; based on the key meteorological factors corresponding to the roadbed to be detected, obtaining the meteorological information of the current detection period.

[0007] Furthermore, based on the geological information, roadbed filling information and settlement information of multiple sample roadbeds and the geological information and roadbed filling information of the roadbed to be detected, the key meteorological factors corresponding to the roadbed to be detected are determined, including: calculating the feature similarity of any two sample roadbeds based on the geological information and roadbed filling information of any two sample roadbeds; dividing the multiple sample roadbeds into multiple roadbed units based on the feature similarity of any two sample roadbeds; for each roadbed unit, calculating the settlement influence coefficient of the roadbed unit corresponding to each meteorological factor based on the meteorological information and settlement information of the multiple sample roadbeds included in the roadbed unit, and determining the key meteorological factors corresponding to the roadbed unit based on the settlement influence coefficient of the roadbed unit corresponding to each meteorological factor; determining the target roadbed unit from the multiple roadbed units based on the geological information and roadbed filling information of the roadbed to be detected; and determining the key meteorological factors corresponding to the roadbed to be detected based on the key meteorological factors corresponding to the target roadbed unit.

[0008] Furthermore, based on the settlement-related information of the roadbed to be inspected and the meteorological information of the current inspection period, the settlement risk value of the roadbed to be inspected in the current inspection period is determined, including: for each roadbed unit, based on the geological information, roadbed filling information, operating load information, meteorological information and settlement information of multiple sample roadbeds included in the roadbed unit, a training sample is generated, and a risk prediction model corresponding to the roadbed unit is established; and the settlement risk value of the roadbed to be inspected in the current inspection period is determined by the risk prediction model corresponding to the target roadbed unit according to the settlement-related information of the roadbed to be inspected and the meteorological information of the current inspection period.

[0009] Furthermore, according to the settlement risk value of the roadbed to be detected in the current detection cycle, a laser detection scheme for the current detection cycle is determined, including: according to the settlement risk value of the roadbed to be detected in the current detection cycle, the laser detection frequency and the single laser detection point density are determined, wherein the laser detection scheme for the current detection cycle at least includes the laser detection frequency and the single laser detection point density.

[0010] Furthermore, according to the laser detection scheme of the current detection cycle, the laser detection information of the roadbed to be detected is collected, including: S11, according to the density of single laser detection points, the ranging values ​​of multiple laser detection points of the roadbed to be detected are collected; S12, according to the ranging values ​​of multiple laser detection points of the roadbed to be detected, the height difference value is calculated; S13, according to the height difference value, the density of the supplementary laser detection points is determined; S14, according to the density of the supplementary laser detection points, the ranging values ​​of multiple laser detection points of the roadbed to be detected are supplemented; S15, according to the supplementary ranging values ​​of multiple laser detection points of the roadbed to be detected, the height difference value is updated; S16, according to the height difference value before the update and the height difference value after the update corresponding to the current iteration, it is judged whether it is necessary to supplement the detection again, if so, execute S13, if not, complete the collection.

[0011] Furthermore, based on the laser detection information, it is determined whether to perform image acquisition, including: grouping multiple laser detection points of the roadbed to be detected according to the planar position of each laser detection point to determine the point group; for each point group, determining the roadbed area corresponding to the point group according to the planar position of each laser detection point included in the point group; for each roadbed area, determining the settlement information of the roadbed area according to the ranging value of each laser detection point included in the point group; and determining whether to perform image acquisition based on the settlement information and area of ​​each roadbed area.

[0012] Further, the real-time image of the subgrade to be detected is a depth map; based on the real-time image of the subgrade to be detected and the laser detection information, a settlement detection result of the subgrade to be detected is generated, including: obtaining an unsettled image of the subgrade to be detected; based on the laser detection information, generating a settlement difference image of the subgrade to be detected according to the real-time image and the unsettled image of the subgrade to be detected; and generating a settlement detection result of the subgrade to be detected according to the settlement difference image of the subgrade to be detected.

[0013] Further, generating a settlement difference image of the subgrade to be detected according to the real-time image and the unsettled image of the subgrade to be detected includes: extracting a first region of interest image from the unsettled image of the subgrade to be detected according to the settlement information of each subgrade region; extracting a second region of interest image from the real-time image of the subgrade to be detected according to the settlement information of each subgrade region; for each pixel in the first region of interest image, calculating the difference between the gray value of the pixel and the gray value of the pixel at the corresponding position in the second region of interest image; and generating a settlement difference image of the subgrade to be detected according to the difference of the gray values corresponding to each pixel in the first region of interest image.

[0014] The present invention provides a subgrade settlement detection system based on image data processing, including: an information acquisition module for acquiring settlement-related information of the subgrade to be detected, where the settlement-related information at least includes geological information, subgrade filling information, and operating load information; the information acquisition module is further used to acquire meteorological information of the current detection period; a risk prediction module for determining a settlement risk value of the subgrade to be detected in the current detection period according to the settlement-related information of the subgrade to be detected and the meteorological information of the current detection period; a laser detection module for determining a laser detection scheme for the current detection period according to the settlement risk value of the subgrade to be detected in the current detection period, and collecting laser detection information of the subgrade to be detected according to the laser detection scheme for the current detection period; an image acquisition module for determining whether to perform image acquisition according to the laser detection information, and if it is determined to perform image acquisition, acquiring a real-time image of the subgrade to be detected; and a settlement analysis module for generating a settlement detection result of the subgrade to be detected based on the real-time image and the laser detection information of the subgrade to be detected.

[0015] Compared with the prior art, the subgrade settlement detection method and system based on image data processing provided by the present invention at least have the following beneficial effects:

[0016] 1. Combining geological information, roadbed filling information, operating load information and meteorological information, it can more comprehensively assess the settlement risk of the roadbed. Through laser detection schemes and real-time image acquisition, the slight deformation and settlement of the roadbed can be captured, thereby improving the accuracy of detection. The laser detection scheme is dynamically adjusted according to the settlement risk value to avoid unnecessary comprehensive detection, thereby saving time and resources. The automated process of image acquisition and processing reduces manual intervention and improves the overall detection efficiency. Through accurate risk assessment and dynamic adjustment of the detection scheme, frequent comprehensive detection is avoided, thereby reducing the detection cost. The automated and intelligent detection process reduces labor costs and improves the overall economic benefits. It can customize the detection scheme according to the characteristics and environmental conditions of different roadbeds, and has strong adaptability. Whether it is a mountainous area with complex geological conditions or a plain area, it can provide effective settlement detection means. Through image processing and data analysis, the settlement detection results can be presented in an intuitive way, which is easy for decision makers to understand. The analysis results based on the data can provide a scientific basis for subsequent maintenance plans and engineering improvements.

[0017] 2. By analyzing the geology, filler, load and settlement information of multiple sample roadbeds and comparing them with the roadbed to be tested, the key meteorological factors that have the greatest impact on the settlement of the roadbed to be tested can be determined. This avoids blindly collecting all meteorological information and improves the pertinence and efficiency of information collection. The determination of key meteorological factors is based on the actual settlement data and meteorological information of the sample roadbed, so it can more accurately reflect the impact of meteorological conditions on roadbed settlement. This helps to more accurately predict the settlement risk of the roadbed to be tested in the current detection cycle. By grouping the sample roadbed into roadbed units and calculating the settlement influence coefficient of the meteorological factor for each unit, it is possible to identify which meteorological factors have a significant impact on the roadbed settlement. This helps to give priority to these key meteorological factors during the detection process, thereby optimizing the allocation of detection resources. After determining the key meteorological factors, relevant meteorological information for the current detection cycle can be collected more specifically, reducing unnecessary information collection work. At the same time, settlement prediction based on key meteorological factors can generate test results faster and improve detection efficiency. Key meteorological factors can be dynamically determined based on the geological and filler characteristics of different roadbeds and actual settlement data. This enhances adaptability, making it applicable to roadbed settlement detection in different geographical environments and climatic conditions. It can make decisions based on actual data, such as determining key meteorological factors, optimizing detection plans, etc. This helps to achieve data-driven decision-making and improve the scientificity and accuracy of decision-making.

[0018] 3. Dynamically adjust the laser detection frequency and the density of single laser detection points according to the settlement risk value to avoid unnecessary frequent detection or waste of resources caused by too dense detection points. Increase the detection frequency and point density in areas with higher settlement risks to ensure that small settlement changes can be captured. By iteratively supplementing detection points and updating height difference values, the settlement of the roadbed can be more accurately reflected, reducing errors caused by insufficient or uneven distribution of detection points. By iteratively supplementing detection points multiple times, ensure that sufficient quantity and density of laser detection data are collected to provide complete data support for subsequent analysis. Each iteration performs additional detection based on the previous height difference value, gradually approaching the actual settlement situation and improving the accuracy of the data.

[0019] 4. The depth map can reflect the depth information of the surface of the object, that is, the distance from each point to the observation point. In the roadbed settlement detection, the depth map can intuitively show the settlement of the roadbed surface, making the detection results more intuitive and easy to understand. By comparing the non-settled image and the real-time image, the settlement differential image is generated, which can clearly show the settlement area and settlement degree of the roadbed. This graphical representation helps engineers to identify and understand the settlement more quickly. By calculating the gray value difference between the corresponding pixels in the first area of ​​interest image and the second area of ​​interest image, the degree of settlement can be quantified. This method is more accurate and reliable than traditional visual inspection. Through image processing and data analysis technology, an automated and intelligent detection process is realized. From the acquisition of the depth map to the generation of the settlement differential image, and then to the output of the settlement detection results, the whole process can be completed efficiently, reducing manual intervention and errors. Compared with the traditional settlement detection method, this method can generate detection results faster, reducing detection time and cost. At the same time, since the detection process is more automated and intelligent, it also reduces the requirements for the skills and experience of the detection personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] This specification will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0021] Figure 1 is a schematic flow chart of a roadbed settlement detection method based on image data processing according to some embodiments of this specification;

[0022] Figure 2 It is a schematic diagram of a process of collecting laser detection information of a roadbed to be detected according to some embodiments of this specification;

[0023] Figure 3 It is a module schematic diagram of a roadbed settlement detection system based on image data processing according to some embodiments of this specification. DETAILED DESCRIPTION

[0024] In order to more clearly illustrate the technical solutions of the embodiments of this specification, the following is a brief introduction to the drawings required for the description of the embodiments. Obviously, the drawings described below are only some examples or embodiments of this specification. For ordinary technicians in this field, this specification can also be applied to other similar scenarios based on these drawings without creative work. Unless it is obvious from the language environment or otherwise explained, the same reference numerals in the figures represent the same structure or operation.

[0025] Figure 1 is a flow chart of a roadbed settlement detection method based on image data processing according to some embodiments of this specification, such as Figure 1 As shown, the roadbed settlement detection method based on image data processing may include the following steps.

[0026] Step 110, obtaining settlement related information of the roadbed to be detected.

[0027] In some embodiments, the settlement-related information includes at least geological information, roadbed filling information, and operational load information.

[0028] Geological information mainly involves the soil type, soil layer structure, geological structure, groundwater level, and physical and mechanical properties of rock and soil. This information is crucial for evaluating the stability of the foundation and predicting the amount of settlement. For example, in a highway project, the foundation soil is mainly low-plastic clay, which has high compressibility and is therefore prone to settlement when subjected to load. At the same time, the soil layer structure is also an important consideration, such as the stratification of the soil layer, the thickness and properties of each soil layer, which will affect the occurrence and development of settlement. The stability of the geological structure has a direct impact on the foundation settlement. For example, in fault zones or karst development areas, the stability of the foundation will be seriously threatened. In addition, changes in the groundwater level will also affect the settlement of the foundation. When the groundwater level drops, the effective stress in the soil increases, resulting in soil compression and settlement.

[0029] Geological information can be obtained using physical instruments and methods, such as seismic exploration, geoelectric exploration, gravity exploration, magnetic exploration, etc.

[0030] The roadbed filler information mainly includes the type, properties, compaction degree and filling method of the filler. This information is of great significance for controlling the settlement of the roadbed. Different types of fillers have different physical and mechanical properties, such as density, compressibility, shear strength, etc. For example, using highly compressible soil as filler will cause a large settlement of the roadbed. Therefore, when selecting fillers, it is necessary to consider their properties and their impact on settlement. The compaction degree of the filler is an important factor affecting the settlement of the roadbed. The higher the compaction degree, the lower the compressibility of the filler, and the settlement is correspondingly reduced. In addition, the filling method will also affect the settlement. For example, filling in layers and compacting layer by layer can effectively control the settlement.

[0031] Surveys can be conducted at the roadbed construction site to understand the source, type, and properties of the filler. Filler samples can be collected for laboratory testing to obtain the filler's density, compressibility, shear strength, and other physical and mechanical properties.

[0032] Operational load information mainly involves the type, size, and duration of loads that roads or buildings bear during use. This information is of great significance for evaluating the long-term stability of the foundation and predicting the amount of settlement. Different types of loads have different effects on the foundation. For example, static loads (such as the deadweight of the building) and dynamic loads (such as traffic loads) have different effects on the settlement of the foundation. At the same time, the size of the load also directly affects the amount of settlement. The greater the load, the stronger the compression on the foundation, and the amount of settlement increases accordingly. The duration of the load will also affect the settlement of the foundation. Foundations that bear large loads for a long time are prone to consolidation settlement, while short-term loads may lead to elastic settlement. Therefore, the duration of the load needs to be considered when evaluating foundation settlement.

[0033] Conduct actual monitoring during the operation of roads or buildings to record the actual load conditions. Sensors, monitoring equipment and other means can be used for real-time monitoring and data recording. For roads and other transportation facilities, conduct traffic flow surveys to obtain traffic load information. The survey content includes vehicle type, number, speed and load size.

[0034] Step 120, obtaining meteorological information of the current detection period.

[0035] In some embodiments, step 120 specifically includes:

[0036] Acquire geological information, roadbed filling information, operating load information, meteorological information and settlement information of multiple sample roadbeds, wherein the settlement information may include settlement characteristics, such as settlement area, maximum settlement depth, etc., and the meteorological information may include characteristic values ​​(such as maximum value, minimum value, variance, etc.) of multiple meteorological factors (such as temperature, humidity, rainfall, etc.) in the environment where the sample roadbed is located;

[0037] Determine the key meteorological factors corresponding to the roadbed to be tested according to the geological information, roadbed filling material information and settlement information of multiple sample roadbeds and the geological information and roadbed filling material information of the roadbed to be tested;

[0038] According to the key meteorological factors corresponding to the roadbed to be detected, the meteorological information of the current detection period is obtained, wherein the meteorological information may include the characteristic value (eg, maximum value, minimum value, variance, etc.) corresponding to each key meteorological factor.

[0039] In some embodiments, based on the geological information, roadbed filling material information and settlement information of multiple sample roadbeds and the geological information and roadbed filling material information of the roadbed to be tested, the key meteorological factors corresponding to the roadbed to be tested are determined, including:

[0040] According to the geological information and roadbed filling information of any two sample roadbeds, the characteristic similarity of any two sample roadbeds is calculated;

[0041] According to the feature similarity of any two sample roadbeds, the multiple sample roadbeds are divided into multiple roadbed units. For example, a clustering algorithm (for example, K-means clustering, hierarchical clustering, mean shift clustering, etc.) is used to divide the multiple sample roadbeds into multiple roadbed units according to the feature similarity of any two sample roadbeds.

[0042] For each roadbed unit, according to the meteorological information and settlement information of multiple sample roadbeds included in the roadbed unit, the settlement influence coefficient of each meteorological factor corresponding to the roadbed unit is calculated, and according to the settlement influence coefficient of each meteorological factor corresponding to the roadbed unit, the key meteorological factor corresponding to the roadbed unit is determined, for example, the meteorological factor with a settlement influence coefficient greater than a settlement influence coefficient threshold is used as the key meteorological factor corresponding to the roadbed unit;

[0043] Based on the geological information of the roadbed to be detected and the roadbed filling material information, a target roadbed unit is determined from multiple roadbed units;

[0044] Based on the key meteorological factors corresponding to the target roadbed unit, the key meteorological factors corresponding to the roadbed to be detected are determined. For example, the key meteorological factors corresponding to the target roadbed unit are unioned to obtain the key meteorological factors corresponding to the roadbed to be detected.

[0045] Specifically, key geological features are extracted from geological information, such as soil type code, soil bearing capacity grade, groundwater level, etc. These features are digitized to facilitate subsequent calculations. Key filler features are extracted from roadbed filler information, such as filler type code, filler density, filler strength, etc. Similarly, these features are also digitized.

[0046] According to the geological key features extracted from the geological information of the two sample roadbeds and the filler key characteristics extracted from the roadbed filler information, the feature similarity of the two sample roadbeds is calculated based on the similarity calculation method (Euclidean distance, cosine similarity, correlation coefficient (such as Pearson correlation coefficient), Jaccard similarity coefficient, etc.).

[0047] The settlement influence coefficient of the roadbed unit corresponding to the meteorological factor can be calculated according to the following formula:

[0048]

[0049] Among them, ψ i is the settlement influence coefficient of the roadbed unit corresponding to the i-th meteorological factor, ψ (i,(m,n)) is the sub-sedimentation influence coefficient of the nth eigenvalue of the ith meteorological factor on the mth settlement feature, N is the total number of eigenvalues ​​of the meteorological factor, M is the total number of settlement features, and F (i,(e,n)) is the nth eigenvalue of the ith meteorological factor corresponding to the eth sample roadbed, F (e,m) is the mth settlement feature corresponding to the eth sample roadbed, and E is the total number of sample roadbeds.

[0050] For each roadbed unit, the geological key features and filler key features of each sample roadbed included in the roadbed unit can be averaged to obtain the central geological key features and central filler key features of the roadbed unit.

[0051] The geological key features and filler key features of the roadbed to be detected can be extracted from the geological information of the roadbed to be detected and the roadbed filler information. The feature similarity between the roadbed to be detected and the roadbed unit is calculated based on the central geological key features and central filler key features of the roadbed unit and the geological key features and filler key features of the roadbed to be detected. The roadbed unit with a feature similarity greater than a feature similarity threshold is taken as the target roadbed unit.

[0052] Step 130, determining the settlement risk value of the roadbed to be detected in the current detection period according to the settlement related information of the roadbed to be detected and the meteorological information of the current detection period.

[0053] In some embodiments, step 130 specifically includes:

[0054] For each roadbed unit, based on geological information, roadbed filling information, operating load information, meteorological information and settlement information of multiple sample roadbeds included in the roadbed unit, a training sample is generated, and a risk prediction model corresponding to the roadbed unit is established, wherein the risk prediction model may be a convolutional neural network model;

[0055] The settlement risk value of the roadbed to be detected in the current detection period is determined by the risk prediction model corresponding to the target roadbed unit according to the settlement related information of the roadbed to be detected and the meteorological information of the current detection period.

[0056] Step 140, determining a laser detection scheme for the current detection cycle according to the settlement risk value of the roadbed to be detected in the current detection cycle.

[0057] In some embodiments, step 140 specifically includes:

[0058] According to the settlement risk value of the roadbed to be inspected in the current inspection cycle, the laser inspection frequency and the single laser inspection point density are determined, wherein the laser inspection scheme of the current inspection cycle at least includes the laser inspection frequency and the single laser inspection point density.

[0059] Specifically, the laser detection frequency and the single laser detection point density can be determined based on the following formula according to the settlement risk value of the roadbed to be detected in the current detection cycle:

[0060]

[0061] Among them, f T is the laser detection frequency of the current detection cycle, R T is the settlement risk value of the current detection period, R 0 is the preset settlement risk value, f 0 is the preset laser detection frequency, ρ T is the single laser detection point density of the current detection cycle, ρ 01 is the preset single laser detection point density, [] is the rounding operation.

[0062] Step 150, collecting laser detection information of the roadbed to be detected according to the laser detection scheme of the current detection cycle.

[0063] Figure 2 is a schematic diagram of a process for collecting laser detection information of a roadbed to be detected according to some embodiments of this specification, such as Figure 2 As shown, in some embodiments, step 150 includes:

[0064] S11, collecting ranging values ​​of multiple laser detection points of the roadbed to be detected according to the density of single laser detection points;

[0065] S12, calculating the height difference value according to the ranging values ​​of multiple laser detection points of the roadbed to be detected;

[0066] S13, determining the density of the supplementary laser detection points according to the height difference value;

[0067] S14, according to the density of the supplemented laser detection points, additionally collect the ranging values ​​of multiple laser detection points of the roadbed to be detected;

[0068] S15, updating the height difference value according to the supplementary collected ranging values ​​of multiple laser detection points of the roadbed to be detected;

[0069] S16. Determine whether additional detection is needed based on the height difference value before and after the update corresponding to the current iteration. If so, execute S13. If not, complete the collection.

[0070] Specifically, the height difference value can be calculated according to the following formula:

[0071]

[0072] Among them, D t is the updated height difference value corresponding to the t-th iteration, Q is the total number of laser detection points collected corresponding to the t-th iteration, H q is the distance measurement value of the qth laser detection point collected corresponding to the tth iteration, H p is the distance measurement value of the p-th laser detection point collected corresponding to the t-th iteration.

[0073] The density of the additional laser detection points can be determined according to the following formula:

[0074]

[0075] Among them, ρ t+1 is the density of the supplementary laser detection points corresponding to the t+1th iteration, D t-1 is the updated height difference value corresponding to the t-1th iteration, △D 0 is the preset difference, ρ 02 The density of laser detection points added to the preset.

[0076] The distance measurement values ​​of all laser detection points collected in the current iteration can be substituted into the calculation formula of the height difference value to update the height difference value.

[0077] If the difference between the height difference value before updating and the height difference value after updating corresponding to the current iteration is less than the difference threshold, it is determined that no additional detection is required.

[0078] Step 160: Determine whether to perform image acquisition based on the laser detection information.

[0079] In some embodiments, step 160 specifically includes:

[0080] According to the plane position of each laser detection point, multiple laser detection points of the roadbed to be detected are grouped to determine the point group;

[0081] For each point group, according to the plane position of each laser detection point included in the point group, determine the roadbed area corresponding to the point group;

[0082] For each roadbed area, the settlement information of the roadbed area is determined according to the ranging value of each laser detection point included in the point group;

[0083] Determine whether to perform image acquisition based on the settlement information and area of ​​each roadbed area.

[0084] Specifically, for any two laser detection points, the planar distance between the two laser detection points is calculated according to the planar positions of the two laser detection points. A clustering algorithm is used to group multiple laser detection points of the roadbed to be detected according to the planar distance between the two laser detection points to determine the point group.

[0085] For each roadbed area, the mean distance value and the variance of the distance value of the roadbed area are calculated according to the distance value of each laser detection point included in the point group, wherein the settlement information of the roadbed area may include the mean distance value and the variance of the distance value of the roadbed area.

[0086] When the mean of the ranging values ​​of the roadbed area is greater than the mean threshold of the ranging values ​​and / or the variance of the ranging values ​​is greater than the variance threshold of the ranging values, the roadbed area is regarded as a suspected subsidence area. When the area of ​​the suspected subsidence area is greater than the area threshold, the suspected subsidence area is regarded as a subsidence area. When there is at least one subsidence area, it is determined to perform image acquisition.

[0087] Step 170: If it is determined to perform image acquisition, a real-time image of the roadbed to be detected is acquired, and a settlement detection result of the roadbed to be detected is generated based on the real-time image of the roadbed to be detected and the laser detection information.

[0088] In some embodiments, the real-time image of the roadbed to be inspected is a depth map. A depth map is a special image in which the value of each pixel represents the depth or distance information of the pixel in three-dimensional space. A depth map is usually represented in the form of a grayscale image, in which the grayscale value of the pixel corresponds to the distance from each point in the scene to the observer. The higher the grayscale value, the farther the distance; the lower the grayscale value, the closer the distance.

[0089] In some embodiments, based on the real-time image of the roadbed to be detected and the laser detection information, generating the settlement detection result of the roadbed to be detected includes:

[0090] Acquire an unsettled image of the roadbed to be detected, wherein the unsettled image of the roadbed to be detected may be a depth map of the roadbed to be detected when no settlement occurs after the roadbed to be detected is built;

[0091] Based on the laser detection information, a settlement differential image of the roadbed to be detected is generated according to the real-time image and the non-settled image of the roadbed to be detected;

[0092] The settlement detection result of the roadbed to be detected is generated according to the settlement differential image of the roadbed to be detected.

[0093] In some embodiments, based on the laser detection information, generating a settlement differential image of the roadbed to be detected according to the real-time image and the non-settled image of the roadbed to be detected includes:

[0094] Extracting a first region of interest image from a non-settled image of the roadbed to be detected according to the settlement information of each roadbed area, wherein the first region of interest image may be an image of a region corresponding to a settlement area determined according to the settlement information of each roadbed area in the non-settled image of the roadbed to be detected;

[0095] Extracting a second region of interest image from the real-time image of the roadbed to be detected according to the settlement information of each roadbed area, wherein the second region of interest image may be an image of an area corresponding to the settlement area determined according to the settlement information of each roadbed area in the real-time image of the roadbed to be detected;

[0096] For each pixel in the first region of interest image, calculating the difference between the grayscale value of the pixel and the grayscale value of the pixel at the corresponding position in the second region of interest image;

[0097] According to the difference in grayscale values ​​corresponding to each pixel in the first area of ​​interest image, a settlement differential image of the roadbed to be inspected is generated, wherein the grayscale value of the pixel in the settlement differential image is the difference between the grayscale value of the pixel at the corresponding position in the first area of ​​interest image and the grayscale value of the pixel at the corresponding position in the second area of ​​interest image.

[0098] Specifically, for each subsidence area, the subsidence depth mean and the subsidence uniformity may be calculated according to the subsidence difference image corresponding to the subsidence area.

[0099] For example, the sedimentation depth mean and sedimentation uniformity can be calculated from the sedimentation difference image corresponding to the sedimentation area according to the following formula:

[0100]

[0101] Among them, μ k is the mean settlement depth of the kth settlement area, G (k,h) is the gray value of the hth pixel of the sedimentation differential image corresponding to the kth sedimentation area, H is the total number of pixels included in the sedimentation differential image corresponding to the kth sedimentation area, α k is the settlement uniformity of the kth settlement area, γ is a preset parameter, and γ is greater than 0.

[0102] The settlement detection result of the roadbed to be detected may include the mean settlement depth and settlement uniformity of each settlement area.

[0103] Figure 3 is a schematic diagram of a module of a roadbed settlement detection system based on image data processing according to some embodiments of this specification, such as Figure 3 As shown, the roadbed settlement detection system based on image data processing may include an information acquisition module, a risk prediction module, a laser detection module, an image acquisition module and a settlement analysis module.

[0104] An information acquisition module is used to acquire settlement-related information of the roadbed to be detected, wherein the settlement-related information at least includes geological information, roadbed filling material information and operating load information;

[0105] The information acquisition module is also used to obtain meteorological information of the current detection period;

[0106] The risk prediction module is used to determine the settlement risk value of the roadbed to be detected in the current detection period according to the settlement related information of the roadbed to be detected and the meteorological information of the current detection period;

[0107] The laser detection module is used to determine the laser detection scheme of the current detection cycle according to the settlement risk value of the roadbed to be detected in the current detection cycle, and collect laser detection information of the roadbed to be detected according to the laser detection scheme of the current detection cycle;

[0108] An image acquisition module is used to determine whether to perform image acquisition based on the laser detection information. If it is determined that image acquisition is to be performed, a real-time image of the roadbed to be detected is acquired;

[0109] The settlement analysis module is used to generate settlement detection results of the roadbed to be detected based on the real-time image of the roadbed to be detected and laser detection information.

[0110] The roadbed settlement detection system based on image data processing can be used to execute the roadbed settlement detection method based on image data processing, which will not be described in detail here.

[0111] Finally, it should be understood that the embodiments described in this specification are only used to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, as an example and not a limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly introduced and described in this specification.

Claims

1. A roadbed settlement detection method based on image data processing, characterized in that: include: Acquire settlement-related information of the roadbed to be detected, wherein the settlement-related information at least includes geological information, roadbed filling material information and operating load information; Get the meteorological information of the current detection period; Determine the settlement risk value of the roadbed to be detected in the current detection period according to the settlement related information of the roadbed to be detected and the meteorological information of the current detection period; Determine the laser detection scheme for the current detection cycle according to the settlement risk value of the roadbed to be detected in the current detection cycle; According to the laser detection scheme of the current detection cycle, collect laser detection information of the roadbed to be detected; Determine whether to perform image acquisition based on laser detection information; If it is determined to perform image acquisition, a real-time image of the roadbed to be detected is acquired, and a settlement detection result of the roadbed to be detected is generated based on the real-time image of the roadbed to be detected and the laser detection information.

2. The roadbed settlement detection method based on image data processing according to claim 1 is characterized in that: Get the meteorological information of the current detection period, including: Obtain geological information, roadbed filling information, operating load information, meteorological information and settlement information of multiple sample roadbeds; Determine the key meteorological factors corresponding to the roadbed to be tested according to the geological information, roadbed filling material information and settlement information of multiple sample roadbeds and the geological information and roadbed filling material information of the roadbed to be tested; According to the key meteorological factors corresponding to the roadbed to be inspected, the meteorological information of the current inspection period is obtained.

3. The roadbed settlement detection method based on image data processing according to claim 2 is characterized in that: According to the geological information, roadbed filling material information and settlement information of multiple sample roadbeds and the geological information and roadbed filling material information of the roadbed to be tested, the key meteorological factors corresponding to the roadbed to be tested are determined, including: According to the geological information and roadbed filling information of any two sample roadbeds, the characteristic similarity of any two sample roadbeds is calculated; According to the feature similarity of any two sample roadbeds, the plurality of sample roadbeds are divided into a plurality of roadbed units; For each roadbed unit, according to the meteorological information and settlement information of multiple sample roadbeds included in the roadbed unit, the settlement influence coefficient of each meteorological factor corresponding to the roadbed unit is calculated, and according to the settlement influence coefficient of each meteorological factor corresponding to the roadbed unit, the key meteorological factor corresponding to the roadbed unit is determined; Determine a target roadbed unit from the plurality of roadbed units based on geological information and roadbed filling material information of the roadbed to be detected; Based on the key meteorological factors corresponding to the target roadbed unit, the key meteorological factors corresponding to the roadbed to be tested are determined.

4. The roadbed settlement detection method based on image data processing according to claim 3 is characterized in that: Determining the settlement risk value of the roadbed to be detected in the current detection period according to the settlement related information of the roadbed to be detected and the meteorological information of the current detection period, including: For each roadbed unit, based on the geological information, roadbed filling information, operating load information, meteorological information and settlement information of multiple sample roadbeds included in the roadbed unit, a training sample is generated, and a risk prediction model corresponding to the roadbed unit is established; The settlement risk value of the roadbed to be detected in the current detection period is determined by the risk prediction model corresponding to the target roadbed unit according to the settlement related information of the roadbed to be detected and the meteorological information of the current detection period.

5. The roadbed settlement detection method based on image data processing according to any one of claims 1 to 4, characterized in that: According to the settlement risk value of the roadbed to be inspected in the current inspection cycle, the laser inspection plan for the current inspection cycle is determined, including: According to the settlement risk value of the roadbed to be detected in the current detection cycle, the laser detection frequency and the single laser detection point density are determined, wherein the laser detection scheme of the current detection cycle at least includes the laser detection frequency and the single laser detection point density.

6. The roadbed settlement detection method based on image data processing according to claim 5 is characterized in that: According to the laser detection scheme of the current detection cycle, the laser detection information of the roadbed to be detected is collected, including: S11, collecting ranging values ​​of multiple laser detection points of the roadbed to be detected according to the density of single laser detection points; S12, calculating the height difference value according to the ranging values ​​of multiple laser detection points of the roadbed to be detected; S13, determining the density of the supplementary laser detection points according to the height difference value; S14, according to the density of the supplemented laser detection points, additionally collect the ranging values ​​of multiple laser detection points of the roadbed to be detected; S15, updating the height difference value according to the supplementary collected ranging values ​​of multiple laser detection points of the roadbed to be detected; S16. Determine whether additional detection is needed based on the height difference value before and after the update corresponding to the current iteration. If so, execute S13. If not, complete the collection.

7. The roadbed settlement detection method based on image data processing according to claim 6 is characterized in that: According to the laser detection information, determine whether to perform image acquisition, including: According to the plane position of each laser detection point, multiple laser detection points of the roadbed to be detected are grouped to determine the point group; For each point group, according to the plane position of each laser detection point included in the point group, determine the roadbed area corresponding to the point group; For each roadbed area, the settlement information of the roadbed area is determined according to the ranging value of each laser detection point included in the point group; Determine whether to perform image acquisition based on the settlement information and area of ​​each roadbed area.

8. The roadbed settlement detection method based on image data processing according to claim 7 is characterized in that: The real-time image of the roadbed to be detected is a depth map; Based on the real-time image and laser detection information of the roadbed to be detected, the settlement detection results of the roadbed to be detected are generated, including: Acquire an unsettled image of the roadbed to be inspected; Based on the laser detection information, a settlement differential image of the roadbed to be detected is generated according to the real-time image and the non-settled image of the roadbed to be detected; The settlement detection result of the roadbed to be detected is generated according to the settlement differential image of the roadbed to be detected.

9. The roadbed settlement detection method based on image data processing according to claim 8 is characterized in that: Based on the laser detection information, a settlement differential image of the roadbed to be detected is generated according to the real-time image and the non-settled image of the roadbed to be detected, including: Extracting a first region of interest image from a non-settled image of the roadbed to be detected according to the settlement information of each roadbed area; Extracting a second region of interest image from the real-time image of the roadbed to be detected according to the settlement information of each roadbed area; For each pixel in the first region of interest image, calculating the difference between the grayscale value of the pixel and the grayscale value of the pixel at the corresponding position in the second region of interest image; A settlement difference image of the roadbed to be detected is generated according to the difference of the grayscale value corresponding to each pixel in the first region of interest image.

10. The roadbed settlement detection system based on image data processing is characterized by: The method for detecting roadbed settlement based on image data processing according to any one of claims 1 to 9 comprises: An information acquisition module, used to acquire settlement-related information of the roadbed to be detected, wherein the settlement-related information at least includes geological information, roadbed filling material information and operating load information; The information acquisition module is also used to obtain meteorological information of the current detection period; A risk prediction module, used to determine the settlement risk value of the roadbed to be detected in the current detection period according to the settlement related information of the roadbed to be detected and the meteorological information of the current detection period; The laser detection module is used to determine the laser detection scheme of the current detection cycle according to the settlement risk value of the roadbed to be detected in the current detection cycle, and collect laser detection information of the roadbed to be detected according to the laser detection scheme of the current detection cycle; An image acquisition module is used to determine whether to perform image acquisition based on the laser detection information. If it is determined that image acquisition is to be performed, a real-time image of the roadbed to be detected is acquired; The settlement analysis module is used to generate settlement detection results of the roadbed to be detected based on the real-time image of the roadbed to be detected and laser detection information.

Images