A method for extracting road surface potholes and a vehicle-mounted road surface pothole detection system

Abstract

The application discloses a kind of road surface pit extraction method and vehicle-mounted road surface pit detection system, the method includes: the laser center line extraction of road surface infrared structured light image is carried out, obtains one-dimensional profile data;One-dimensional profile data is carried out Gaussian smoothing processing, obtains road surface reference line;One-dimensional profile data is subtracted from road surface reference line, obtains the feature response of each point;Based on feature response, the instantaneous energy feature of each point is calculated using Teager-Kaiser energy operator;Based on one-dimensional profile data, the profile difference feature of each point is extracted using first-order local difference operator;The instantaneous energy feature of each point and profile difference feature are combined to generate enhanced signal;The peak point of enhanced signal is used as the candidate center point of pit, and each candidate center point is used as starting point, according to the preset boundary expansion condition, the boundary is expanded point by point to left and right sides, until the preset boundary expansion condition is no longer satisfied, stop, to determine the preliminary boundary of pit.The application can realize all-day road surface pit detection.

Description

Technical Field

[0001] This invention relates to the field of road safety detection technology, specifically to a method for extracting road potholes and a vehicle-mounted road pothole detection system, which is suitable for vehicle-mounted dynamic detection scenarios. Background Technology

[0002] With the rapid development of the transportation industry, the safe operation of airport runways and highways faces increasingly severe challenges. The presence of potholes can lead to serious safety accidents, such as tire blowouts or loss of control collisions on highways. While existing pothole detection technologies have seen some application, they still have significant limitations and urgently need improvement. In highway scenarios, the main monitoring methods (such as driver visual inspection and vehicle-mounted radar) have large blind spots, and sensor performance degrades under extreme environments (such as high temperatures and strong sunlight), resulting in a high false alarm rate.

[0003] Existing methods for detecting road potholes can be broadly categorized as follows: 1. Methods based on radar detection combined with photoelectric sensors. These methods primarily detect potholes and determine their location by transmitting and receiving radio signals, and then use image processing technology for real-time automatic detection and identification. This type of method involves a large equipment system and has high maintenance costs. 2. Methods based on photoelectric sensors. Although the cost is lower, they are greatly affected by lighting and weather conditions, making it difficult to meet all-weather processing requirements. 3. Existing laser detection equipment is mostly used for static road surface smoothness measurement and lacks the ability to reconstruct 3D point clouds and provide real-time alarms in dynamic vehicle scenarios. Summary of the Invention

[0004] The main objective of this invention is to provide a method for extracting road potholes and a vehicle-mounted road pothole detection system to achieve all-day road pothole detection.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: In a first aspect, the present invention provides a method for extracting road surface potholes, the method comprising: Laser centerline extraction is performed on the infrared structured light image of the road surface to obtain one-dimensional contour data; High-resolution one-dimensional contour data The road surface baseline is obtained by smoothing the surface. Subtract the one-dimensional profile data from the road surface baseline to obtain the characteristic response of each point; Based on the characteristic response, the instantaneous energy characteristics of each point are calculated using the Teager-Kaiser energy operator; Based on one-dimensional contour data, the contour difference features of each point are extracted using a first-order local difference operator. By combining the instantaneous energy characteristics and contour difference characteristics of each point, an enhanced signal is generated; The peak point of the enhanced signal is used as the candidate center point of the pit. Starting from each candidate center point, the boundary is expanded point by point to the left and right sides according to the preset boundary expansion conditions until the preset boundary expansion conditions are no longer met, so as to determine the preliminary boundary of the pit. The preset boundary expansion conditions include: the feature response is greater than zero, or the absolute value of the contour difference feature is less than the preset local difference threshold, or the absolute value of the instantaneous energy feature is greater than the preset energy threshold.

[0006] Following the above technical solution, the first-order local difference operator calculates the contour difference features of each point according to the following formula: ; In the formula, For the contour data value of the j-th point, For preset length, Let be the contour difference feature of the j-th point.

[0007] Following the above technical solution, the method further includes: Median filtering is applied to one-dimensional contour data to obtain smoothed data. Based on the obtained smoothed data, the contour difference features of each point are extracted using a first-order local difference operator.

[0008] Following the above technical solution, the enhanced signal is calculated according to the following formula: ; In the formula, For the contour difference feature of the j-th point, The instantaneous energy characteristics of the j-th point, Let be the enhanced signal value at point j.

[0009] Following the above technical solution, the method further includes: When the absolute value of the instantaneous energy characteristic is less than the preset energy threshold, the instantaneous energy characteristic of that point is set to zero.

[0010] Following the above technical solution, the method further includes: When the absolute value of the contour difference feature is less than the preset local difference threshold, the contour difference feature of that point is set to zero.

[0011] Following the above technical solution, the method further includes: When the width of the initial boundary of the pit is greater than the preset width threshold, the maximum value of the feature response within the initial boundary range of the pit is used as the reference point. Within the preset search range to the left and right of the reference point, the left minimum point and right minimum point of the feature response are found again, and the left minimum point and right minimum point are updated as the starting boundary and ending boundary of the pit.

[0012] Following the above technical solution, the method further includes: The pit depth is determined based on the initial and final boundaries of the pit, combined with the corresponding feature responses: ; In the formula, As the starting boundary, To terminate the boundary, The characteristic response of the corresponding point, The depth of the pit; For each pit, the overall severity is calculated based on the pit depth and the corresponding contour difference features and instantaneous energy features: ; In the formula, The contour difference feature term represents the sum of the contour difference features of all points within the boundary of the pit / groove. For instantaneous energy characteristic term, it represents the sum of instantaneous energy characteristics of all points within the boundary of the pit; α , β , γ These are the weighting coefficients; For overall severity; The median absolute deviation is calculated based on the overall severity of all potholes. ; An adaptive threshold is set based on the absolute deviation of the median. : ; in, This represents the average severity of all potholes. k >0 indicates an adjustable parameter; Based on adaptive threshold Filter out pits whose overall severity exceeds the adaptive threshold.

[0013] Following the above technical solution, the method further includes: The above steps are repeated for multiple consecutive frames of infrared structured light images of the road surface. The pothole boundaries extracted from each frame are combined in frame order to generate a three-dimensional point cloud model of the road surface, and the spatial location and geometric parameters of the potholes are output. The geometric parameters include length, width, depth, area and volume.

[0014] Secondly, the present invention also provides a vehicle-mounted road pothole detection system, the system comprising: The image acquisition module is used to emit structured light onto the road surface and acquire infrared structured light images of the road surface; The vehicle positioning module is used to acquire the vehicle's location information and associate the location information with each frame of the road surface infrared structured light image; The image processing module is used to execute the road surface pothole extraction method described in the first aspect, to extract the pothole boundaries of each frame of the road surface infrared structured light image, and to combine the pothole boundaries extracted from each frame in the frame order with the corresponding position information to generate two-dimensional contour data of the road surface and potholes. The data parsing module is used to parse two-dimensional contour data and construct a three-dimensional point cloud model of the road surface.

[0015] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects: This invention extracts one-dimensional contour data from infrared structured light images of the road surface and obtains a road surface baseline using Gaussian smoothing. It then calculates feature response, instantaneous energy features, and contour difference features to generate an enhanced signal. Based on the peak points of the enhanced signal, candidate pothole centers are determined. Finally, using boundary expansion conditions—including feature response greater than zero, absolute value of contour difference features less than a threshold, and absolute value of instantaneous energy features greater than a threshold—the initial pothole boundary is obtained by expanding point by point. This scheme achieves real-time pothole localization in dynamic vehicle scenarios, effectively suppresses random noise, exhibits high sensitivity to shallow and small potholes, and is unaffected by ambient light, providing all-weather detection capability.

[0016] This invention addresses the issue of wide pits by using the maximum value of the feature response as a reference point within the initial boundary range, and then searching for the minimum value of the feature response within a preset search range to its left and right, thus updating the initial and final boundaries of the pit. This boundary correction mechanism solves the problem of boundary shrinkage caused by insufficient initial boundary search range for wide pits, significantly improving the detection accuracy and boundary positioning reliability of large-sized pits.

[0017] This invention constructs a comprehensive pothole severity index by fusing the sum of pothole depth, contour difference features, and instantaneous energy features. An adaptive threshold is set based on the median absolute deviation of the comprehensive pothole severity of all potholes to filter potholes with a comprehensive severity exceeding the threshold. This effectively distinguishes between significant potholes that require maintenance and minor road defects that do not require treatment, reducing the false alarm rate and improving the practicality and robustness of the system. Attached Figure Description

[0018] Figure 1 This is a schematic flowchart of a road surface pothole extraction method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a vehicle-mounted road pothole detection system according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the laser centerline extraction result according to an embodiment of the present invention; Figure 4 This is a schematic diagram of one-dimensional contour data according to an embodiment of the present invention; Figure 5This is a schematic diagram of the pit extraction result of one-dimensional contour data according to an embodiment of the present invention; Figure 6 This is a schematic diagram of the original point cloud of road surface potholes according to an embodiment of the present invention; Figure 7 This is a schematic diagram of the road surface pothole extraction results according to an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. All other embodiments obtained by those skilled in the art based on the embodiments provided by this invention without inventive effort are within the scope of protection of this invention.

[0020] Obviously, the accompanying drawings described below are merely some examples or embodiments of the present invention. Those skilled in the art can apply the present invention to other similar scenarios based on these drawings without any inventive effort. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this invention, modifications to design, manufacturing, or production based on the technical content disclosed in this invention are merely conventional technical means and should not be construed as insufficient disclosure of the present invention.

[0021] In this invention, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this invention may be combined with other embodiments without conflict.

[0022] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "a," "an," "an," "the," and similar words used in this invention do not indicate quantity limitation and may indicate singular or plural. The terms "comprising," "including," "having," and any variations thereof used in this invention are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units not listed, or may include other steps or units inherent to these processes, methods, products, or devices. The terms "connected," "linked," "coupled," and similar words used in this invention are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "A plurality" used in this invention refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships may exist; for example, "A and / or B" can represent: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following objects have an "or" relationship. The terms "first," "second," and "third" used in this invention are merely to distinguish similar objects and do not represent a specific ordering of the objects.

[0023] This invention provides a method for extracting road potholes and a vehicle-mounted road pothole detection system, which can realize real-time detection and positioning of road potholes around the clock, and is not easily affected by the lighting environment. It has high detection accuracy and improves the accuracy of road pothole detection.

[0024] Example 1: like Figure 1 As shown, this embodiment provides a method for extracting road potholes, including the following steps: S110: Extract the laser centerline from the infrared structured light image of the road surface to obtain one-dimensional contour data.

[0025] Specifically, the laser centerline is extracted from the acquired infrared structured light image of the road surface, such as... Figure 3 As shown, the one-dimensional contour data TempData corresponding to each frame of the image is obtained. Missing values ​​in TempData are then filled in using linear interpolation, ensuring that the data length of TempData equals the pixel width of the image. In practical applications, the image size can be 1920×100, so the filled-in one-dimensional contour data contains 1920 sampling points. Figure 4 Only sampling points in the range of 800-1920 are shown.

[0026] S120: Gaussian smoothing is applied to the one-dimensional contour data to obtain the road surface baseline.

[0027] For the one-dimensional contour data TempData, use a standard deviation of [size missing]. σ b Gaussian smoothing is performed using a Gaussian kernel to obtain the road surface baseline. B x It is used to depict the overall trend of the road surface. The formula for calculating Gaussian smoothing is: ; Among them, standard deviation σ b The preferred value range is 50 to 200, and the specific value can be adjusted according to the road surface smoothness.

[0028] S130: Subtract the one-dimensional profile data from the road surface baseline to obtain the characteristic response of each point.

[0029] Compare the original one-dimensional contour data TempData with the road surface baseline B x Calculate the characteristic response by subtracting point by point. S x Its expression is: ; This characteristic response S x It is used to characterize the degree of deviation of local potholes on the road surface from the overall road surface reference.

[0030] S140: Based on the characteristic response, the instantaneous energy characteristics of each point are calculated using the Teager-Kaiser energy operator.

[0031] The Teager-Kaiser Energy Operator (TKEO) was used to analyze the characteristic response. S x Calculate instantaneous energy characteristics : ; In the formula, The instantaneous energy characteristics of the j-th point, Let be the characteristic response of the j-th point.

[0032] This energy operator can capture the instantaneous change characteristics of the pit profile signal, thereby improving the sensitivity to shallow or small pits.

[0033] Furthermore, to improve the accuracy of subsequent pit extraction, minute fluctuations can be suppressed. Specifically, when the instantaneous energy characteristics... The absolute value is less than the preset energy threshold. δ T When the instantaneous energy characteristic of that point is set to zero, the influence of non-significant local changes is eliminated, and its formula can be expressed as: ; In the formula, δ T The preset energy threshold can be set to 1.0 in practical applications.

[0034] S150: Based on one-dimensional contour data, the contour difference features of each point are extracted using a first-order local difference operator.

[0035] First, median filtering is applied to the one-dimensional contour data to obtain smoothed data. To suppress random noise, a first-order local difference operator is then constructed, and the contour difference features of each sampling point are calculated according to the following formula. D x : ; In the formula, The median-filtered contour data value of the j-th sampling point, where L is the preset model length. Let be the contour difference feature of the j-th sampling point.

[0036] Furthermore, to eliminate the interference of noise on subsequent boundary detection, it is necessary to suppress small changes. Specifically, when the absolute value of the contour difference feature... Less than the preset local difference threshold δ D When the contour difference feature of that point is set to zero, its formula can be expressed as: ; In the formula, δ D As a local difference threshold, in practical applications δ D Version 2.0 is acceptable.

[0037] S160: Combine the instantaneous energy characteristics and contour difference characteristics of each point to generate an enhanced signal.

[0038] To enhance the pitting features, this invention combines a first-order local difference operator with an energy operator to generate an enhanced signal. Specifically, the contour difference features of each sampling point are... With instantaneous energy characteristics Multiplying these together generates the enhanced signal, enhanced_signal, which is calculated using the following formula: ; In the formula, For the contour difference feature of the j-th point, The instantaneous energy characteristics of the j-th point, Let be the enhanced signal value at point j.

[0039] This enhanced signal simultaneously reflects the intensity of local changes and instantaneous energy characteristics of the crater, providing a basis for subsequent crater center point detection and boundary expansion.

[0040] S170: The peak point of the enhanced signal is used as the candidate center point of the pit, and the boundary is expanded point by point to the left and right sides according to the preset boundary expansion conditions, starting from each candidate center point, until the preset boundary expansion conditions are no longer met, so as to determine the preliminary boundary of the pit.

[0041] Specifically, the peak points of the enhanced signal (enhanced_signal) are identified, and these peak points are used as candidate center points for the pits. P c For each candidate center point P c The boundary is expanded point by point to the left and right sides, and the expansion stops when the preset boundary expansion condition is no longer met. The preset boundary expansion condition is: ; In other words, the boundary continues to expand outward as long as the current sampling point meets any of the above conditions; when none of the three conditions are met, the expansion stops, thus determining the preliminary left and right boundaries of the pit. This method ensures that the boundary covers the local variation area of ​​the pit while suppressing random noise interference.

[0042] Thus far, this embodiment has completed the extraction of the preliminary boundaries of potholes in a single-frame infrared structured light image of a road surface through steps S110 to S170, as follows: Figure 5 As shown.

[0043] It should be noted that the step numbers do not represent a fixed execution order. For example, step S150 can be executed before step S120, as long as the input data required for each step is ready. Those skilled in the art can adjust the order of the steps according to actual needs, and such adjustments are all within the scope of protection of this invention.

[0044] Example 2: Based on Example 1, this embodiment further modifies the initial boundary of the pit to improve the detection accuracy of wide pits.

[0045] For the initial boundary of the pit determined in Example 1, calculate its width. W ,when W Greater than the preset width threshold W t When this happens, perform the following boundary correction operations: First, within the initial boundary range of the pit, we search for characteristic responses. S x The maximum value point is used as a reference point. X c; Then, at the reference point X c The preset search range on the left R Within this range, the minimum point of the characteristic response is re-found, serving as the corrected left boundary. l new Similarly, the preset search range is on the right. R Within, find the minimum point of the characteristic response, which serves as the corrected right boundary. r new : ; The corrected left boundary l new and right boundary r new These serve as the actual starting and ending boundaries of the pit, respectively.

[0046] By correcting the boundaries as described above, the problem of shrinkage of large-width pit boundaries caused by insufficient initial boundary search range can be effectively avoided, thereby improving detection accuracy.

[0047] Example 3: This embodiment provides a pothole severity assessment and adaptive screening method. This method can be executed based on the pothole boundary (preliminary boundary or corrected boundary) obtained in Embodiment 1 or Embodiment 2. It is mainly used to suppress noise interference and false alarms of minor defects, and to distinguish between significant potholes that require maintenance and minor pavement defects that do not require treatment.

[0048] For each detected pit, severity assessment and screening are performed according to the following steps: Step 310: Construct a comprehensive pothole severity index.

[0049] First, based on the initial boundary of the pit. and termination boundary The pit depth is determined by combining the corresponding characteristic responses. : ; In the formula, The characteristic response of the corresponding point; Then, the overall severity is calculated based on the pit depth and the corresponding contour difference features and instantaneous energy features. : ; In the formula, The contour difference feature term represents the sum of the contour difference features of all points within the boundary of the pit / groove. For instantaneous energy characteristic term, it represents the sum of instantaneous energy characteristics of all points within the boundary of the pit; α ,β , γ These are weighting coefficients, typically set to 0.7, 0.2, and 0.2 respectively, and can be adjusted according to the actual application.

[0050] Step S320: Adaptive threshold filtering.

[0051] First, calculate the median absolute deviation of the overall severity of all potholes. : ; In the formula, The overall severity of each pothole.

[0052] Then, set the adaptive threshold, as shown in the following expression: ; In the formula, This is the average severity of all potholes. k >0 is an adjustable parameter used to control the strictness of pit retention.

[0053] Finally, the pits are filtered based on an adaptive threshold, resulting in a set of filtered pits. P final It can be represented as: ; That is, retain pits with a comprehensive severity greater than the adaptive threshold, and remove pits with a comprehensive severity less than or equal to the adaptive threshold. Noise or minor defects. Finally, according to P final The starting coordinates of each pit are determined, and the actual position of the pit on the laser centerline is output.

[0054] By employing the adaptive threshold filtering described above, the false alarm rate can be effectively reduced, and the reliability of pothole detection can be improved. This embodiment can be used in conjunction with Embodiment 1 or Embodiment 2 to output filtered pothole information and its spatial location.

[0055] Example 4: This embodiment provides a vehicle-mounted road pothole detection system for executing the road pothole extraction method described in any one of embodiments one to three, thereby achieving real-time detection and three-dimensional reconstruction of road potholes in dynamic vehicle scenarios.

[0056] like Figure 2 As shown, the system includes an image acquisition module 410, a vehicle positioning module 420, an image processing module 430, and a data parsing module 440.

[0057] The image acquisition module 410 is mounted on the vehicle body and is used to emit line structured light onto the road surface and acquire infrared structured light images of the road surface. The image acquisition module includes an infrared structured light emitter, an on-board encoder, and an infrared camera. The infrared structured light emitter (i.e., a line laser emitter) has a wavelength of 800nm ​​and projects a line of structured light onto the road surface at an angle of 30° to 50° with respect to the ground. The infrared camera, mounted vertically downwards, is equipped with an 800nm ​​band filter and is used to acquire road surface images containing structured light. The wheel encoder, mounted on the wheels, outputs pulse signals to trigger camera acquisition, ensuring that images are acquired at fixed spatial intervals during vehicle movement.

[0058] The vehicle positioning module 420, such as a GPS positioning module, is used to acquire the vehicle's location information (latitude, longitude, mileage, etc.) in real time and to spatiotemporally correlate the location information with each frame of the collected road surface infrared structured light image to provide a basis for the spatial positioning of potholes.

[0059] The image processing module 430 is connected to the image acquisition module and the positioning module, and is used to execute the road surface pothole extraction method described in any one of Embodiments 1 to 3. Specifically, this module extracts the laser centerline of each frame of infrared structured light image to obtain one-dimensional contour data, and then extracts the preliminary boundary of the pothole according to steps S120 to S170 of Embodiment 1. It can further perform boundary correction according to Embodiment 2 and severity assessment and adaptive filtering according to Embodiment 3. During the processing, the image processing module simultaneously receives the position information from the positioning module, associates the pothole boundary extracted in each frame with the position information of that frame, and generates two-dimensional contour data of the road surface and pothole with geographic coordinates.

[0060] The data parsing module 440 is used to parse the two-dimensional contour data output by the image processing module and construct a three-dimensional point cloud model of the road surface, such as... Figure 6 As shown in the diagram. Simultaneously, this module can also measure the three-dimensional dimensions (length, width, depth), area, and volume of the pit in real time.

[0061] Optionally, the system can also be equipped with a data storage module to store detected pit information, including the spatial location, geometric parameters, severity level, corresponding frame image data, and 3D point cloud model of the pit. Figure 7 As shown in the image. In addition, the system includes an alarm module that, when the overall severity of detected potholes exceeds a preset alarm threshold, issues different levels of audible and visual alarm signals and uploads the pothole information to a cloud server or notifies maintenance personnel.

[0062] In actual operation, when the vehicle is in motion, the wheel encoder triggers the infrared camera to collect road structure light images at fixed spatial intervals (e.g., one frame per centimeter); the image processing module processes each frame of image in real time and extracts the pothole boundaries; the vehicle positioning module synchronously records the position information of the current frame; the data parsing module stitches the contour data of each frame in sequence and displays the three-dimensional point cloud of the road surface in real time; when a significant pothole is detected, an early warning is issued, and the data storage module saves the relevant data to facilitate subsequent maintenance decisions.

[0063] The vehicle-mounted road pothole detection system described in this embodiment integrates the depression depth, contour change intensity, and energy characteristics within the pothole area to construct a pothole severity evaluation index. Based on an adaptive threshold strategy using median absolute deviation, the system filters pothole detection results to suppress noise interference and minor defects. It can achieve real-time detection and 3D reconstruction of road potholes in all-weather, vehicle-mounted dynamic scenarios, and has the advantages of high detection accuracy, strong real-time performance, and low susceptibility to lighting conditions.

[0064] In summary, this invention provides a method for extracting road potholes and a vehicle-mounted road pothole detection system, which can extract the three-dimensional contour features of road potholes in real time, 24 / 7, solving the problems of low efficiency and poor real-time performance of traditional detection methods. It is suitable for safety maintenance scenarios of airport runways, highways and urban roads.

[0065] Furthermore, the present invention also provides a computer device / apparatus / system, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the above-described method.

[0066] The present invention also provides a computer-readable storage medium having a computer program / instructions stored thereon, which, when executed by a processor, implement the steps of the above-described method.

[0067] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, systems, and computer program products according to embodiments of the invention. It should be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more blocks of the flowchart illustrations and / or block diagrams.

[0068] These computer program instructions may also be stored in a computer-readable storage medium capable of directing a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more blocks in a block diagram.

[0069] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable data processing apparatus to produce a computer-implemented process, thereby providing steps for implementing the functions specified in one or more flowcharts and / or one or more blocks in a block diagram.

[0070] It should be noted that, depending on the implementation needs, the various steps / components described in this invention can be broken down into more steps / components, or two or more steps / components or parts of the operation of steps / components can be combined into new steps / components to achieve the purpose of this invention.

[0071] Those skilled in the art will readily understand that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for extracting road potholes, characterized in that, The method includes: Laser centerline extraction is performed on the infrared structured light image of the road surface to obtain one-dimensional contour data; Gaussian smoothing is applied to the one-dimensional contour data to obtain the road surface baseline; Subtract the one-dimensional profile data from the road surface baseline to obtain the characteristic response of each point; Based on the characteristic response, the instantaneous energy characteristics of each point are calculated using the Teager-Kaiser energy operator; Based on one-dimensional contour data, the contour difference features of each point are extracted using a first-order local difference operator. By combining the instantaneous energy characteristics and contour difference characteristics of each point, an enhanced signal is generated; The peak point of the enhanced signal is used as the candidate center point of the pit. Starting from each candidate center point, the boundary is expanded point by point to the left and right sides according to the preset boundary expansion conditions until the preset boundary expansion conditions are no longer met, so as to determine the preliminary boundary of the pit. The preset boundary expansion conditions include: the feature response is greater than zero, or the absolute value of the contour difference feature is less than the preset local difference threshold, or the absolute value of the instantaneous energy feature is greater than the preset energy threshold. Among them, the Teager-Kaiser energy operator is used to analyze the characteristic response. S x Calculate instantaneous energy characteristics : ； In the formula, The instantaneous energy characteristics of the j-th point are... The characteristic response of the j-th point; The first-order local difference operator calculates the profile difference features at each point according to the following formula: ； In the formula, For the contour data value of the j-th point, For preset length, The contour difference feature of the j-th point; The enhanced signal is calculated using the following formula: ； In the formula, For the contour difference feature of the j-th point, The instantaneous energy characteristics of the j-th point are... Let be the enhanced signal value at point j.

2. The method for extracting road potholes according to claim 1, characterized in that, The method also includes: Median filtering is applied to one-dimensional contour data to obtain smoothed data; Based on the obtained smoothed data, the contour difference features of each point are extracted using a first-order local difference operator.

3. The method for extracting road potholes according to claim 1, characterized in that, The method also includes: When the absolute value of the instantaneous energy characteristic is less than the preset energy threshold, the instantaneous energy characteristic of that point is set to zero.

4. The method for extracting road potholes according to claim 1, characterized in that, The method also includes: When the absolute value of the contour difference feature is less than the preset local difference threshold, the contour difference feature of that point is set to zero.

5. The method for extracting road potholes according to claim 1, characterized in that, The method also includes: When the width of the initial boundary of the pit is greater than the preset width threshold, the maximum value of the feature response within the initial boundary range of the pit is used as the reference point. Within the preset search range to the left and right of the reference point, the left minimum point and right minimum point of the feature response are found again, and the left minimum point and right minimum point are updated as the starting boundary and ending boundary of the pit.

6. The method for extracting road potholes according to claim 1 or 5, characterized in that, The method also includes: The pit depth is determined based on the initial and final boundaries of the pit, combined with the corresponding feature responses: ； In the formula, As the starting boundary, To terminate the boundary, The characteristic response of the corresponding point, The depth of the pit; For each pit, the overall severity is calculated based on the pit depth and the corresponding contour difference features and instantaneous energy features: ； In the formula, The contour difference feature term represents the sum of the contour difference features of all points within the boundary of the pit / groove. For instantaneous energy characteristic term, it represents the sum of instantaneous energy characteristics of all points within the boundary of the pit; α , β , γ These are the weighting coefficients; For overall severity; The median absolute deviation is calculated based on the overall severity of all potholes. ; An adaptive threshold is set based on the absolute deviation of the median. : ； in, This represents the average severity of all potholes. k >0 indicates an adjustable parameter; Based on adaptive threshold Filter out pits whose overall severity exceeds the adaptive threshold.

7. The method for extracting road potholes according to any one of claims 1 to 5, characterized in that, The method also includes: The above steps are repeated for multiple consecutive frames of infrared structured light images of the road surface. The pothole boundaries extracted from each frame are combined in frame order to generate a three-dimensional point cloud model of the road surface, and the spatial location and geometric parameters of the potholes are output. The geometric parameters include length, width, depth, area and volume.

8. A vehicle-mounted road pothole detection system, characterized in that, The system includes: The image acquisition module is used to emit structured light onto the road surface and acquire infrared structured light images of the road surface; An image processing module is used to execute the road surface pothole extraction method according to any one of claims 1 to 7, so as to extract the pothole boundaries of each frame of road surface infrared structured light image; The positioning module is used to acquire the vehicle's location information and associate the location information with the pothole boundaries of each frame of the road surface infrared structured light image; The data storage module is used to combine the pothole boundaries extracted from each frame with the location information in frame order to generate a 3D point cloud model of the road surface.

Images