Road inspection method and related device based on vehicle-mounted laser radar

Abstract

The present invention discloses a road inspection method and related devices based on a vehicle-mounted laser radar, wherein the method includes: emitting a laser beam by controlling the emission frequency, power, angle and direction of a laser transmitter; receiving a signal reflected from the emitted laser beam to obtain three-dimensional data of the road surface; performing a filtering preprocessing operation on the three-dimensional data, extracting the features of obstacles in the preprocessed three-dimensional data according to a feature extraction model; judging whether an obstacle exists according to an adaptive filtering and non-local filtering algorithm, and outputting the location information of the obstacle. The present invention judges whether an obstacle exists according to the adaptive filtering and non-local filtering algorithms, does not need to rely on manual inspection, improves the efficiency of road inspection, can timely discover hidden dangers on the road, and is of great significance for ensuring road safety.

Description

Technical Field

[0001] The present invention relates to the technical field of road inspection, and in particular to a road inspection method and system based on a vehicle-mounted laser radar, electronic equipment and computing equipment. Background Art

[0002] Intelligent driving technology is increasingly being used in the automotive industry, and LiDAR, as the core sensor of intelligent driving technology, is also increasingly being used. Among them, road obstacle detection in autonomous driving is crucial, because accurate obstacle identification is a prerequisite for safe vehicle driving.

[0003] Existing obstacle detection methods mainly include: Ultrasonic detection: ultrasonic sensors are used to emit sound waves to the surroundings and receive the reflected sound waves. By analyzing the propagation time of the sound waves and the amplitude of the echo, the distance and shape of the obstacle can be determined. This method has high accuracy and reliability at low speeds, but is not suitable for high-speed movement and complex environments. Computer vision detection: computer vision identifies obstacles by analyzing images taken by the camera. This method is low-cost and can provide rich environmental information. However, it is greatly affected by lighting and line of sight conditions and is sensitive to interference such as shadows and reflections. Laser or millimeter wave radar detection: laser radar obtains detailed three-dimensional information about the surrounding environment by emitting laser beams and measuring the time it takes to reflect back. Millimeter wave radar uses electromagnetic waves in the millimeter wave band to detect obstacles. This method can provide high-precision and high-resolution obstacle location and shape information. However, there are also some disadvantages in the vehicle-mounted laser radar obstacle inspection technology. The data processing is difficult: the amount of three-dimensional point cloud data obtained by the laser radar is very large, and it needs to be processed and analyzed efficiently and accurately, so the data processing technology is required to be high.

[0004] Therefore, the problem of how to improve the road inspection efficiency of vehicle-mounted lidar needs to be solved urgently. Summary of the invention

[0005] In view of the above problems, the present invention is proposed to provide a road inspection method and system based on vehicle-mounted laser radar, electronic equipment and computing equipment that overcome the problem of low detection efficiency of the above-mentioned road inspection method based on vehicle-mounted laser radar.

[0006] According to one aspect of the present invention, a road inspection method based on a vehicle-mounted laser radar is provided, comprising:

[0007] The laser beam is emitted by controlling the emission frequency, power, angle and direction of the laser emitter;

[0008] receiving a signal reflected from the emitted laser beam to obtain three-dimensional data of the road surface;

[0009] Performing a filtering preprocessing operation on the three-dimensional data, and extracting features of obstacles in the preprocessed three-dimensional data according to a feature extraction model;

[0010] The adaptive filtering and non-local filtering algorithms are used to determine whether there are obstacles and output the location information of the obstacles.

[0011] In an optional manner, extracting the features of obstacles from the preprocessed three-dimensional data according to the feature extraction model further includes:

[0012] The feature extraction model adopts a self-attention mechanism to generate a weighted coefficient for each position by calculating the correlation between different positions in the three-dimensional data;

[0013] The weighting coefficient of each position is applied to the three-dimensional data to obtain a weighted feature representation, and multi-level feature extraction of the three-dimensional data is achieved by stacking multiple layers of self-attention mechanisms.

[0014] In an optional manner, the determining whether there is an obstacle according to the adaptive filtering and non-local filtering algorithms further includes:

[0015] Determine whether the difference and similarity between the three-dimensional data after adaptive filtering and non-local filtering and the original three-dimensional data exceeds or falls below a preset threshold;

[0016] If the difference or similarity exceeds or falls below a threshold, it is determined that an obstacle exists.

[0017] In an optional manner, the calculation formula of the adaptive filtering is:

[0018]

[0019] Among them, W(n+1) is the weight vector value at time n+1, W(n) is the weight vector value at time n, e(n) is the error vector, μ is a preset constant that controls the speed and amplitude of weight update, d(n) is the value of the target vector at time n, X(n) is the value of the input vector at time n, and X(n) is T is the transpose of X(n).

[0020] In an optional manner, the calculation formula of the non-local filtering algorithm is:

[0021]

[0022] Where d(i,j) is the pixel value of the output image at the coordinate position (i,j), f(k,l) is the pixel value of the input image at the coordinate position (k,l), m is the mean of the pixel values, g(k,l) is the image block similar to f(k,l), w(k,l,i,j) is the similarity weight, and Z is the normalization constant.

[0023] In an optional manner, the filtering preprocessing operation on the three-dimensional data further includes:

[0024] Applying Fourier transform to convert the three-dimensional data from the time domain to the frequency domain, and performing low-pass filtering on the three-dimensional data in the frequency domain to remove high-frequency noise;

[0025] An inverse Fourier transform is applied to convert the filtered three-dimensional data from the frequency domain back to the time domain, and the filtered three-dimensional data is output.

[0026] In an optional manner, the expression of the Fourier transform is:

[0027]

[0028] The expression of the inverse Fourier transform is:

[0029]

[0030] Among them, X(k) is a complex sequence in the frequency domain, x(n) is a complex sequence in the time domain, w(k×n) is a complex sequence of the Nth root of unity, and N is the length of the vector.

[0031] According to another aspect of the present invention, a road inspection system based on a vehicle-mounted laser radar is provided, comprising:

[0032] A laser radar transmitting module, used for transmitting a laser beam, comprising a laser transmitter, a control device and an adjusting device, wherein the laser transmitter is used for transmitting a laser beam, the control device is used for controlling the transmitting frequency and power of the laser transmitter, and the adjusting device is used for adjusting the angle and direction of the laser transmitter;

[0033] A laser radar receiving module, used to receive the signal reflected by the emitted laser beam to obtain three-dimensional data of the road surface;

[0034] The road recognition module includes a preprocessing unit and a recognition unit. The preprocessing unit is used to perform a filtering preprocessing operation on the three-dimensional data. The recognition unit is used to extract the features of obstacles in the preprocessed three-dimensional data according to a real-time feature extraction model, determine whether there are obstacles according to adaptive filtering and non-local filtering algorithms, and output the location information of the obstacles.

[0035] According to another aspect of the present invention, there is provided an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned road inspection method based on vehicle-mounted laser radar when executing the program.

[0036] According to another aspect of the present invention, there is provided a computing device, comprising: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus;

[0037] The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the above-mentioned road inspection method based on vehicle-mounted laser radar.

[0038] According to the solution provided by the present invention, a laser beam is emitted by controlling the emission frequency, power, angle and direction of a laser emitter; a signal reflected from the emitted laser beam is received to obtain three-dimensional data of the road surface; a filtering preprocessing operation is performed on the three-dimensional data, and features of obstacles in the preprocessed three-dimensional data are extracted according to a feature extraction model; and whether an obstacle exists is determined according to an adaptive filtering and non-local filtering algorithm, and the location information of the obstacle is output. The present invention determines whether an obstacle exists according to an adaptive filtering and non-local filtering algorithm, without relying on manual inspection, thereby improving the efficiency of road inspections, and being able to promptly discover hidden dangers on the road, which is of great significance for ensuring road safety.

[0039] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0041] Figure 1 A schematic diagram showing a process flow of a road inspection method based on a vehicle-mounted laser radar according to an embodiment of the present invention is shown;

[0042] Figure 2 A functional structure diagram of a road inspection system based on a vehicle-mounted laser radar according to an embodiment of the present invention is shown;

[0043] Figure 3A schematic diagram of the structure of a computing device according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0044] The exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present invention and to enable the scope of the present invention to be fully communicated to those skilled in the art.

[0045] Figure 1 FIG. 1 is a flow chart showing a road inspection method based on a vehicle-mounted laser radar according to an embodiment of the present invention. Figure 1 As shown, the following steps are included:

[0046] Step S101, emitting a laser beam by controlling the emission frequency, power, angle and direction of the laser emitter.

[0047] By controlling the emission frequency, power, angle and direction of the laser transmitter, precise control of the laser beam can be achieved. For example, by controlling the frequency and power of the laser transmitter (LiDAR), high-precision and high-resolution distance measurement and terrain mapping can be achieved, and by adjusting the angle and direction, scanning and imaging of a specific area can be achieved. By controlling the frequency and direction of the laser transmitter, high-speed and long-distance communication can be achieved.

[0048] Step S102, receiving a signal reflected from the emitted laser beam to obtain three-dimensional data of the road surface.

[0049] By receiving the signal reflected from the emitted laser beam, three-dimensional data of the road surface can be obtained. For example, by emitting a laser beam to the road surface and measuring the time it takes to be reflected back, the distance between the laser and the road surface can be calculated. By continuously moving the lidar or rotating the transmitter, three-dimensional data of the road surface can be obtained.

[0050] Step S103: performing a filtering preprocessing operation on the three-dimensional data, and extracting features of obstacles from the preprocessed three-dimensional data according to a feature extraction model.

[0051] Performing filtering preprocessing operations on the three-dimensional data can improve the accuracy and reliability of the data, thereby better extracting the features of obstacles. Common filtering preprocessing methods include median filtering, Gaussian filtering, bilateral filtering, etc., which can eliminate noise and outliers, smooth data, and thus improve the accuracy of subsequent processing. When extracting the features of obstacles, the preprocessed three-dimensional data can be processed according to the feature extraction model. The feature extraction model can identify the edges, shapes, sizes and other features of obstacles, thereby providing a basis for subsequent obstacle identification.

[0052] In an optional manner, the filtering preprocessing operation on the three-dimensional data further includes:

[0053] Applying Fourier transform to convert the three-dimensional data from the time domain to the frequency domain, and performing low-pass filtering on the three-dimensional data in the frequency domain to remove high-frequency noise;

[0054] An inverse Fourier transform is applied to convert the filtered three-dimensional data from the frequency domain back to the time domain, and the filtered three-dimensional data is output.

[0055] First, the 3D data is converted from the time domain to the frequency domain using Fourier transform, and then the data is low-pass filtered in the frequency domain to remove high-frequency noise. Finally, the filtered data is converted from the frequency domain back to the time domain using inverse Fourier transform, and the processed 3D data is output.

[0056] For example, in the Fourier transform process, the three-dimensional data xt is converted from the time domain to the frequency domain to obtain the frequency domain representation Xf.

[0057] Xf=∫xte -2πt dt

[0058] During the low-pass filtering process, a low-pass filter is applied to Xf in the frequency domain to filter out high-frequency noise.

[0059] Yf=H Lf ×

[0060] Among them, H Lf is the frequency response of the low-pass filter.

[0061] In the inverse Fourier transform process, the low-pass filtered frequency domain representation Yf is converted back to the time domain to obtain the filtered three-dimensional data yt

[0062] yt=∫Yfe -2πf df

[0063] In an optional manner, the expression of the Fourier transform is:

[0064]

[0065] The expression of the inverse Fourier transform is:

[0066]

[0067] Among them, X(k) is a complex sequence in the frequency domain, which represents the representation of the signal in the frequency domain, x(n) is a complex sequence in the time domain, which represents the representation of the signal in the time domain, w(k×n) is a complex sequence of the Nth root of unity, and N is the length of the vector, that is, the sequence length or the number of data points.

[0068] In this embodiment, the Fourier transform process and the inverse Fourier transform process are optimized, and the time domain signal x(n) is defined by a filter w(k×n) to obtain a frequency domain signal X(k). By optimizing the Fourier transform and inverse Fourier transform processes, the speed and accuracy of signal processing can be improved. Since the overhead of complex operations is large in the calculation of Fourier transform and inverse transform, in this embodiment, real number operations are used instead of complex operations to reduce the overhead of floating-point operations.

[0069] In an optional manner, extracting the features of obstacles from the preprocessed three-dimensional data according to the feature extraction model further includes:

[0070] The feature extraction model adopts a self-attention mechanism to generate a weighted coefficient for each position by calculating the correlation between different positions in the three-dimensional data;

[0071] The weighting coefficient of each position is applied to the three-dimensional data to obtain a weighted feature representation, and multi-level feature extraction of the three-dimensional data is achieved by stacking multiple layers of self-attention mechanisms.

[0072] Specifically, the self-attention mechanism is used to make the feature extraction model focus on different parts of the input data, and a weight is assigned to each part according to the correlation between these parts, so that the model can understand and take into account all the information in the input data, not just the directly related parts. By calculating the correlation of different positions in the three-dimensional data to analyze the features of different positions in the three-dimensional data, it can help the model understand the inherent structure and relationship of the data. By applying these weighting coefficients to the original three-dimensional data, a weighted representation of each position can be obtained. For important positions, the weighting coefficient will be larger, while for less important positions, the weighting coefficient will be reduced. By stacking multiple layers of self-attention mechanisms, the features of three-dimensional data can be analyzed at multiple levels, which enhances the representation ability of the model and enables it to understand the complex structure and pattern of the data more deeply. Multi-layer self-attention mechanisms can extract multi-level features from three-dimensional data, not only extracting basic surface features, but also further extracting deeper surface features.

[0073] Step S104: determine whether there is an obstacle based on the adaptive filtering and non-local filtering algorithms, and output the location information of the obstacle.

[0074] In this embodiment, adaptive filtering and non-local filtering algorithms can be used for obstacle detection and positioning, wherein the adaptive filtering algorithm can adjust its own parameters according to the input signal and the expected output signal to minimize the error between the two. In obstacle detection, the adaptive filter is used to process the data of the laser radar. By comparing the actual measurement data with the expected data without obstacles, the existence of obstacles can be detected, and the location information of the obstacles can be extracted by analyzing the output of the filter. The non-local filtering algorithm achieves image smoothing and detail retention by comparing each pixel in the image with the surrounding pixels and adjusting the weights according to the similarity. In obstacle detection, non-local filtering can detect the edge of the obstacle by comparing the difference between each pixel and the surrounding pixels.

[0075] In an optional manner, the determining whether there is an obstacle according to the adaptive filtering and non-local filtering algorithms further includes:

[0076] Determine whether the difference and similarity between the three-dimensional data after adaptive filtering and non-local filtering and the original three-dimensional data exceeds or falls below a preset threshold;

[0077] If the difference or similarity exceeds or falls below a threshold, it is determined that an obstacle exists.

[0078] When performing obstacle detection, it is possible to determine whether there is an obstacle by comparing the difference and similarity between the three-dimensional data after adaptive filtering and non-local filtering and the original three-dimensional data. Specifically, first, it is necessary to calculate the difference between the three-dimensional data after adaptive filtering and non-local filtering and the original three-dimensional data. In order to determine whether there is an obstacle, it is necessary to set a threshold of the difference, which can be determined according to actual needs. If the calculated difference exceeds the preset threshold, then there may be an obstacle. According to the comparison result between the calculated difference and the threshold, it can be determined whether there is an obstacle. If the difference exceeds the threshold, it can be considered that there is an obstacle; if the similarity is lower than the threshold, it can also be considered that there is an obstacle. Finally, the judgment result is output. If there is an obstacle, the location information and other relevant information of the obstacle are output; if there is no obstacle, the corresponding result is output. It should be noted that this obstacle detection method based on difference and similarity may be affected by sensor accuracy, environmental noise and other factors, resulting in misjudgment or missed judgment. Therefore, in practical applications, multiple measurements and calibrations are required to improve the accuracy and reliability of detection.

[0079] In an optional manner, the calculation formula of the adaptive filtering is:

[0080]

[0081] Among them, W(n+1) is the weight vector value at time n+1, W(n) is the weight vector value at time n, e(n) is the error vector, μ is a preset constant that controls the speed and amplitude of weight update, d(n) is the value of the target vector at time n, X(n) is the value of the input vector at time n, and X(n) is T is the transpose of X(n).

[0082] In this embodiment, the weight vector W(n+1) at the next moment is updated according to the current weight vector W(n), the target vector d(n), the input vector X(n) and the error vector e(n). Wherein μ is a factor that controls the update speed and amplitude.

[0083] In an optional manner, the calculation formula of the non-local filtering algorithm is:

[0084]

[0085] Where d(i,j) is the pixel value of the output image at the coordinate position (i,j), f(k,l) is the pixel value of the input image at the coordinate position (k,l), m is the mean of the pixel values, g(k,l) is the image block similar to f(k,l), w(k,l,i,j) is the similarity weight, and Z is the normalization constant.

[0086] In this embodiment, the similarity between pixels in the three-dimensional data is used for filtering, not only the surrounding pixels are considered, but also other similar pixel blocks are considered, and different weights are assigned according to the similarity. Among them, w(k,l,i,j) is the similarity weight, and the weight of the pixel block similar to the current pixel will be greater; otherwise, the weight will be smaller. The square of the modulus used to calculate the image block similar to fk,l represents the energy or amplitude of the image block.

[0087] According to the solution provided by the present invention, a laser beam is emitted by controlling the emission frequency, power, angle and direction of a laser emitter; a signal reflected from the emitted laser beam is received to obtain three-dimensional data of the road surface; a filtering preprocessing operation is performed on the three-dimensional data, and features of obstacles in the preprocessed three-dimensional data are extracted according to a feature extraction model; and whether an obstacle exists is determined according to an adaptive filtering and non-local filtering algorithm, and the location information of the obstacle is output. The present invention determines whether an obstacle exists according to an adaptive filtering and non-local filtering algorithm, without relying on manual inspection, thereby improving the efficiency of road inspections, and being able to promptly discover hidden dangers on the road, which is of great significance for ensuring road safety.

[0088] Figure 2 The functional structure diagram of the road inspection system based on the vehicle-mounted laser radar according to an embodiment of the present invention is shown, including:

[0089] A laser radar transmitting module, used for transmitting a laser beam, comprising a laser transmitter, a control device and an adjusting device, wherein the laser transmitter is used for transmitting a laser beam, the control device is used for controlling the transmitting frequency and power of the laser transmitter, and the adjusting device is used for adjusting the angle and direction of the laser transmitter;

[0090] A laser radar receiving module, used to receive the signal reflected by the emitted laser beam to obtain three-dimensional data of the road surface;

[0091] The road recognition module includes a preprocessing unit and a recognition unit. The preprocessing unit is used to perform a filtering preprocessing operation on the three-dimensional data. The recognition unit is used to extract the features of obstacles in the preprocessed three-dimensional data according to a real-time feature extraction model, determine whether there are obstacles according to adaptive filtering and non-local filtering algorithms, and output the location information of the obstacles.

[0092] Figure 3 The schematic diagram of the structure of the computing device embodiment of the present invention is shown. The specific embodiment of the present invention does not limit the specific implementation of the computing device.

[0093] like Figure 3 As shown, the computing device may include: a processor 302 , a communications interface 304 , a memory 306 , and a communications bus 308 .

[0094] The processor 302, the communication interface 304, and the memory 306 communicate with each other through the communication bus 308. The communication interface 304 is used to communicate with other devices such as a client or other server network elements. The processor 302 is used to execute the program 310, which can specifically execute the operations corresponding to the above-mentioned road inspection method based on the vehicle-mounted laser radar.

[0095] Specifically, the program 310 may include program codes, which include computer operation instructions.

[0096] The processor 302 may be a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention. The one or more processors included in the computing device may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0097] The memory 306 is used to store the program 310. The memory 306 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0098] The program 310 may be specifically configured to enable the processor 302 to perform the following operations:

[0099] The laser beam is emitted by controlling the emission frequency, power, angle and direction of the laser emitter;

[0100] receiving a signal reflected from the emitted laser beam to obtain three-dimensional data of the road surface;

[0101] Performing a filtering preprocessing operation on the three-dimensional data, and extracting features of obstacles in the preprocessed three-dimensional data according to a feature extraction model;

[0102] The adaptive filtering and non-local filtering algorithms are used to determine whether there are obstacles and output the location information of the obstacles.

[0103] In an optional manner, extracting the features of obstacles from the preprocessed three-dimensional data according to the feature extraction model further includes:

[0104] The feature extraction model adopts a self-attention mechanism to generate a weighted coefficient for each position by calculating the correlation between different positions in the three-dimensional data;

[0105] The weighting coefficient of each position is applied to the three-dimensional data to obtain a weighted feature representation, and multi-level feature extraction of the three-dimensional data is achieved by stacking multiple layers of self-attention mechanisms.

[0106] In an optional manner, the determining whether there is an obstacle according to the adaptive filtering and non-local filtering algorithms further includes:

[0107] Determine whether the difference and similarity between the three-dimensional data after adaptive filtering and non-local filtering and the original three-dimensional data exceeds or falls below a preset threshold;

[0108] If the difference or similarity exceeds or falls below a threshold, it is determined that an obstacle exists.

[0109] In an optional manner, the calculation formula of the adaptive filtering is:

[0110]

[0111] Among them, W(n+1) is the weight vector value at time n+1, W(n) is the weight vector value at time n, e(n) is the error vector, μ is a preset constant that controls the speed and amplitude of weight update, d(n) is the value of the target vector at time n, X(n) is the value of the input vector at time n, and X(n) is T is the transpose of X(n).

[0112] In an optional manner, the calculation formula of the non-local filtering algorithm is:

[0113]

[0114] Where d(i,j) is the pixel value of the output image at the coordinate position (i,j), f(k,l) is the pixel value of the input image at the coordinate position (k,l), m is the mean of the pixel values, g(k,l) is the image block similar to f(k,l), w(k,l,i,j) is the similarity weight, and Z is the normalization constant.

[0115] In an optional manner, the filtering preprocessing operation on the three-dimensional data further includes:

[0116] Applying Fourier transform to convert the three-dimensional data from the time domain to the frequency domain, and performing low-pass filtering on the three-dimensional data in the frequency domain to remove high-frequency noise;

[0117] An inverse Fourier transform is applied to convert the filtered three-dimensional data from the frequency domain back to the time domain, and the filtered three-dimensional data is output.

[0118] In an optional manner, the expression of the Fourier transform is:

[0119]

[0120] The expression of the inverse Fourier transform is:

[0121]

[0122] Among them, X(k) is a complex sequence in the frequency domain, x(n) is a complex sequence in the time domain, w(k×n) is a complex sequence of the Nth root of unity, and N is the length of the vector.

[0123] According to the solution provided by the present invention, a laser beam is emitted by controlling the emission frequency, power, angle and direction of a laser emitter; a signal reflected from the emitted laser beam is received to obtain three-dimensional data of the road surface; a filtering preprocessing operation is performed on the three-dimensional data, and features of obstacles in the preprocessed three-dimensional data are extracted according to a feature extraction model; and whether an obstacle exists is determined according to an adaptive filtering and non-local filtering algorithm, and the location information of the obstacle is output. The present invention determines whether an obstacle exists according to an adaptive filtering and non-local filtering algorithm, without relying on manual inspection, thereby improving the efficiency of road inspections, and being able to promptly discover hidden dangers on the road, which is of great significance for ensuring road safety.

[0124] The algorithm or display provided herein is not inherently related to any particular computer, virtual system or other equipment. Various general purpose systems can also be used together with the teachings based on this. According to the above description, it is obvious to construct the structure required for this type of system. In addition, the embodiment of the present invention is not directed to any specific programming language yet. It should be understood that various programming languages ​​can be utilized to realize the content of the present invention described herein, and the description made to specific languages ​​above is for disclosing the best mode of the present invention.

[0125] In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.

[0126] Similarly, it should be understood that in order to streamline the present invention and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the present invention, the various features of the embodiments of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, this disclosed method should not be interpreted as reflecting the following intention: that the claimed invention requires more features than the features explicitly recited in each claim. More specifically, as reflected in the claims below, inventive aspects lie in less than all the features of the individual embodiments disclosed above. Therefore, the claims that follow the specific embodiment are hereby expressly incorporated into the specific embodiment, with each claim itself serving as a separate embodiment of the present invention.

[0127] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and in addition they may be divided into a plurality of submodules or subunits or subcomponents. Except that at least some of such features and / or processes or units are mutually exclusive, all features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed in this manner may be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature providing the same, equivalent or similar purpose.

[0128] In addition, those skilled in the art will appreciate that, although some embodiments herein include certain features included in other embodiments but not other features, the combination of features of different embodiments is meant to be within the scope of the present invention and form different embodiments. For example, in the claims below, any one of the claimed embodiments may be used in any combination.

[0129] The various component embodiments of the present invention can be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or digital signal processor (DSP) can be used in practice to implement some or all functions of some or all components according to embodiments of the present invention. The present invention can also be implemented as a device or apparatus program (e.g., computer program and computer program product) for executing part or all of the methods described herein. Such a program implementing the present invention can be stored on a computer-readable medium, or can have the form of one or more signals. Such a signal can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0130] It should be noted that the above embodiments illustrate the present invention rather than limit it, and that those skilled in the art may devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets shall not be construed as a limitation on the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention may be implemented by means of hardware comprising a number of different elements and by means of a suitably programmed computer. In a unit claim enumerating a number of devices, several of these devices may be embodied by the same hardware item. The use of the words first, second, and third, etc. does not indicate any order. These words may be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be understood as limitations on the order of execution.

Claims

1. A road inspection method based on vehicle-mounted laser radar, characterized in that: include: The laser beam is emitted by controlling the emission frequency, power, angle and direction of the laser emitter; receiving a signal reflected from the emitted laser beam to obtain three-dimensional data of the road surface; Performing a filtering preprocessing operation on the three-dimensional data, and extracting features of obstacles in the preprocessed three-dimensional data according to a feature extraction model; Determine whether the difference and similarity between the three-dimensional data after adaptive filtering and non-local filtering and the original three-dimensional data exceeds or falls below a preset threshold; If the difference exceeds a preset threshold or the similarity is lower than a preset threshold, it is determined that an obstacle exists and the location information of the obstacle is output; The calculation formula of the adaptive filtering is: Among them, W(n+1) is the weight vector value at time n+1, W(n) is the weight vector value at time n, e(n) is the error vector, μ is a preset constant that controls the speed and amplitude of weight update, d(n) is the value of the target vector at time n, X(n) is the value of the input vector at time n, and X(n) is T is the transpose of X(n); The calculation formula of the non-local filtering algorithm is: Where d(i,j) is the pixel value of the output image at the coordinate position (i,j), f(k,l) is the pixel value of the input image at the coordinate position (k,l), m is the mean of the pixel values, g(k,l) is the image block similar to f(k,l), w(k,l,i,j) is the similarity weight, and Z is the normalization constant.

2. The road inspection method based on vehicle-mounted laser radar according to claim 1 is characterized in that: Extracting the features of obstacles from the preprocessed three-dimensional data according to the feature extraction model further comprises: The feature extraction model adopts a self-attention mechanism to generate a weighted coefficient for each position by calculating the correlation between different positions in the three-dimensional data; The weighting coefficient of each position is applied to the three-dimensional data to obtain a weighted feature representation, and multi-level feature extraction of the three-dimensional data is achieved by stacking multiple layers of self-attention mechanisms.

3. The road inspection method based on vehicle-mounted laser radar according to claim 1 is characterized in that: The filtering preprocessing operation on the three-dimensional data further includes: Applying Fourier transform to convert the three-dimensional data from the time domain to the frequency domain, and performing low-pass filtering on the three-dimensional data in the frequency domain to remove high-frequency noise; An inverse Fourier transform is applied to convert the filtered three-dimensional data from the frequency domain back to the time domain, and the filtered three-dimensional data is output.

4. The road inspection method based on vehicle-mounted laser radar according to claim 3 is characterized in that: The expression of the Fourier transform is: The expression of the inverse Fourier transform is: Among them, X(k) is a complex sequence in the frequency domain, x(n) is a complex sequence in the time domain, w(k×n) is a complex sequence of the Nth root of unity, and N is the length of the vector.

5. A road inspection system based on vehicle-mounted laser radar, characterized in that: include: A laser radar transmitting module, used for transmitting a laser beam, comprising a laser transmitter, a control device and an adjusting device, wherein the laser transmitter is used for transmitting a laser beam, the control device is used for controlling the transmitting frequency and power of the laser transmitter, and the adjusting device is used for adjusting the angle and direction of the laser transmitter; A laser radar receiving module, used to receive the signal reflected by the emitted laser beam to obtain three-dimensional data of the road surface; A road recognition module includes a preprocessing unit and a recognition unit, wherein the preprocessing unit is used to perform a filtering preprocessing operation on the three-dimensional data, and the recognition unit is used to extract the features of obstacles in the preprocessed three-dimensional data according to a real-time feature extraction model, and judge whether the difference and similarity between the three-dimensional data after adaptive filtering and non-local filtering and the original three-dimensional data exceed or are lower than a preset threshold; if the difference exceeds the preset threshold or the similarity is lower than the preset threshold, it is judged that an obstacle exists, and the location information of the obstacle is output; The calculation formula of the adaptive filtering is: Among them, W(n+1) is the weight vector value at time n+1, W(n) is the weight vector value at time n, e(n) is the error vector, μ is a preset constant that controls the speed and amplitude of weight update, d(n) is the value of the target vector at time n, X(n) is the value of the input vector at time n, and X(n) is T is the transpose of X(n); The calculation formula of the non-local filtering algorithm is: Where d(i,j) is the pixel value of the output image at the coordinate position (i,j), f(k,l) is the pixel value of the input image at the coordinate position (k,l), m is the mean of the pixel values, g(k,l) is the image block similar to f(k,l), w(k,l,i,j) is the similarity weight, and Z is the normalization constant.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, it implements the road inspection method based on vehicle-mounted laser radar as described in any one of claims 1-4.

7. A computing device comprising: A processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus; The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the road inspection method based on vehicle-mounted laser radar as described in any one of claims 1-4.

Images