Novel track geometric parameter detection method

Abstract

The invention belongs to the technical field of track detection. The invention provides a novel track geometric parameter detection method. The method comprises the following steps: determining and controlling a high-definition camera to rotate to a shooting angle based on an irradiation angle of solar rays; performing track identification on the shot high-definition image data to obtain two track sections, performing foreign matter detection on each track section, and predicting the attribute category of the foreign matter based on the detected foreign matter features; dividing the track sections into smooth track sections and foreign matter track sections based on the distribution positions of the foreign matters, extracting first track geometric parameters of the smooth track sections by using an image recognition technology, and calculating second track geometric parameters of the foreign matter track sections based on attribute categories; and aligning and integrating the first track geometric parameters and the second track geometric parameters to obtain track geometric parameters. According to the invention, the dynamic adjustment of the detection view of the camera and the smooth extraction of the track geometric parameters of the foreign matter coverage area are realized.

Description

Technical Field

[0001] The present invention relates to the field of track detection technology, and in particular to a novel method for detecting track geometric parameters. Background Art

[0002] The detection of track geometry parameters is crucial to ensure safe, smooth and comfortable train operation. These parameters include gauge, track direction, height, level, superelevation, triangular pits, etc., which directly reflect the quality of the track. With the increase in train speed, especially the development of high-speed railways, the requirements for high track smoothness are getting higher and higher, because even a small unevenness may cause serious safety problems under high-speed operation conditions.

[0003] Detecting the geometric parameters of the track based on machine vision and image processing technology is currently a more commonly used detection method. However, when the above detection method is implemented, the detection angle of the detection camera is fixed, and its corresponding detection field of view is also fixed, which is not conducive to improving the detection efficiency of the track geometric parameters. Although the efficient detection method based on dynamic field of view is widely demanded in the field, the existing technology has not yet provided relevant solutions. At the same time, there are often foreign objects on the track, and the presence of foreign objects will affect the image extraction of the track geometric parameters. The existing technology also has no effective solution to this. The technical purpose of the present invention is to solve at least one of the above technical problems. Summary of the invention

[0004] In this regard, the present invention provides a new track geometry parameter detection method, system, electronic device, computer storage medium and computer program product to solve the above technical problems.

[0005] The present invention discloses a novel method for detecting track geometric parameters, which is applied to an automatic detection vehicle, wherein the automatic detection vehicle comprises a high-definition camera and a light sensor; the method comprises the following steps:

[0006] The light sensor detects the irradiation angle of sunlight in the current area, determines the shooting angle of the high-definition camera based on the irradiation angle of sunlight, and controls the high-definition camera to rotate to the shooting angle; wherein different shooting angles correspond to shooting fields of different sizes;

[0007] Performing track recognition on the high-definition image data obtained by the high-definition camera at the shooting angle to obtain two track segments, and performing foreign object detection on each of the track segments, and predicting the attribute category of the foreign object based on the foreign object features obtained by the detection;

[0008] Based on the distribution position of the foreign matter, the track segment is divided into a smooth track segment and a foreign matter track segment, the first track geometry parameter of each smooth track segment is extracted using image recognition technology, and the second track geometry parameter of the foreign matter track segment is calculated based on the attribute category;

[0009] The first orbital geometric parameters and the second orbital geometric parameters are aligned and integrated to obtain orbital geometric parameters.

[0010] In some embodiments, the determining of the shooting angle of the high-definition camera based on the illumination angle of sunlight includes:

[0011] Determine the direction data of the track based on the high-precision map data, and calculate the reflection angle range of the reflected light of the sunlight at the illumination angle on the track according to the direction data and the illumination angle of the sunlight;

[0012] Based on the three-dimensional size of the automatic inspection vehicle and the three-dimensional layout of the high-definition camera on the automatic inspection vehicle, the shooting angle that does not overlap with the reflection angle range is determined within the angle adjustment range of the high-definition camera, and the difference between the shooting angle and the current shooting angle of the high-definition camera is the smallest.

[0013] In some embodiments, the performing foreign object detection on each track segment and predicting the attribute category of the foreign object based on the detected foreign object features includes:

[0014] Extracting a first appearance feature of each of the track segments, performing a continuity analysis of the appearance feature based on the first appearance feature, and identifying a region where the appearance feature is discontinuous as the foreign matter;

[0015] Extracting a second appearance feature of the foreign object, the second appearance feature includes a foreign object appearance feature and a foreign object structural feature, predicting the foreign object category of the foreign object based on the foreign object appearance feature and the foreign object structural feature, and predicting the geometric fullness and hardness of the foreign object based on the foreign object category and the foreign object structural feature;

[0016] The foreign matter category, the geometric fullness and the hardness are used as the attribute categories of the foreign matter.

[0017] In some embodiments, the predicting of the geometric fullness and hardness of the foreign body based on the foreign body category and the foreign body structural features includes:

[0018] Predicting a first geometric structure of the foreign object under normal conditions based on the foreign object category and the foreign object structural features, and extracting a real-time second geometric structure of the foreign object from the foreign object structural features;

[0019] The variation range of the geometric structure of the foreign body is calculated based on the first geometric structure and the second geometric structure, and the geometric fullness is determined based on the variation range;

[0020] The dynamic characteristics of the foreign object are extracted from the structural characteristics of the foreign object, the dynamic value of the foreign object is calculated based on the dynamic characteristics, and the hardness of the foreign object is determined based on the dynamic value.

[0021] In some embodiments, the automatic detection vehicle further includes an ultrasonic detector, and the second track geometric parameter of the foreign object track segment is calculated based on the attribute category, including:

[0022] If the hardness is lower than the hardness threshold, the second track geometric parameters of the foreign object track segment are predicted based on the two first track geometric parameters adjacent to the foreign object track segment above and below;

[0023] If the hardness is above the hardness threshold, then:

[0024] If the geometric fullness is lower than the fullness threshold, the second track geometric parameters of the foreign object track segment are predicted based on the two first track geometric parameters adjacent to the foreign object track segment above and below;

[0025] If the geometric fullness is higher than the fullness threshold, the ultrasonic detector is used to detect the foreign object track segment to extract the second track geometric parameters.

[0026] In some embodiments, the predicting of the first geometric structure of the foreign object under normal conditions based on the foreign object category and the foreign object structural features includes:

[0027] The foreign object category and the foreign object structural features are input into a geometric structure prediction model, and the geometric structure prediction model outputs the first geometric structure of the foreign object predicted by it under normal conditions; wherein, the geometric structure prediction model is obtained by fine-tuning the BERT large model.

[0028] The present invention also discloses a novel track geometry parameter detection system, which is applied to an automatic detection vehicle, wherein the automatic detection vehicle comprises a high-definition camera and a light sensor; the system comprises a processing module and a storage module, wherein the processing module implements the following method steps by calling and executing the computer code in the storage module:

[0029] The light sensor detects the irradiation angle of sunlight in the current area, determines the shooting angle of the high-definition camera based on the irradiation angle of sunlight, and controls the high-definition camera to rotate to the shooting angle; wherein different shooting angles correspond to shooting fields of different sizes;

[0030] Performing track recognition on the high-definition image data obtained by the high-definition camera at the shooting angle to obtain two track segments, and performing foreign object detection on each of the track segments, and predicting the attribute category of the foreign object based on the foreign object features obtained by the detection;

[0031] Based on the distribution position of the foreign matter, the track segment is divided into a smooth track segment and a foreign matter track segment, the first track geometry parameter of each smooth track segment is extracted using image recognition technology, and the second track geometry parameter of the foreign matter track segment is calculated based on the attribute category;

[0032] The first orbital geometric parameters and the second orbital geometric parameters are aligned and integrated to obtain orbital geometric parameters.

[0033] The present invention also discloses an electronic device, comprising: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor executes the computer program to implement the method described in any of the preceding items.

[0034] The present invention also discloses a computer storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement any of the above methods.

[0035] The present invention also discloses a computer program product. When the computer program product is run on a terminal, the terminal implements any of the above methods when executing the computer program product.

[0036] The beneficial effects of the present invention are:

[0037] (1) The present invention selects a suitable shooting angle of a high-definition camera by analyzing the angle of sunlight, thereby realizing dynamic adjustment of its shooting field of view, which can effectively improve the extraction efficiency of orbital geometric parameters.

[0038] (2) The present invention analyzes and identifies the attribute categories of foreign objects on the track, and calculates the track geometry parameters of the track section covered by the foreign objects based on the attribute categories, thereby achieving smooth detection of the track geometry parameters when the track is covered by foreign objects, which can greatly improve the adaptability of track geometry parameter detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0040] Figure 1 is a schematic diagram of a scenario of using an automatic detection vehicle to detect track geometric parameters disclosed in an embodiment of the present invention;

[0041] Figure 2 It is a flow chart of a novel method for detecting track geometric parameters disclosed in an embodiment of the present invention;

[0042] Figure 3 It is a structural schematic diagram of a new type of track geometry parameter detection system disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0043] The following is a description of the implementation of the present application by specific specific embodiments. People familiar with the technology can easily understand other advantages and effects of the present application from the contents disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0044] In addition, the technical features involved in the different embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0045] like Figure 1 , Figure 2 As shown, an embodiment of the present invention discloses a novel method for detecting track geometric parameters, which is applied to an automatic detection vehicle, wherein the automatic detection vehicle includes a high-definition camera and a light sensor; the method includes the following steps:

[0046] The light sensor detects the irradiation angle of sunlight in the current area, determines the shooting angle of the high-definition camera based on the irradiation angle of sunlight, and controls the high-definition camera to rotate to the shooting angle; wherein different shooting angles correspond to shooting fields of different sizes;

[0047] Performing track recognition on the high-definition image data obtained by the high-definition camera at the shooting angle to obtain two track segments, and performing foreign object detection on each of the track segments, and predicting the attribute category of the foreign object based on the foreign object features obtained by the detection;

[0048] Based on the distribution position of the foreign matter, the track segment is divided into a smooth track segment and a foreign matter track segment, the first track geometry parameter of each smooth track segment is extracted using image recognition technology, and the second track geometry parameter of the foreign matter track segment is calculated based on the attribute category;

[0049] The first orbital geometric parameters and the second orbital geometric parameters are aligned and integrated to obtain orbital geometric parameters.

[0050] The present invention constructs an automatic inspection vehicle, which is equipped with a high-definition camera and a light sensor. The best shooting angle that the high-definition camera should take can be determined according to the irradiation angle of sunlight detected by the light sensor. When the high-definition camera is adjusted to the best shooting angle, the shooting field of the camera is adjusted. Specifically, the smaller the shooting angle is, the closer the upper edge of the shooting field of the high-definition camera is to the automatic inspection vehicle. At this time, the shooting field of view is smaller (i.e., close-up), and can only cover a short-distance track; the larger the shooting angle is, the farther the upper edge of the shooting field of the high-definition camera is from the automatic inspection vehicle. At this time, the shooting field of view is larger (i.e., long-distance), and can cover a longer distance and a longer track.

[0051] At the same time, after the high-definition camera obtains high-definition image data at the above shooting angle, it identifies the track contained therein and obtains the left and right track segments. In addition, foreign object detection is performed on each track segment, and the attribute category of the foreign object is analyzed based on the detected foreign object characteristics, and the track segment is divided into a smooth track segment and a foreign object track segment based on the distribution position of the foreign object. The foreign object track segment is the area where foreign objects exist on the track.

[0052] Next, the first track geometry parameters of each smooth track segment are extracted using image recognition technology, and the second track geometry parameters of the foreign object track segment are calculated based on the attribute category of the foreign object obtained above. The track geometry parameters include at least one of the track gauge, track surface flatness, track direction, etc.

[0053] Finally, the first orbital geometric parameters and the second orbital geometric parameters obtained above are aligned and integrated to obtain the orbital geometric parameters of the left and right orbits corresponding to the current high-definition image data.

[0054] Therefore, on the one hand, the present invention selects a suitable shooting angle of a high-definition camera by analyzing the angle of sunlight, thereby realizing dynamic adjustment of its shooting field of view, which can effectively improve the extraction efficiency of track geometry parameters. On the other hand, the present invention analyzes and identifies the attribute category of foreign objects on the track, and calculates the track geometry parameters of the track section covered by the foreign objects based on the attribute category, thereby realizing the smooth detection of track geometry parameters when there are foreign objects covering the track, which can greatly improve the adaptability of track geometry parameter detection.

[0055] It should be noted that the automatic inspection vehicle used in the present invention may be driven fully automatically or under manual remote control, and the present invention does not limit this. Figure 1Although the power system, processing system, communication system and other functional components of the automatic detection vehicle are not shown, the configuration of these functional components is a mature prior art in the field and will not be described in detail in the present invention. The adjustment of the shooting angle of the high-definition camera can be achieved based on a multi-axis pan-tilt platform.

[0056] In some embodiments, the determining of the shooting angle of the high-definition camera based on the illumination angle of sunlight includes:

[0057] Determine the direction data of the track based on the high-precision map data, and calculate the reflection angle range of the reflected light of the sunlight at the illumination angle on the track according to the direction data and the illumination angle of the sunlight;

[0058] Based on the three-dimensional size of the automatic inspection vehicle and the three-dimensional layout of the high-definition camera on the automatic inspection vehicle, the shooting angle that does not overlap with the reflection angle range is determined within the angle adjustment range of the high-definition camera, and the difference between the shooting angle and the current shooting angle of the high-definition camera is the smallest.

[0059] In an embodiment of the present invention, the automatic inspection vehicle can determine the orientation data of the currently inspected track based on the high-precision map data to establish a three-dimensional coordinate system based on the track, and combine the illumination angle of the sunlight and the three-dimensional coordinate system to infer the reflection angle range of the sunlight reflected on the track at the illumination angle. Due to the long-term friction contact with the train wheels, the top surface of the track is very smooth and has a strong light reflection ability, and the excessively strong reflected sunlight will appear as overexposure points in the high-definition image data taken by the high-definition camera, making it impossible to extract the geometric parameters of the track.

[0060] In this regard, the present invention is based on the above-mentioned three-dimensional coordinate system, and based on the three-dimensional size of the automatic inspection vehicle and the three-dimensional layout position of the high-definition camera on the automatic inspection vehicle, the coordinates of the lens of the high-definition camera can be determined, and then a number of shooting angles that do not overlap with the above-mentioned reflection angle range can be screened out from the angle adjustment range of the high-definition camera, and then the shooting angle with the smallest difference with the current shooting angle of the high-definition camera is screened out. This can effectively reduce the travel of the high-definition camera to adjust the shooting angle, reduce the waiting time of the automatic inspection vehicle (stopping or slowing down is required when adjusting the shooting angle), and improve the detection efficiency of track geometry parameters.

[0061] In some embodiments, the performing foreign object detection on each track segment and predicting the attribute category of the foreign object based on the detected foreign object features includes:

[0062] Extracting a first appearance feature of each of the track segments, performing a continuity analysis of the appearance feature based on the first appearance feature, and identifying a region where the appearance feature is discontinuous as the foreign matter;

[0063] Extracting a second appearance feature of the foreign object, the second appearance feature includes a foreign object appearance feature and a foreign object structural feature, predicting the foreign object category of the foreign object based on the foreign object appearance feature and the foreign object structural feature, and predicting the geometric fullness and hardness of the foreign object based on the foreign object category and the foreign object structural feature;

[0064] The foreign matter category, the geometric fullness and the hardness are used as the attribute categories of the foreign matter.

[0065] In an embodiment of the present invention, the upper surface of the track is usually smooth, and the geometric features of the track extracted by image detection technology (such as the edge line of the upper surface of the track) should be continuous and smooth, but if there is a foreign object, the corresponding area will become discontinuous and not smooth. In this regard, the present invention performs a continuity analysis of the appearance features of the extracted first appearance features of the track segment, and when the appearance features are found to be discontinuous, the corresponding area is identified as a foreign object. The foreign object recognition method of the present invention has a particularly small computational load, and does not require pre-setting of features for various types of foreign objects, which can greatly improve the efficiency of foreign object recognition, and can support the automatic inspection vehicle to detect and extract track geometric parameters at a higher operating speed, thereby improving the detection efficiency of track geometric parameters.

[0066] After the foreign object is identified, the second appearance feature of the foreign object is extracted. The second appearance feature includes the appearance feature of the foreign object and the structural feature of the foreign object. Among them, the appearance feature of the foreign object mainly includes the color, texture, reflectivity, etc. of the foreign object, and the structural feature of the foreign object mainly refers to the geometric structure size of the foreign object, the length of the part protruding from the upper surface of the track, and the dynamic characteristics of the foreign object, such as whether the foreign object has a large amplitude of swing. Based on the above-mentioned foreign object appearance features and foreign object structural features, the foreign object category can be predicted, such as metal sheets, paper, plastic bottles, milk cartons, etc.; then, based on the attribute category and foreign object structural features, the geometric fullness and hardness of the foreign object can be predicted; finally, the foreign object category, geometric fullness and hardness obtained are used as the attribute category of the foreign object.

[0067] In some embodiments, the predicting of the geometric fullness and hardness of the foreign body based on the foreign body category and the foreign body structural features includes:

[0068] Predicting a first geometric structure of the foreign object under normal conditions based on the foreign object category and the foreign object structural features, and extracting a real-time second geometric structure of the foreign object from the foreign object structural features;

[0069] The variation range of the geometric structure of the foreign body is calculated based on the first geometric structure and the second geometric structure, and the geometric fullness is determined based on the variation range;

[0070] The dynamic characteristics of the foreign object are extracted from the structural characteristics of the foreign object, the dynamic value of the foreign object is calculated based on the dynamic characteristics, and the hardness of the foreign object is determined based on the dynamic value.

[0071] In the embodiment of the present invention, in order to facilitate the subsequent analysis of the geometric parameters of the track in the area where the foreign matter exists, it is necessary to predetermine the geometric fullness and hardness of the foreign matter. The geometric fullness refers to the change range between the current geometric structure of the foreign matter and its original geometric structure; the hardness refers to the hardness of the foreign matter itself, for example, a metal sheet has high hardness and a paper has low hardness.

[0072] Specifically, first, the first geometric structure of the foreign object under normal conditions is predicted based on the attribute category and structural characteristics of the foreign object obtained above. For example, a foreign object is identified as a milk carton, and further considering its geometric structure size (which belongs to the structural characteristics of the foreign object) can predict the first geometric structure of the milk carton when it is normal and intact. The real-time second geometric structure of the foreign object can also be directly extracted from the structural characteristics of the foreign object, and then by comparing the geometric structure change amplitude between the first geometric structure and the second geometric structure, the geometric fullness of the foreign object can be determined based on the change amplitude. Obviously, the larger the change amplitude, the lower its geometric fullness, and vice versa. Among them, the change amplitude can, for example, be equal to the ratio of the volume corresponding to the second geometric structure divided by the volume corresponding to the first geometric structure.

[0073] Next, the dynamic features of the foreign object are extracted from the structural features of the foreign object. Based on the dynamic features, the dynamic value of the foreign object can be calculated. The dynamic features refer to the position change amplitude and change direction of the reference point of the foreign object. The larger the change amplitude and the more change directions, the lower the hardness, such as paper, and vice versa, such as metal sheets. The hardness of the foreign object can then be determined based on the dynamic value.

[0074] In some embodiments, the automatic detection vehicle further includes an ultrasonic detector, and the second track geometric parameter of the foreign object track segment is calculated based on the attribute category, including:

[0075] If the hardness is lower than the hardness threshold, the second track geometric parameters of the foreign object track segment are predicted based on the two first track geometric parameters adjacent to the foreign object track segment above and below;

[0076] If the hardness is above the hardness threshold, then:

[0077] If the geometric fullness is lower than the fullness threshold, the second track geometric parameters of the foreign object track segment are predicted based on the two first track geometric parameters adjacent to the foreign object track segment above and below;

[0078] If the geometric fullness is higher than the fullness threshold, the ultrasonic detector is used to detect the foreign object track segment to extract the second track geometric parameters.

[0079] In the embodiment of the present invention, the automatic inspection vehicle is also equipped with an ultrasonic detector (such as Figure 1 As shown, the ultrasonic detector is arranged in an integrated manner with the high-definition camera), and the ultrasonic waves emitted by the ultrasonic detector can penetrate the foreign objects covered on the track, thereby detecting the geometric parameters of the track, and then judging whether there are pits and cracks on the track (which may be caused by the train wheels rolling over the foreign objects with higher hardness). Therefore, the present invention determines the detection and determination method of the second track geometric parameters of the foreign object track section based on the hardness and geometric fullness of the foreign object.

[0080] Specifically, if the hardness of the foreign object is lower than the hardness threshold, it means that it is a soft material such as paper or milk carton. Even if the train wheels run over it, it will not cause damage to the track. At this time, the second track geometric parameters of the foreign object track segment can be obtained by directly performing line fitting on the two first track geometric parameters adjacent to each other in the upper and lower parts of the foreign object track segment. For line fitting, for example, the edge lines of the upper surfaces of the two smooth track segments adjacent to each other in the upper and lower parts are detected in advance, and the lower end point of the upper edge line is connected with the upper end point of the lower edge line, which constitutes the edge line of the foreign object track segment.

[0081] If the hardness of the foreign object is higher than the hardness threshold, there are defects such as pits and cracks on the track due to the foreign object being crushed by the train wheels, which affect the safety of the track. In this regard, further analysis is conducted to determine whether the geometric fullness of the foreign object is lower than the fullness threshold. If it is lower than the fullness threshold, it indicates that the foreign object has undergone a large degree of deformation after being crushed by the train wheels. Obviously, although its hardness is higher than the hardness threshold (the hardness threshold is usually set low), it is still low for the high-hardness track and is not enough to cause damage to the track. At this point, the aforementioned method is still used to indirectly infer the second track geometric parameters of the foreign object track segment, that is, the two first track geometric parameters adjacent to the upper and lower foreign object track segment are directly linearly fitted to obtain the second track geometric parameters of the foreign object track segment.

[0082] However, if it is higher than the fullness threshold, it means that the foreign object has not been deformed to a large extent after being crushed by the train wheels. Obviously, its own hardness is high enough, and the crushing of the train wheels may cause damage to the track. At this time, it is not appropriate to use the above-mentioned calculation method, but to start the ultrasonic detector to detect the track section with foreign objects, so as to directly extract the second track geometry parameters of the track section covered by foreign objects through ultrasonic guided wave technology.

[0083] Since the scanning operation when detecting the track geometry parameters based on ultrasonic guided wave technology needs to be performed one by one, the detection efficiency is relatively low. Therefore, the present invention makes a decision based on the hardness and geometric fullness of the foreign body and uses image recognition technology or ultrasonic guided wave technology to detect the geometry parameters of the track in the area where the foreign body is located. This can significantly improve the efficiency of track detection while effectively reducing the probability of missed detection of damaged track sections.

[0084] In some embodiments, the predicting of the first geometric structure of the foreign object under normal conditions based on the foreign object category and the foreign object structural features includes:

[0085] The foreign object category and the foreign object structural features are input into a geometric structure prediction model, and the geometric structure prediction model outputs the first geometric structure of the foreign object predicted by it under normal conditions; wherein, the geometric structure prediction model is obtained by fine-tuning the BERT large model.

[0086] In an embodiment of the present invention, in order to improve the accuracy of predicting the first geometric structure of a foreign object under normal circumstances, the present invention constructs a geometric structure prediction model. The geometric structure prediction model can be a dedicated vertical model, such as a model constructed based on algorithms such as convolutional networks and Transformers. However, the construction and training of such models require a lot of manpower and material resources, and the cost is high. In this regard, the present invention preferably uses the existing, general BERT large model to construct the above-mentioned geometric structure prediction model, that is, uses the collected small sample data to fine-tune the BERT large model to obtain a local model that can analyze and predict the first geometric structure of a foreign object under normal circumstances. In addition, the knowledge distillation technology can be used to further slim down the local model to reduce the size of the model, thereby improving the convenience of embedding the model into the automatic inspection vehicle.

[0087] like Figure 3 As shown, the embodiment of the present invention also discloses a novel track geometry parameter detection system, which is applied to an automatic detection vehicle, wherein the automatic detection vehicle includes a high-definition camera and a light sensor; the system includes a processing module and a storage module, wherein the processing module implements the following method steps by calling and executing the computer code in the storage module:

[0088] The light sensor detects the irradiation angle of sunlight in the current area, determines the shooting angle of the high-definition camera based on the irradiation angle of sunlight, and controls the high-definition camera to rotate to the shooting angle; wherein different shooting angles correspond to shooting fields of different sizes;

[0089] Performing track recognition on the high-definition image data obtained by the high-definition camera at the shooting angle to obtain two track segments, and performing foreign object detection on each of the track segments, and predicting the attribute category of the foreign object based on the foreign object features obtained by the detection;

[0090] Based on the distribution position of the foreign matter, the track segment is divided into a smooth track segment and a foreign matter track segment, the first track geometry parameter of each smooth track segment is extracted using image recognition technology, and the second track geometry parameter of the foreign matter track segment is calculated based on the attribute category;

[0091] The first orbital geometric parameters and the second orbital geometric parameters are aligned and integrated to obtain orbital geometric parameters.

[0092] An embodiment of the present invention further discloses an electronic device, comprising: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor executes the computer program to implement the method described in the above embodiment.

[0093] An embodiment of the present invention further discloses a computer storage medium, wherein the computer storage medium stores a computer program, and the computer program is executed by a processor to implement the method described in the above embodiment.

[0094] The embodiment of the present invention further discloses a computer program product. When the computer program product is run on a terminal, the terminal implements the method described in the above embodiment.

[0095] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, 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 device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0096] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1A function specified in one or more boxes.

[0097] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0098] The above description is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention.

Claims

1. A new method for detecting track geometry parameters, characterized in that , applied to an automatic inspection vehicle, the automatic inspection vehicle includes a high-definition camera and a light sensor; the method includes the following steps: The light sensor detects the irradiation angle of sunlight in the current area, determines the shooting angle of the high-definition camera based on the irradiation angle of sunlight, and controls the high-definition camera to rotate to the shooting angle; wherein different shooting angles correspond to shooting fields of different sizes; Performing track recognition on the high-definition image data obtained by the high-definition camera at the shooting angle to obtain two track segments, and performing foreign object detection on each of the track segments, and predicting the attribute category of the foreign object based on the foreign object features obtained by the detection; Based on the distribution position of the foreign matter, the track segment is divided into a smooth track segment and a foreign matter track segment, the first track geometry parameter of each smooth track segment is extracted using image recognition technology, and the second track geometry parameter of the foreign matter track segment is calculated based on the attribute category; The first orbital geometric parameters and the second orbital geometric parameters are aligned and integrated to obtain orbital geometric parameters.

2. A novel track geometry parameter detection method according to claim 1, characterized in that: Determining the shooting angle of the high-definition camera based on the illumination angle of the sun's rays includes: Determine the direction data of the track based on the high-precision map data, and calculate the reflection angle range of the reflected light of the sunlight at the illumination angle on the track according to the direction data and the illumination angle of the sunlight; Based on the three-dimensional size of the automatic inspection vehicle and the three-dimensional layout of the high-definition camera on the automatic inspection vehicle, the shooting angle that does not overlap with the reflection angle range is determined within the angle adjustment range of the high-definition camera, and the difference between the shooting angle and the current shooting angle of the high-definition camera is the smallest.

3. A novel track geometry parameter detection method according to claim 2, characterized in that: Performing foreign object detection on each track segment and predicting the attribute category of the foreign object based on the detected foreign object features includes: Extracting a first appearance feature of each of the track segments, performing a continuity analysis of the appearance feature based on the first appearance feature, and identifying a region where the appearance feature is discontinuous as the foreign matter; Extracting a second appearance feature of the foreign object, the second appearance feature includes a foreign object appearance feature and a foreign object structural feature, predicting the foreign object category of the foreign object based on the foreign object appearance feature and the foreign object structural feature, and predicting the geometric fullness and hardness of the foreign object based on the foreign object category and the foreign object structural feature; The foreign matter category, the geometric fullness and the hardness are used as the attribute categories of the foreign matter.

4. A novel track geometry parameter detection method according to claim 3, characterized in that: The geometric fullness and hardness of the foreign body are predicted based on the foreign body category and the foreign body structural characteristics, including: Predicting a first geometric structure of the foreign object under normal conditions based on the foreign object category and the foreign object structural features, and extracting a real-time second geometric structure of the foreign object from the foreign object structural features; The variation range of the geometric structure of the foreign body is calculated based on the first geometric structure and the second geometric structure, and the geometric fullness is determined based on the variation range; The dynamic characteristics of the foreign object are extracted from the structural characteristics of the foreign object, the dynamic value of the foreign object is calculated based on the dynamic characteristics, and the hardness of the foreign object is determined based on the dynamic value.

5. A novel method for detecting track geometric parameters according to claim 4, characterized in that: The automatic detection vehicle further includes an ultrasonic detector, and the second track geometric parameter of the foreign object track segment is calculated based on the attribute category, including: If the hardness is lower than the hardness threshold, the second track geometric parameters of the foreign object track segment are predicted based on the two first track geometric parameters adjacent to the foreign object track segment above and below; If the hardness is above the hardness threshold, then: If the geometric fullness is lower than the fullness threshold, the second track geometric parameters of the foreign object track segment are predicted based on the two first track geometric parameters adjacent to the foreign object track segment above and below; If the geometric fullness is higher than the fullness threshold, the ultrasonic detector is used to detect the foreign object track segment to extract the second track geometric parameters.

6. A novel method for detecting track geometric parameters according to claim 5, characterized in that: Predicting a first geometric structure of the foreign object under normal conditions based on the foreign object category and the foreign object structural characteristics includes: The foreign object category and the foreign object structural features are input into a geometric structure prediction model, and the geometric structure prediction model outputs the first geometric structure of the foreign object predicted by it under normal conditions; wherein, the geometric structure prediction model is obtained by fine-tuning the BERT large model.

7. A new type of track geometry parameter detection system, applied to an automatic detection vehicle, the automatic detection vehicle includes a high-definition camera and a light sensor; the system includes a processing module and a storage module; the characteristics are: The processing module implements the following method steps by calling and executing the computer code in the storage module: The light sensor detects the irradiation angle of sunlight in the current area, determines the shooting angle of the high-definition camera based on the irradiation angle of sunlight, and controls the high-definition camera to rotate to the shooting angle; wherein different shooting angles correspond to shooting fields of different sizes; Performing track recognition on the high-definition image data obtained by the high-definition camera at the shooting angle to obtain two track segments, and performing foreign object detection on each of the track segments, and predicting the attribute category of the foreign object based on the foreign object features obtained by the detection; Based on the distribution position of the foreign matter, the track segment is divided into a smooth track segment and a foreign matter track segment, the first track geometry parameter of each smooth track segment is extracted using image recognition technology, and the second track geometry parameter of the foreign matter track segment is calculated based on the attribute category; The first orbital geometric parameters and the second orbital geometric parameters are aligned and integrated to obtain orbital geometric parameters.

8. An electronic device comprising: At least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor executes the computer program to implement the method according to any one of claims 1 to 6.

9. A computer storage medium storing a computer program, characterized in that: The computer program is executed by a processor to implement the method according to any one of claims 1 to 6.

10. A computer program product comprising a computer program stored on a non-transitory computer readable medium, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images