Line feature region identification method and system for track geometric detection, and electronic equipment

Abstract

The invention belongs to the technical field of rail transit, particularly relates to a line feature region identification method and system for rail geometric detection and electronic equipment, and aims to solve the problem of insufficient reliability of a monitoring result of line feature information of the rail geometric detection. The method comprises the following steps: determining a to-be-identified track line feature image; inputting a track line feature image to be recognized into the line feature region recognition model to obtain a line feature region recognition result output by the line feature region recognition model; wherein the line feature region identification model is obtained based on sample data of a line feature image and region image annotation training of the sample data; the line feature region recognition model is used for performing multi-feature judgment on fusion data after space-time synchronization of the to-be-recognized track line feature image and the track geometric detection waveform data to obtain a line feature region recognition result. According to the invention, high reliability of monitoring results can be realized.

Description

Background Art

[0002] At present, the track line feature detection is mainly achieved by using the ALD on the track geometry detection beam carried by the detection vehicle, and cooperating with the on-board detection equipment to collect and process the sensor data. Its technical solution is mainly based on the system structure of "ALD-data acquisition industrial computer-processing system", as shown in the attached figure. Figure 2 As shown, data collection is achieved through ALD, data packaging and storage are achieved through industrial computers and data acquisition cards, and data processing and alignment are completed through algorithms in the processing system.

[0003] The existing technical solution is based on the principle of using the signal waveform output after removing the interference signal from the ALD level signal to display the output result. This requires the detection personnel to determine the line characteristic information at the current location through the characteristic signal waveform, which requires high experience of the personnel, and the waveform characteristics may not be able to truly reflect the line characteristic results. It requires on-site compounding or combined with other sensors for comprehensive analysis to make a judgment. It is more complicated to use, less reliable and the results are not intuitive. Summary of the invention

[0004] In order to solve the above-mentioned problems in the prior art, namely, the insufficient reliability of the monitoring results of the track geometry detection line characteristic information and the high experience requirement for the detection personnel, the present invention provides a line characteristic area identification method for track geometry detection, which achieves high reliability of the monitoring results.

[0005] In one aspect of the present invention, a method for identifying a line feature area for track geometry detection is provided, comprising: Determining a track line feature image to be identified; Inputting the track line feature image to be identified into a line feature region identification model to obtain a line feature region identification result output by the line feature region identification model; Wherein, the line feature region recognition model is obtained after training based on sample data of line feature images and regional image annotation of the sample data; The line feature area recognition model is used to perform multiple feature judgments on the fused data of the track line feature image to be identified and the track geometry detection waveform data after time and space synchronization to obtain the line feature area recognition result.

[0006] In some preferred embodiments, the line feature region recognition model includes a data acquisition unit, a feature judgment unit and a region recognition unit; The line feature area recognition model is used to perform multiple feature judgments on the fused data of the track line feature image to be identified and the track geometry detection waveform data after time and space synchronization to obtain the line feature area recognition result, including: Acquire, by the data acquisition unit, fused data of the track line feature image to be identified and the track geometry detection waveform data after time and space synchronization; The fused data is input into the feature judgment unit, and the feature judgment unit performs multiple feature judgments to filter out feature areas of corresponding types of features in the track line feature image to be identified; The feature area of ​​the corresponding type of feature in the track line feature image to be identified is input into the area identification unit, and the line feature area identification result is output.

[0007] In some preferred embodiments, before the data acquisition unit acquires the fused data of the track line feature image to be identified and the track geometry detection waveform data after time and space synchronization, the method includes: The track line feature image to be identified is collected by an image sensor, and the track geometry detection waveform data is collected by an eddy current meter; The eddy current meter is driven to start working by a trigger signal sent by an encoder, and a level signal is returned as a timestamp to achieve time synchronization between the track line feature image to be identified and the track geometry detection waveform data; The column data of the track line feature image to be identified, the track geometry detection waveform data and the spatial position of the real line are matched one by one to achieve spatial synchronization of the track line feature image to be identified and the track geometry detection waveform data.

[0008] In some preferred embodiments, the step of making the column data of the track line feature image to be identified, the track geometry detection waveform data and the spatial position of the real line correspond one to one to achieve spatial synchronization of the track line feature image to be identified and the track geometry detection waveform data comprises: Based on the number of triggers per revolution of the encoder and the wheel diameter and the line mileage during data collection, the data collection position of any frame of the image sensor is obtained to establish a spatial position correspondence between the characteristic image of the track line to be identified and the real line; The collected track geometry detection waveform data is offset according to the frame number based on the correction offset to obtain the alignment relationship between the track geometry detection waveform data and the middle position of the track line feature image to be identified; wherein the correction offset is calculated based on the number of triggers per circle of the encoder and the wheel diameter and the projection distance between the image sensor and the eddy current meter on the horizontal plane; Based on the coordinate correction value, the position of the track geometry detection waveform data aligned with the center position of the image sensor is corrected to obtain the corresponding relationship between the column data of the track line feature image to be identified and the track geometry detection waveform data; wherein the coordinate correction value is calculated based on the actual distance of the image sensor imaging projection relationship and the data sampling interval.

[0009] In some preferred embodiments, the correction offset is calculated based on the number of triggers per revolution of the encoder and the wheel diameter and the projection distance between the image sensor and the eddy current meter on the horizontal plane, and the formula is as follows: Q = (L1 × n) / (π × d); Among them, Q is the correction offset, L1 is the projection distance between the image sensor and the eddy current meter on the horizontal plane, n is the number of triggers per circle of the encoder, and d is the wheel diameter of the encoder.

[0010] In some preferred implementations, the coordinate correction value is calculated based on the actual distance of the image sensor imaging projection relationship and the data sampling interval, and the formula is as follows: R = (ΔL × n) / (π × d); Wherein, R is the coordinate correction value, ΔL is the actual distance of the imaging projection relationship of the image sensor, n is the number of triggers per circle of the encoder, and d is the wheel diameter of the encoder.

[0011] In some preferred embodiments, the step of inputting the fused data into the feature judgment unit, and performing multiple feature judgments by the feature judgment unit to screen out feature areas of corresponding types of features in the track line feature image to be identified includes: Based on the amplitude characteristics, amplitude characteristics, and rise and fall characteristics of the corresponding curves of the characteristic regions of preset standard types, the corresponding type characteristics of the fused data are screened once; Based on the width features corresponding to the feature areas in the curve, the fused data after the primary screening is secondary screened for features of corresponding types, so as to obtain feature areas of corresponding types of features in the feature image of the track line to be identified; wherein the width features corresponding to the feature areas in the curve are calculated based on the spatial position correspondence between the feature image of the track line to be identified and the real line and the fixed physical size of the line features.

[0012] In some preferred embodiments, the eddy current meter is arranged in the middle of the detection beam of the track geometry detection equipment; the encoder is arranged on the axle of the wheel of the track geometry detection equipment; the eddy current meter and the image sensor are located on the same horizontal plane; the track geometry detection equipment includes an urban transportation track geometry detection vehicle.

[0013] Another aspect of the present invention provides a line feature area recognition system for track geometry detection, comprising: An image determination unit, used to determine a track line feature image to be identified; A region recognition unit, used for inputting the track line feature image to be recognized into a line feature region recognition model, and obtaining a line feature region recognition result output by the line feature region recognition model; Wherein, the line feature region recognition model is obtained after training based on sample data of line feature images and regional image annotation of the sample data; The line feature area recognition model is used to perform multiple feature judgments on the fused data of the track line feature image to be identified and the track geometry detection waveform data after time and space synchronization to obtain the line feature area recognition result.

[0014] According to a third aspect of the present invention, an electronic device is provided, comprising: at least one processor; and a memory communicatively coupled to at least one of the processors; The memory stores instructions executable by the processor, and the instructions are used to be executed by the processor to implement the line feature area recognition method for track geometry detection.

[0015] Beneficial effects of the present invention: The present invention provides analysis of image data synchronized with waveform signals, which is compatible with the existing data analysis and processing system, and can provide relatively accurate image information, improve data reliability, and reduce the experience requirements for operators. It can intuitively display the line status at the current location, so that track maintenance personnel can accurately and conveniently associate the collected regional feature data with the on-site situation, which greatly simplifies the process of line feature identification and reduces the difficulty of identification, and is of great significance for solving the objective problems existing on-site. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings: Figure 1 It is a flow chart of a line feature area recognition method for track geometry detection; Figure 2 is a schematic diagram of an existing track line feature detection device; Figure 3 is a schematic diagram of the improved track geometry detection device of the present invention; Figure 4 It is a structural diagram of the line feature area recognition model; Figure 5 It is a structural diagram of the line feature area recognition system for track geometry detection; Figure 6 It is a schematic diagram of the structure of a computer system of a server for implementing the method, system, and electronic device embodiments of the present application; Figure 7 It is a schematic diagram of an image after the image sensor captures the image, performs grayscale processing and removes irrelevant areas; Figure 8 It is a schematic diagram of the feature area of ​​the image captured by the image sensor and identified by the depth model; Fig. 9 It is a schematic diagram of converting the image into waveform information by column projection; Fig.10 It is a schematic diagram of the waveform characteristics of ALD data in different sections; Fig.11 It is a schematic diagram of the line characteristics of the positive line area identified in the actual detection; Fig.12 It is a schematic diagram of the line characteristics of the turnout area identified in the actual inspection; Fig.13 It is a schematic diagram of the line characteristics of the transponder area identified in the actual detection; Fig.14 It is a detailed flow chart of the line feature region recognition method for geometric detection. DETAILED DESCRIPTION

[0017] The present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the relevant invention are shown in the accompanying drawings.

[0018] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0019] The present invention combines image sensors and traditional eddy current sensors to perform data fusion analysis and processing to improve the reliability of detection results, enhance the credibility of data and the persuasiveness of detection results, and reduce the experience requirements for detection personnel. The details are as follows: The present invention relates to a related device mounted on a train for urban transportation track geometry detection (i.e., track geometry detection equipment), which is composed of an ALD, an image sensor, an encoder, a data acquisition industrial computer, a data analysis industrial computer, and a system power supply. Figure 3 The functions of each component are as follows: (1) ALD (eddy current sensor, i.e. Figure 3The eddy current meter in the instrument is installed in the middle of the detection beam and uses the principle that the eddy current in the object to be measured generates an induced electromagnetic field for measurement. When the current in the excitation coil is passed, an alternating magnetic field is generated, causing the generation of eddy currents in the object to be measured. The eddy currents cause the loss of magnetic flux due to their own damping effect. This in turn causes a change in the electrical level, thereby obtaining the original data of the characteristic information.

[0020] (2) Image sensor (i.e. Figure 3 Visual sensor in: Using a frame-triggered high-speed camera and shooting vertically downward, high-speed and high-precision image acquisition can be achieved.

[0021] (3) Encoder: It is installed on the axle of the wheel of the track geometry detection equipment and is used to convert the angular displacement of the wheel into a pulse signal.

[0022] (4) Data acquisition industrial computer (i.e. Figure 3 Real-time acquisition processor in): converts the pulse signal provided by the encoder into a trigger signal, sends a trigger signal to each sensor, and collects and stores the data information of each sensor.

[0023] (5) Data analysis industrial computer (i.e. Figure 3 Data analysis server in the vehicle): obtain, analyze and process the data information stored in the real-time acquisition processor through the vehicle LAN.

[0024] (6) System power supply: Provide power for each sensor. The data acquisition industrial computer, data analysis industrial computer, and system power supply are all installed in the scale structure of the train for urban transportation track geometry detection; Based on the above device, the present invention provides a line feature area recognition method for track geometry detection, comprising: Determining a track line feature image to be identified; Inputting the track line feature image to be identified into a line feature region identification model to obtain a line feature region identification result output by the line feature region identification model; Wherein, the line feature region recognition model is obtained after training based on sample data of line feature images and regional image annotation of the sample data; The line feature area recognition model is used to perform multiple feature judgments on the fused data of the track line feature image to be identified and the track geometry detection waveform data after time and space synchronization to obtain the line feature area recognition result.

[0025] In order to more clearly explain the line feature area recognition method of track geometry detection of the present invention, the following is combined with Figure 1 , Fig.14 Each step in the embodiment of the present invention is described in detail.

[0026] The line feature area recognition method for track geometry detection of the first embodiment of the present invention is as follows: Figure 1 As shown, it includes step S101-step S102, and each step is described in detail as follows: Step S101, determining a track line feature image to be identified; Step S102, inputting the track line feature image to be identified into a line feature region recognition model, and obtaining a line feature region recognition result output by the line feature region recognition model; Wherein, the line feature region recognition model is obtained after training based on sample data of line feature images and regional image annotation of the sample data; Specifically, different areas have obvious line feature images. By analyzing the data sets obtained by collecting a large amount of sample data in the corresponding feature intervals, deep learning annotation training is used to identify the above sample data interval images, thereby obtaining a deep learning model that can preliminarily judge the line characteristics.

[0027] The line feature area recognition model is used to perform multiple feature judgments on the fused data of the track line feature image to be identified and the track geometry detection waveform data after time and space synchronization to obtain the line feature area recognition result.

[0028] The present invention provides analysis of image data synchronized with waveform signals, which is compatible with the existing data analysis and processing system, and can provide relatively accurate image information, improve data reliability, and reduce the experience requirements for operators. It can intuitively display the line status at the current location, so that track maintenance personnel can accurately and conveniently associate the collected regional feature data with the on-site situation, which greatly simplifies the process of line feature identification and reduces the difficulty of identification, and is of great significance for solving the objective problems existing on-site.

[0029] Although the various steps in the above embodiment are described in the above-mentioned order, those skilled in the art can understand that in order to achieve the effect of this embodiment, different steps do not have to be executed in such an order. They can be executed simultaneously (in parallel) or in a reverse order. These simple changes are within the scope of protection of the present invention.

[0030] Based on the above embodiments, Figure 4 As shown, the line feature region recognition model includes a data acquisition unit 201, a feature judgment unit 202 and a region recognition unit 203; The line feature area recognition model is used to perform multiple feature judgments on the fused data of the track line feature image to be identified and the track geometry detection waveform data after time and space synchronization to obtain the line feature area recognition result, including: The data acquisition unit 201 acquires the fused data of the track line feature image to be identified and the track geometry detection waveform data after time and space synchronization; The fused data is input into the feature judgment unit 202, and the feature judgment unit 202 performs multiple feature judgments to filter out feature areas of corresponding types of features in the track line feature image to be identified; The feature region of the corresponding type of feature in the track line feature image to be identified is input into the region identification unit 203, and the line feature region identification result is output.

[0031] Based on the above embodiment, before the data acquisition unit acquires the fused data of the track line feature image to be identified and the track geometry detection waveform data after time and space synchronization, it includes: The track line feature image to be identified is collected by an image sensor, and the track geometry detection waveform data is collected by an eddy current meter; The eddy current meter is driven to start working by a trigger signal sent by an encoder, and a level signal is returned as a timestamp to achieve time synchronization between the track line feature image to be identified and the track geometry detection waveform data; Specifically, when the vehicle moves, the encoder installed on the wheel axle sends out a pulse signal, which is converted into a trigger signal after frequency division. The visual sensor starts working when it receives the trigger signal and returns a level signal as a timestamp to achieve time synchronization.

[0032] The column data of the track line feature image to be identified, the track geometry detection waveform data and the spatial position of the real line are matched one by one to achieve spatial synchronization of the track line feature image to be identified and the track geometry detection waveform data.

[0033] Specifically, we establish (1) the spatial position relationship between the camera data and the real line; (2) the corresponding relationship between the middle position of each frame image and the ALD data; and (3) the corresponding relationship between each column of data in a frame image and the ALD data. Based on the above corresponding relationships, any column of camera data, ALD data and the spatial position of the real line can be matched to achieve spatial synchronization.

[0034] Based on the above embodiment, the column data of the track line feature image to be identified, the track geometry detection waveform data and the spatial position of the real line are made to correspond one to one to achieve spatial synchronization of the track line feature image to be identified and the track geometry detection waveform data, including: Based on the number of triggers per revolution of the encoder and the wheel diameter and the line mileage during data collection, the data collection position of any frame of the image sensor is obtained to establish a spatial position correspondence between the characteristic image of the track line to be identified and the real line; Specifically, according to the encoder principle, if the number of encoder triggers per revolution is n and the wheel diameter is d, and the line mileage when the acquisition starts is recorded as X0, then the camera's mth frame data acquisition position is X0+(m×π×d) / n, thereby establishing the spatial position relationship between the camera data and the actual line.

[0035] The collected track geometry detection waveform data is offset according to the frame number based on the correction offset, so as to achieve the alignment relationship between the track geometry detection waveform data and the middle position of the track line characteristic image to be identified; wherein the correction offset is calculated based on the number of triggers per circle of the encoder and the wheel diameter and the projection distance between the image sensor and the eddy current meter on the horizontal plane; Specifically, the projection distance between the visual sensor camera lens position and the ALD installation position on the horizontal plane is L1. The number of encoder triggers per revolution (n) and the wheel diameter (d) can be used to calculate the length trigger times to be (L1×n) / (π×d), that is, the frame difference between the camera shooting position and the ALD measurement position is (L1×n) / (π×d). From the above, it can be seen that the ALD acquisition data is corrected, and the acquired ALD data is offset according to the frame number. When the correction offset Q is (L1×n) / (π×d), the ALD data can be aligned with the middle position of the camera captured image.

[0036] Based on the coordinate correction value, the position of the track geometry detection waveform data aligned with the center position of the image sensor is corrected to obtain the corresponding relationship between the column data of the track line feature image to be identified and the track geometry detection waveform data; wherein the coordinate correction value is calculated based on the actual distance of the image sensor imaging projection relationship and the data sampling interval.

[0037] Specifically, the position of the ALD data aligned with the center of the camera image is denoted as A0. Because the relative position of the camera and the rail is determined by the structure of the acquisition device and can be regarded as a constant, the relationship between the width of the captured image and the width of the corresponding area on the actual line can be obtained based on the camera imaging projection relationship. Let the number of columns that differ between a column of data in the image and the middle position be Δw, and the actual distance between the two obtained based on the camera imaging projection relationship be ΔL. It is known that the sampling interval should be (π×d) / n, then the ALD data corresponding to the column of data and the coordinate correction value R at A0 are (ΔL×n) / (π×d). Thus, the corresponding relationship between each column of image data and ALD data can be obtained.

[0038] Based on the above embodiment, the step of inputting the fusion data into the feature judgment unit, and performing multiple feature judgments by the feature judgment unit to filter out feature areas of corresponding types of features in the track line feature image to be identified includes: Based on the amplitude characteristics, amplitude characteristics, and rise and fall characteristics of the corresponding curves of the characteristic regions of preset standard types, the corresponding type characteristics of the fused data are screened once; Based on the width features corresponding to the feature areas in the curve, the fused data after the primary screening is secondary screened for features of corresponding types, so as to obtain feature areas of corresponding types of features in the feature image of the track line to be identified; wherein the width features corresponding to the feature areas in the curve are calculated based on the spatial position correspondence between the feature image of the track line to be identified and the real line and the fixed physical size of the line features.

[0039] Specifically, the photos are taken, grayscale processed and irrelevant areas are removed, such as Figure 7 As shown, the feature range of interest is taken, and the feature area output result can be identified through deep learning model processing, such as Figure 8 shown.

[0040] In addition, the grayscale image is vertically projected, and the image as a whole is moved to the left by dw and upward by dh. Then, for any (h,w)∈image pixel range, the (h,w) pixel of the new image and the (h+dh,w+dw) of the original image should be exactly the same. That is, the hth row of the new image and the h+dhth row of the original image should be exactly the same, and the wth column of the new image and the w+dwth column of the original image should be exactly the same. That is, this method can be used to obtain a grayscale waveform with the same trend as the original image.

[0041] The formulas for row grayscale projection and column grayscale projection are as follows: ; In the above formula, Gk(i,j) represents the grayscale value of the pixel at (i,j) of the k-th frame image, Gk(i) represents the grayscale value of the pixel in the i-th row of the image, and Gk(j) represents the grayscale value of the pixel in the j-th column of the image.

[0042] Therefore, the above image can be transformed into waveform information by column projection to obtain the projection result, such as Fig. 9 shown.

[0043] It can be seen that the grayscale projection waveforms of different feature segments are significantly different from the positive line. Data collection can be performed on various feature areas manually selected on the line, and this can be used as standard data for feature recognition to establish a recognition algorithm: First, the obtained data is denoised to remove isolated points to avoid interference points affecting the results; then the general positive line area is used as a benchmark to perform max-min standardization on the curve. The amplitude characteristics, amplitude characteristics, and rise and fall characteristics of the corresponding curve of this type of area can be determined based on the above standard area data; from the above, it can be seen that the spatial synchronization process has established the relationship between each column of image data and the actual physical position, and the physical size of the line feature is fixed, so the width characteristics corresponding to the feature area in the curve can be calculated; finally, two-step screening is performed according to the above characteristics, and all areas that meet the corresponding characteristics are the corresponding feature areas. Traditional feature area judgment mainly relies on the fact that the ALD data will have obvious characteristic values ​​when passing through different sections, and the line characteristics can be roughly screened according to the waveform characteristics. Fig.10 As shown; In actual detection, the deep learning model is first used to determine and select the interval of interest, and then the image of the area is processed by grayscale projection to obtain the grayscale projection curve of the target area. The recognition algorithm can determine the line characteristics. At the same time, the recognition algorithm is compatible with the existing ALD waveform and participates in the feature area judgment together, so that the line characteristics of the section can be more accurately identified, such as Fig.11 , 12 , as shown in 13.

[0044] The line feature area recognition system for track geometry detection of the second embodiment of the present invention is as follows: Figure 5 As shown, including: An image determination unit 301 is used to determine a track line feature image to be identified; The region identification unit 302 is used to input the track line feature image to be identified into a line feature region identification model to obtain a line feature region identification result output by the line feature region identification model; Wherein, the line feature region recognition model is obtained after training based on sample data of line feature images and regional image annotation of the sample data; The line feature area recognition model is used to perform multiple feature judgments on the fused data of the track line feature image to be identified and the track geometry detection waveform data after time and space synchronization to obtain the line feature area recognition result.

[0045] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process and related instructions of the system described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0046] It should be noted that the line feature area identification system for track geometry detection provided in the above embodiment is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the modules or steps in the embodiments of the present invention can be decomposed or combined. For example, the modules in the above embodiment can be combined into one module, or further divided into multiple sub-modules to complete all or part of the functions described above. The names of the modules and steps involved in the embodiments of the present invention are only for distinguishing the modules or steps, and are not regarded as improper limitations of the present invention.

[0047] An electronic device according to a third embodiment of the present invention includes: at least one processor; and a memory communicatively coupled to at least one of the processors; The memory stores instructions executable by the processor, and the instructions are used to be executed by the processor to implement the line feature area recognition method for track geometry detection.

[0048] A fourth embodiment of the present invention is a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to be executed by the computer to implement the line feature area recognition method for track geometry detection.

[0049] Technicians in the relevant technical field can clearly understand that, for the convenience and brevity of description, the specific working process and related instructions of the electronic device and computer-readable storage medium described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0050] Those skilled in the art should be able to appreciate that the modules and method steps of each example described in conjunction with the embodiments disclosed herein can be implemented with electronic hardware, computer software, or a combination of the two, and the programs corresponding to the software modules and method steps can be placed in random access memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, or any other form of storage medium known in the technical field. In order to clearly illustrate the interchangeability of electronic hardware and software, the composition and steps of each example have been generally described in the above description according to the function. Whether these functions are performed in electronic hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0051] Reference below Figure 6, which shows a schematic diagram of the structure of a computer system of a server for implementing the method, system, and device embodiments of the present application. Figure 6 The server shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0052] like Figure 6 As shown, the computer system includes a central processing unit (CPU) 601, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 602 or the program loaded from the storage part 608 to the random access memory (RAM) 603. In the RAM 603, various programs and data required for system operation are also stored. The CPU 601, ROM 602 and RAM 603 are connected to each other through a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0053] The following components are connected to the I / O interface 605: an input section 606 including a keyboard, a mouse, etc.; an output section 607 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as needed. A removable medium 611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 610 as needed so that a computer program read therefrom is installed into the storage section 608 as needed.

[0054] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 609, and / or installed from the removable medium 611. When the computer program is executed by the central processing unit (CPU) 601, the above-mentioned functions defined in the method of the present application are executed. It should be noted that the above-mentioned computer-readable medium of the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, - but not limited to - an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection with one or more conductors, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, an apparatus or a device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in combination with an instruction execution system, an apparatus or a device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0055] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0056] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0057] The terms "first", "second", etc. are used to distinguish similar objects rather than to describe or indicate a particular order or sequence.

[0058] The term "comprise" or any other similar term is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus / device that includes a list of elements includes not only those elements but also other elements not expressly listed, or also includes elements inherent to such process, method, article, or apparatus / device.

[0059] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. A method for identifying line feature areas for track geometry detection, characterized in that: include: Determining a track line feature image to be identified; Inputting the track line feature image to be identified into a line feature region identification model to obtain a line feature region identification result output by the line feature region identification model; Wherein, the line feature region recognition model is obtained after training based on sample data of line feature images and regional image annotation of the sample data; The line feature area recognition model is used to perform multiple feature judgments on the fused data of the track line feature image to be identified and the track geometry detection waveform data after time and space synchronization to obtain the line feature area recognition result.

2. The line feature area recognition method for track geometry detection according to claim 1, characterized in that: The line feature region recognition model includes a data acquisition unit, a feature judgment unit and a region recognition unit; The line feature area recognition model is used to perform multiple feature judgments on the fused data of the track line feature image to be identified and the track geometry detection waveform data after time and space synchronization to obtain the line feature area recognition result, including: Acquire, by the data acquisition unit, fused data of the track line feature image to be identified and the track geometry detection waveform data after time and space synchronization; The fused data is input into the feature judgment unit, and the feature judgment unit performs multiple feature judgments to filter out feature areas of corresponding types of features in the track line feature image to be identified; The feature area of ​​the corresponding type of feature in the track line feature image to be identified is input into the area identification unit, and the line feature area identification result is output.

3. The line feature area recognition method for track geometry detection according to claim 2, characterized in that: Before the data acquisition unit acquires the fused data of the track line feature image to be identified and the track geometry detection waveform data after time and space synchronization, the method includes: The track line feature image to be identified is collected by an image sensor, and the track geometry detection waveform data is collected by an eddy current meter; The eddy current meter is driven to start working by a trigger signal sent by an encoder, and a level signal is returned as a timestamp to achieve time synchronization between the track line feature image to be identified and the track geometry detection waveform data; The column data of the track line feature image to be identified, the track geometry detection waveform data and the spatial position of the real line are matched one by one to achieve spatial synchronization of the track line feature image to be identified and the track geometry detection waveform data.

4. The line feature area recognition method for track geometry detection according to claim 3 is characterized in that: The step of making the column data of the track line feature image to be identified, the track geometry detection waveform data and the spatial position of the real line correspond one to one to achieve spatial synchronization of the track line feature image to be identified and the track geometry detection waveform data includes: Based on the number of triggers per revolution of the encoder and the wheel diameter and the line mileage during data collection, the data collection position of any frame of the image sensor is obtained to establish a spatial position correspondence between the characteristic image of the track line to be identified and the real line; The collected track geometry detection waveform data is offset according to the frame number based on the correction offset to obtain the alignment relationship between the track geometry detection waveform data and the middle position of the track line feature image to be identified; wherein the correction offset is calculated based on the number of triggers per circle of the encoder and the wheel diameter and the projection distance between the image sensor and the eddy current meter on the horizontal plane; Based on the coordinate correction value, the position of the track geometry detection waveform data aligned with the center position of the image sensor is corrected to obtain the corresponding relationship between the column data of the track line feature image to be identified and the track geometry detection waveform data; wherein the coordinate correction value is calculated based on the actual distance of the image sensor imaging projection relationship and the data sampling interval.

5. The line feature area recognition method for track geometry detection according to claim 4 is characterized in that: The correction offset is calculated based on the number of triggers per revolution of the encoder and the wheel diameter and the projection distance of the image sensor and the eddy current meter on the horizontal plane, and the formula is as follows: Q = (L1 × n) / (π × d); Among them, Q is the correction offset, L1 is the projection distance between the image sensor and the eddy current meter on the horizontal plane, n is the number of triggers per circle of the encoder, and d is the wheel diameter of the encoder.

6. The line feature area recognition method for track geometry detection according to claim 4, characterized in that: The coordinate correction value is calculated based on the actual distance of the image sensor imaging projection relationship and the data sampling interval, and its formula is as follows: R = (ΔL × n) / (π × d); Wherein, R is the coordinate correction value, ΔL is the actual distance of the imaging projection relationship of the image sensor, n is the number of triggers per circle of the encoder, and d is the wheel diameter of the encoder.

7. The line feature area recognition method for track geometry detection according to claim 3 is characterized in that: The step of inputting the fused data into the feature judgment unit, and performing multiple feature judgments by the feature judgment unit to screen out feature areas of corresponding types of features in the track line feature image to be identified, includes: Based on the amplitude characteristics, amplitude characteristics, and rise and fall characteristics of the corresponding curves of the characteristic regions of preset standard types, the corresponding type characteristics of the fused data are screened once; Based on the width features corresponding to the feature areas in the curve, the fused data after the primary screening is secondary screened for features of corresponding types, so as to obtain feature areas of corresponding types of features in the feature image of the track line to be identified; wherein the width features corresponding to the feature areas in the curve are calculated based on the spatial position correspondence between the feature image of the track line to be identified and the real line and the fixed physical size of the line features.

8. The line feature area recognition method for track geometry detection according to claim 3 is characterized in that: The eddy current meter is arranged in the middle of the detection beam of the track geometry detection device; the encoder is arranged on the axle of the wheel of the track geometry detection device; the eddy current meter and the image sensor are located on the same horizontal plane; The track geometry detection equipment includes an urban transportation track geometry detection vehicle.

9. A line feature area recognition system for track geometry detection, characterized in that: include: An image determination unit, used to determine a track line feature image to be identified; A region recognition unit, used for inputting the track line feature image to be recognized into a line feature region recognition model, and obtaining a line feature region recognition result output by the line feature region recognition model; Wherein, the line feature region recognition model is obtained after training based on sample data of line feature images and regional image annotation of the sample data; The line feature area recognition model is used to perform multiple feature judgments on the fused data of the track line feature image to be identified and the track geometry detection waveform data after time and space synchronization to obtain the line feature area recognition result.

10. An electronic device, characterized in that: include: at least one processor; and a memory communicatively coupled to at least one of the processors; Wherein, the memory stores instructions executable by the processor, and the instructions are used to be executed by the processor to implement the line feature area identification method for track geometry detection as described in any one of claims 1-8.

Images