A Blind Correction Method and System for Image Distortion of Linear Array Cameras on the Whole Track Train

Abstract

The present invention discloses a method and system for blind correction of image distortion of a whole train linear array camera, which relates to the technical field of rail transit. It solves the problems that the existing train image distortion correction methods have large errors or rely on prior information. The key technical points are as follows: The whole train image of the train is captured by two or more linear array cameras that are not on the same cross-section, and the train images captured by the linear array cameras at multiple angles are used to perform blind correction on themselves; when the solution is applied in the field of rail transit, there is no need to set scale marks on the train to be captured, nor prior information such as standard template images or train component scales and markers. The application premise is more relaxed, and the train image can be directly corrected, realizing efficient and stable blind correction of image distortion of the whole train linear array camera of the rail train.

Description

Technical Field

[0001] The present invention relates to the field of rail transit technology, and more specifically, to a method and system for blindly correcting image distortion of a linear array camera of a whole rail train. Background Art

[0002] In the field of rail transit, since the length of a train is much greater than its width and height, linear array cameras are often the first choice for fine imaging of the train's appearance. If the shooting frequency of a linear array camera is fixed, the linear array camera will inevitably capture distorted train images because it is impossible for a train to pass through the linear array camera at the same agreed calibration speed (i.e., the speed that matches the shooting frequency) every time.

[0003] For the above problems, on the one hand, most of the existing processing methods are to measure the speed of the train and change the shooting frequency of the linear array camera according to its instantaneous movement speed (because the linear array camera module is installed on the trackside, it is impossible to drive the linear array camera to shoot through the train's own speed module). The existing speed measurement forms are divided into two types: magnetic steel speed measurement and radar speed measurement. The magnetic steel speed measurement has high accuracy. Its principle is to calculate the train speed by using the signal time difference generated by the train wheel pair passing through the magnetic steel, so the speed measurement frequency is far from meeting the requirements; the radar speed measurement frequency is higher, but still lower than the linear array camera shooting frequency. In addition, the error percentage of radar speed measurement is high when the train is at low speed. Therefore, using the speed measurement module to perform variable frequency shooting of the train cannot fundamentally solve the distortion problem of the train image taken by the linear array camera. On the other hand, image correction can be performed through algorithms. For example, scale marks (or markers) can be set on the train, and the correction of the train image can be indirectly completed by correcting the markers; or the standard template image method can be used to match the feature information in the standard image with the feature information in the image to be corrected one by one to complete the correction of the train image. However, such schemes rely on a large amount of prior information or preconditions such as standard image templates or other train component scales and markers. Image correction cannot be performed if these prior information or preconditions are not met.

[0004] Therefore, the present application provides a method and system for blind correction of image distortion of a linear array camera of a railway train to solve the above-mentioned problems. Summary of the invention

[0005] The purpose of the present application is to provide a method and system for blind correction of image distortion of a linear array camera of a whole railway train, so as to solve the problem that the existing train image distortion correction methods have large errors or rely on prior information. The present application uses at least two linear array cameras that are not in the same cross-section to perform blind correction on synchronously shot train images, so as to realize efficient and stable train image distortion correction that is independent of prior information.

[0006] The present application first provides a method for blind correction of image distortion of a whole train line array camera, including: S1. Synchronously capturing multiple train images by at least two line array cameras located at different cross-sections, where the cross-section refers to the track cross-section; S2. Using the line array camera that first captures the train image as the reference camera and other line array cameras as offset cameras, measuring the physical offset distance between each offset camera and the reference camera, and solving the image offset distance between each offset camera and the reference camera according to the physical offset distance and the horizontal resolution of the line array camera imaging; S3. Extracting N marker points from the train images captured by each line array camera, where the N marker points in different train images correspond to each other on the train cross-section, extracting the x coordinates of the marker points in the train images of the reference camera and the offset cameras, and marking the x coordinates of the marker points in the train images of each offset camera that are within the coordinate range of the reference camera; S4. Assuming that the correction amount of the extracted x coordinates of the marker points is an unknown quantity, establishing an equal cross-section equation based on the characteristics corresponding to the marker point cross-sections in the train images captured by the offset camera and the reference camera, and establishing a symmetry equation based on the symmetric characteristics at both ends of the train; S5. Establishing a linear equation system according to the equal cross-section equation and the symmetry equation, and solving the linear equation system to obtain the correction amount; S6. Transforming the N marker points in the train image captured by the line array camera based on the correction amount to obtain the train image after distortion correction.

[0007] In a possible implementation manner, the equal cross-section equation is:

[0008] ;

[0009] Wherein, is the x coordinate of the th marker point in the train image of the reference camera 、 The correction amount of the th marker point in the train image of the reference camera 、 Offset_x is the image offset distance between the offset camera and the reference camera, is the x coordinate of the th marker point in the train image of the offset camera, is the th marker point in the train image of the offset camera.

[0010] In a possible implementation manner, for the correction amount of the marker points in the train images of each offset camera that are within the coordinate range of the reference camera, assuming only linear distortion exists, it is expressed as:

[0011] ;

[0012] Among them, is the correction amount of the -th marker point in the offset camera train image. At this time, the -th marker point is within the range of the reference camera coordinates; and are the x coordinates of the j -th and j +1-th marker points in the reference camera train image, and are the correction amounts of the j -th and j +1-th marker points in the reference camera train image, is the x coordinate of the -th marker point in the offset camera train image, located between and .

[0013] In a possible implementation manner, the symmetric equation is:

[0014] ;

[0015] Among them, and are the x coordinates of the marker points of specific components in the first half formation of the reference camera train image, and are and 's correction amounts; and are the x coordinates of the marker points of the same type of specific components in the second half formation of the offset camera train image, and are and 's correction amounts.

[0016] The present application also provides a blind image distortion correction system for a whole vehicle of an orbital train, including: an image acquisition unit for synchronously capturing multiple train images based on at least two linear array cameras located at different cross-sections, where the cross-section refers to the orbital cross-section; an offset calculation unit for taking the linear array camera that first captures the train image as the reference camera and other linear array cameras as offset cameras, measuring the physical offset distance between each offset camera and the reference camera, and solving the image offset distance between each offset camera and the reference camera according to the physical offset distance and the horizontal resolution of the linear array camera imaging; a marker point extraction unit for extracting N marker points from the train images captured by each linear array camera, where the N marker points in different train images correspond to each other on the train cross-section, extracting the x coordinates of the marker points in the train images of the reference camera and the offset cameras, and marking the x coordinates of the marker points in the train images of each offset camera that are within the coordinate range of the reference camera; an equation establishment unit for assuming the correction amount of the extracted x coordinates of the marker points as unknowns, establishing an equation of the same cross-section based on the characteristics corresponding to the marker point cross-sections in the train images captured by the offset camera and the reference camera, and establishing a symmetry equation based on the symmetric characteristics at both ends of the train; an equation solving unit for establishing a linear equation system according to the equation of the same cross-section and the symmetry equation, and solving the linear equation system to obtain the correction amount; a distortion correction unit for transforming the N marker points in the train images captured by the linear array camera based on the correction amount to obtain a train image after distortion correction.

[0017] In a possible implementation manner, in the equation establishment unit, the equation of the same cross-section is:

[0018] ;

[0019] Wherein, is the x coordinate of the th marker point in the train image of the reference camera 、 The correction amount of the th marker point in the train image of the reference camera 、 Offset_x is the image offset distance between the offset camera and the reference camera, is the x coordinate of the th marker point in the train image of the offset camera, is the th marker point in the train image of the offset camera.

[0020] In a possible implementation, in the solution formula establishing unit, for the correction amounts of the marked points in each offset camera train image that are within the reference camera coordinate range, assuming only linear distortion exists, it is expressed by the x coordinate of the marked point in the known reference camera train image, the correction amount of the marked point in the reference camera train image, and the x coordinate of the marked point in the offset camera train image as:

[0021] ;

[0022] wherein, is the correction amount of the th marked point in the offset camera train image, and at this time the th marked point is within the reference camera coordinate range; and are the x coordinates of the j and j +1th marked points in the reference camera train image, and are the correction amounts of the j and j +1th marked points in the reference camera train image, is the x coordinate of the th marked point in the offset camera train image, and is between and .

[0023] In a possible implementation, in the solution formula establishing unit, the symmetric equation is:

[0024] ;

[0025] wherein, and are the x coordinates of the marked points of the specific components in the first half formation of the reference camera train image, and are the correction amounts of and ; and are the x coordinates of the marked points of the same type of specific components in the second half formation of the offset camera train image, and are the correction amounts of and .

[0026] This application also provides a computer storage medium, including computer instructions, which when running on a mobile terminal, cause the mobile terminal to execute the method as described above.

[0027] The present application also provides a computer program product, which, when running on a mobile terminal, causes the mobile terminal to execute the method as described above.

[0028] Compared with the prior art, the present application has the following beneficial effects: by using two or more linear array cameras that are not on the same cross-section to capture the whole train image, and using the train images captured by the linear array cameras from multiple angles to perform blind calibration on itself; when this method is applied to the rail transit field, it is not necessary to set scale marks on the train to be photographed, nor is it necessary to have prior information such as a standard template image or the scale of train components and markers. The application premise is more relaxed, and the train image can be directly calibrated to achieve efficient and stable blind calibration of the distortion of the linear array camera image of the whole rail train. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, form a part of the present application, and do not limit the embodiments of the present invention. In the drawings:

[0030] Figure 1 is a flowchart of the method for blind calibration of the distortion of the linear array camera image of the whole rail train provided by the embodiment of the present application;

[0031] Figure 2 is a layout diagram of the linear array camera;

[0032] Figure 3 is a schematic diagram of the marked points of the train images captured by the reference camera and the offset camera;

[0033] Figure 4 is a schematic diagram of the train images captured by the reference camera and the offset camera after distortion correction;

[0034] Figure 5 is a structural diagram of the system for blind calibration of the distortion of the linear array camera image of the whole rail train provided by the embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0035] In the following, the term "comprising" or "may comprise" that can be used in various embodiments of the present application indicates the presence of the claimed functions, operations, or elements, and does not limit the addition of one or more functions, operations, or elements. In addition, as used in various embodiments of the present application, the terms "comprising", "having" and their cognates are only intended to indicate a specific feature, number, step, operation, element, component, or combination of the foregoing items, and should not be construed as precluding the existence or addition of the possibility of one or more other features, numbers, steps, operations, elements, components, or combinations of the foregoing items.

[0036] The terms used in the various embodiments of the present application are for the purpose of describing specific embodiments only and are not intended to limit the various embodiments of the present application. As used herein, the singular forms are intended to include the plural forms as well, unless the context clearly indicates otherwise. Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of the present application belong. The terms (such as those defined in a commonly used dictionary) will be interpreted as having the same meaning as the contextual meaning in the relevant technical field and will not be interpreted as having an idealized meaning or an overly formal meaning, unless clearly defined in the various embodiments of the present application.

[0037] To make the objectives, technical solutions, and advantages of the present application more clearly understood, the present application will be further described in detail below with reference to the embodiments and the accompanying drawings. The illustrative embodiments and descriptions thereof of the present application are only used to explain the present application and do not limit the present application.

[0038] Please refer to Figure 1 as shown in Figure 1 is a flowchart of a method for blind correction of image distortion of a whole train linear array camera provided by an embodiment of the present application. The method includes: S1. Synchronously capturing multiple train images based on at least two linear array cameras located at different cross-sections, where the cross-section refers to the track cross-section; S2. Using the linear array camera that first captures the train image as the reference camera and other linear array cameras as offset cameras, measuring the physical offset distance between each offset camera and the reference camera, and solving the image offset distance between each offset camera and the reference camera according to the physical offset distance and the horizontal resolution of the linear array camera imaging; S3. Extracting N marker points from the train images captured by each linear array camera, where the N marker points in different train images correspond to each other on the train cross-section, extracting the x coordinates of the marker points in the train images of the reference camera and the offset cameras, and marking the x coordinates of the marker points in the train images of each offset camera that are within the coordinate range of the reference camera; S4. Assuming that the correction amount of the extracted x coordinates of the marker points is an unknown quantity, establishing a same cross-section equation based on the characteristics corresponding to the marker point cross-sections in the train images captured by the offset camera and the reference camera, and establishing a symmetric equation based on the symmetric characteristics at both ends of the train; S5. Establishing a linear equation system according to the same cross-section equation and the symmetric equation, and solving the linear equation system to obtain the correction amount; S6. Transforming the N marker points in the train images captured by the linear array camera based on the correction amount to obtain the train images after distortion correction.

[0039] Specifically, this application uses at least two linear array cameras not in the same cross-section to synchronously capture train images. An in-cross-section equation is established based on the actual physical positions of the linear array cameras in the layout, the positions of the same cross-section of the train captured by each linear array camera, and the assumed correction amount. A constraint equation set is established based on multiple in-cross-section equations and symmetry equations, and then the assumed correction amount is solved. The solved correction amount can be used to simultaneously correct the images captured by these linear array cameras.

[0040] The improvement of this application lies in that the whole train image is captured by two or more linear array cameras not in the same cross-section, and the train images captured by the linear array cameras at multiple angles are used for self blind correction. When this method is applied in the rail transit field, it is not necessary to set scale marks on the train to be captured, nor is it necessary to have prior information such as a standard template image or the scale of train components and markers. The application premise is more relaxed, and the train image can be directly corrected, realizing efficient and stable blind correction of the distortion of the linear array camera images of the whole rail train.

[0041] Taking two linear array cameras as an example (the case of multiple linear array cameras is similar), each step will be described in detail with reference to the accompanying drawings.

[0042] Step S1 is to synchronously capture multiple train images based on at least two linear array cameras located in different cross-sections. The cross-section refers to the rail cross-section. Taking two linear array cameras as an example, before capturing the train images, the two linear array cameras are arranged on both sides of the rail and in different cross-sections, that is, the angle between the connection line of the two linear array cameras and the rail is not perpendicular, as Figure 2 shown. When the train passes by, the two linear array cameras synchronously capture the train and obtain a (whole train) train image respectively. Since the speed of the train cannot meet the requirement of being completely uniform, when the train passes non-uniformly, both linear array cameras will capture distorted train images.

[0043] In step S2, the linear array camera that first captures the train image is used as the reference camera, and other linear array cameras are used as offset cameras. The physical offset distance between each offset camera and the reference camera is measured, and based on the physical offset distance and the horizontal resolution of the linear array camera imaging, the image offset distance between each offset camera and the reference camera is solved. Taking two linear array cameras as an example, the linear array camera that first captures the train image is used as the reference camera, and the other linear array camera is used as the offset camera. The physical offset distance Offset_L (unit: meter) between the offset camera and the reference camera is measured in space. According to the physical offset distance and the horizontal resolution of the linear array camera imaging, the image offset distance Offset_x (unit: image pixel) between the train image captured by the offset camera and the train image captured by the reference camera is obtained in the same proportion.

[0044] Step S3 is to extract N marked points from the train images captured by each linear array camera. The N marked points in different train images correspond to each other on the train cross-section. Extract the x-coordinates of the marked points in the train images of the reference camera and the offset cameras, and mark the x-coordinates of the marked points in the train images of each offset camera that are within the coordinate range of the reference camera. Taking two linear array cameras as an example, the marked point positioning can be performed on the images captured by the two linear array cameras respectively through existing mature marked point positioning algorithms or manual annotation methods, and the x-coordinates of the marked points in the images can be obtained. The marked point positioning needs to ensure that the number of marked points in the train images captured by each linear array camera is the same (for example, each train image captured by a linear array camera has N marked points), and the marked points correspond one-to-one on the same cross-section of the train. Here, one-to-one correspondence means that for a marked point in one image, a marked point that is the same as it or on the same train cross-section can be found in other images. The marked points can be understood as characteristic points on the train, which can be the corner points of the train windows, the corner points of the doors, etc. After the marked point positioning is completed, record the x-coordinates of all marked points in their respective images. As Figure 3 shown, record the coordinates of the marked points in the train image of the reference camera in ascending order as , and the coordinates of the marked points in the train image of the offset camera as . Extract the x-coordinates of all marked points in the train image of the reference camera and the x-coordinates of the marked points in the train images of the offset cameras outside the above coordinate range. Obtain . Here .

[0045] Step S4 is to assume that the correction amount of the extracted x-coordinates of the marked points is an unknown quantity, establish a same cross-section equation based on the characteristics corresponding to the marked point cross-sections in the train images captured by the offset cameras and the reference camera, and establish a symmetry equation based on the symmetric characteristics at both ends of the train. Assume that the correction amount of the extracted marked point is denoted as , and use the correction amount as the unknown quantity. Here, fix the last marked point in the train image of the offset camera, and the correction amount of this marked point is 0, that is Use the image stretching / compression transformation and image offset corresponding to the correction amount to establish a same cross-section equation for the N marked points of the two linear array cameras respectively.

[0046] In a possible implementation manner, the same cross-section equation is:

[0047] (1)

[0048] Among them, is the x-coordinate of the th marked point in the train image of the reference camera 、 Correction amount of the th marker point in the reference camera train image 、 Offset_x is the image offset between the offset camera and the reference camera. is the x coordinate of the th marker point in the offset camera train image. is the correction amount of the th marker point in the offset camera train image.

[0049] In a possible implementation, for the correction amounts of the marker points in each offset camera train image that are within the coordinate range of the reference camera, assuming only linear distortion exists, it is expressed by the known x coordinates of the marker points in the reference camera train image, the correction amounts of the marker points in the reference camera train image, and the x coordinates of the marker points in the offset camera train image as:

[0050] (2)

[0051] Wherein, is the correction amount of the th marker point in the offset camera train image, and at this time the th marker point is within the coordinate range of the reference camera; and are the x coordinates of the j and j +1th marker points in the reference camera train image. and are the correction amounts of the j and j +1th marker points in the reference camera train image. is the x coordinate of the th marker point in the offset camera train image, located between and . The meaning expressed by Formula 2 is that is placed at a position in the image between and . Assuming that only linear distortion exists between and , then the correction amount of can be determined by the unknown quantity and the unknown quantity , and it is an expression of a linear transformation.

[0052] In addition, if the train has the property of head-to-tail symmetry (such as the common six-car formation of A1-B1-C1-C2-B2-A2, where A1 and A2 are the end cars with cabs, B1 and B2 are the near-end cars containing power equipment, and C1 and C2 are the intermediate cars for transporting passengers), a series of symmetry equations can also be established based on the symmetry characteristics of the train. For example, Figure 3 the symmetry between the head and the tail of the train means that the length of a specific component at the head of the train is equal to the length of the same specific component at the tail of the train.

[0053] In a possible implementation, the symmetry equation is:

[0054] (3)

[0055] Wherein, and are the x coordinates of the marked points of the specific components in the first half formation of the reference camera train image, and are and correction amounts; and are the x coordinates of the marked points of the same type of specific components in the second half formation of the offset camera train image, and are and correction amounts. The meaning expressed by this equation is that a certain specific component appears in both the first half and the second half of the train, and the length of its corrected image is equal in terms of pixel points. Here, multiple groups of similar equations can be established based on the symmetry information of the train.

[0056] Step S5 is to establish a linear equation system based on the same cross-section equation and the symmetry equation, and solve the linear equation system to obtain the correction amounts. All the same cross-section equations and symmetry equations are established as a linear equation system, and the linear equation system is solved to obtain the correction amounts assumed in step S4. The same cross-section equation and the symmetry equation are rewritten in the form of matrix rows:

[0057] (4)

[0058] Wherein, and b are constants, is the correction amount. This form of matrix rows is composed of all the formulas (1) and (3). For the correction amount of the marked points not assumed therein,

[0059] When , equation (1) can be written as:

[0060] ;

[0061] When happens, substituting formula (2) into formula (1), it can be written as:

[0062] ;

[0063] Establish all the above matrix row equalities into a system of linear equations in the form of. Here y is . Solve the system of linear equations to obtain the correction amount.

[0064] Step S6 is to transform N marker points in the train image captured by the line array camera based on the correction amount to obtain a distortion-corrected train image. Based on the obtained correction amount , perform coordinate transformation on all the marker points located in step S3 to complete image distortion correction. As Figure 4 shown, after the train images captured by the reference camera and the offset camera are distortion-corrected, the positions of the two images on the same cross-section form a standard one-to-one correspondence, and the distorted images are well restored.

[0065] It can be understood that the method for blind correction of image distortion of an entire track train line array camera provided in this application uses the train images captured by at least two line array cameras not on the same cross-section to perform blind correction on itself, without prior information or preconditions such as a standard image template or other train component scales and markers, and can more efficiently and stably achieve train image distortion correction.

[0066] Please refer to Figure 5 shown Figure 5This is the structural diagram of the blind image distortion correction system for the whole train linear array camera provided by the embodiment of the present application. The system is used to implement the above method and includes: an image acquisition unit for synchronously capturing multiple train images based on at least two linear array cameras located at different cross-sections, where the cross-section refers to the track cross-section; an offset calculation unit for taking the linear array camera that first captures the train image as the reference camera and other linear array cameras as offset cameras, measuring the physical offset distance between each offset camera and the reference camera, and solving the image offset distance between each offset camera and the reference camera according to the physical offset distance and the horizontal resolution of the linear array camera imaging; a marker point extraction unit for extracting N marker points from the train images captured by each linear array camera, where the N marker points in different train images correspond to each other on the train cross-section, extracting the x coordinates of the marker points in the train images of the reference camera and the offset cameras, and marking the x coordinates of the marker points in the train images of each offset camera that are within the coordinate range of the reference camera; an equation establishment unit for assuming the correction amount of the extracted x coordinates of the marker points as unknowns, establishing an equation of the same cross-section based on the characteristics corresponding to the marker point cross-sections in the train images captured by the offset camera and the reference camera, and establishing a symmetry equation based on the symmetric characteristics at both ends of the train; an equation solving unit for establishing a linear equation system according to the equation of the same cross-section and the symmetry equation, and solving the linear equation system to obtain the correction amount; a distortion correction unit for transforming the N marker points in the train images captured by the linear array camera based on the correction amount to obtain the train images after distortion correction.

[0067] In a possible implementation manner, in the equation establishment unit, the equation of the same cross-section is:

[0068] ;

[0069] Wherein, is the x coordinate of the th marker point in the train image of the reference camera 、 The correction amount of the th marker point in the train image of the reference camera 、 Offset_x is the image offset distance between the offset camera and the reference camera, is the x coordinate of the th marker point in the train image of the offset camera, is the th marker point in the train image of the offset camera.

[0070] In a possible implementation manner, in the solution formula establishing unit, for the correction amounts of the marked points in each offset camera train image that are within the reference camera coordinate range, assuming only linear distortion exists, it is expressed by the x coordinate of the marked point in the known reference camera train image, the correction amount of the marked point in the reference camera train image, and the x coordinate of the marked point in the offset camera train image as:

[0071] ;

[0072] wherein, is the correction amount of the th marked point in the offset camera train image, and at this time the th marked point is within the reference camera coordinate range; and are the x coordinates of the j and j +1th marked points in the reference camera train image, and are the correction amounts of the j and j +1th marked points in the reference camera train image, is the x coordinate of the th marked point in the offset camera train image, and is between and .

[0073] In a possible implementation manner, in the solution formula establishing unit, the symmetric equation is:

[0074] ;

[0075] wherein, and are the x coordinates of the marked points of the specific components in the first half formation of the reference camera train image, and are the correction amounts of and ; and are the x coordinates of the marked points of the same type of specific components in the second half formation of the offset camera train image, and are the correction amounts of and .

[0076] It should be noted that the on-vehicle linear array camera image distortion blind correction system for rail trains provided by this application is used to implement the above-mentioned on-vehicle linear array camera image distortion blind correction method for rail trains, corresponds one-to-one with the above method, and has corresponding technical effects, so it will not be elaborated here.

[0077] The embodiments of the present application further provide a computer storage medium, including computer instructions, which, when running on a mobile terminal, cause the mobile terminal to execute the method as described above.

[0078] The embodiments of the present application further provide a computer program product, which, when running on a mobile terminal, causes the mobile terminal to execute the method as described above.

[0079] The specific embodiments described above further elaborate on the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for blind correction of image distortion of a linear array camera of a railway train, characterized in that: include: A plurality of train images are obtained by synchronously photographing at least two linear array cameras located at different sections, wherein the section refers to a track cross section; The linear array camera that first captures the train image is used as the reference camera, and the other linear array cameras are used as offset cameras. The physical offset distance between each offset camera and the reference camera is measured, and the image offset distance between each offset camera and the reference camera is solved according to the physical offset distance and the horizontal resolution of the linear array camera imaging. Extract N marking points from the train images taken by each linear array camera, the N marking points in different train images correspond to each other on the train cross section, extract the x coordinates of the marking points in the train images of the reference camera and the offset camera, and mark the x coordinates of the marking points in the train images of each offset camera within the coordinate range of the reference camera; Assuming that the correction amount of the x-coordinate of the extracted marker point is an unknown quantity, the same section equation is established with the features corresponding to the sections of the marker points in the train images taken by the offset camera and the reference camera, and the symmetry equation is established with the features symmetrical at both ends of the train; Establish a linear equation system according to the same-section equation and the symmetry equation, and solve the linear equation system to obtain a correction amount; Transforming N marking points in the train image captured by the linear array camera based on the correction amount to obtain a distortion-corrected train image; Among them, the same cross-section equation is: ； in, is the first The x-coordinate of the marker point 、 The first Correction amount for each marker point 、 Offset_x is the image offset between the offset camera and the reference camera. is the offset camera train image The x-coordinates of the marker points, is the offset camera train image The correction amount for each marker point; The symmetric equation is: ； in, and is the x-coordinate of the marker point of the specific component of the first half of the train in the reference camera image, and for and The amount of correction; and is the x-coordinate of the marker point of the same type of specific component in the second half of the group in the offset camera train image, and for and The amount of correction.

2. According to claim 1, a method for blind correction of image distortion of a linear array camera of a railway train, characterized in that: Assuming that only linear distortion exists, the correction amount of the marked points in the reference camera train image, the correction amount of the marked points in the reference camera train image, and the x-coordinate of the marked points in the offset camera train image are expressed as: ； in, is the offset camera train image The correction amount of the marking point, The marker points are within the coordinate range of the reference camera; and is the first j and j +1 x coordinate of the marker point, and is the first j and j +1 mark point correction, is the offset camera train image The x-coordinates of the marker points are located at and between.

3. A blind correction system for image distortion of a linear array camera of a railway train, characterized in that: A method for blindly correcting image distortion of a linear array camera of a whole railway train according to any one of claims 1 to 2, comprising: An image acquisition unit, configured to obtain a plurality of train images based on synchronous shooting by at least two linear array cameras located at different sections, wherein the section refers to a track cross section; An offset obtaining unit is used to use the linear array camera that first captures the train image as a reference camera and other linear array cameras as offset cameras, measure the physical offset distance between each offset camera and the reference camera, and calculate the image offset distance between each offset camera and the reference camera according to the physical offset distance and the horizontal resolution of the linear array camera imaging; A marker point extraction unit is used to extract N marker points from the train images taken by each linear array camera, wherein the N marker points in different train images correspond to each other on the train cross section, extract the x coordinates of the marker points in the train images of the reference camera and the offset camera, and mark the x coordinates of the marker points in the train images of each offset camera that are within the coordinate range of the reference camera; A solution formula establishment unit is used to assume that the correction amount of the extracted marker point x-coordinate is an unknown amount, establish a same-section equation with features corresponding to the marker point cross section in the train image taken by the offset camera and the reference camera, and establish a symmetric equation with features symmetrical at both ends of the train; A solution-solving unit, used for establishing a linear equation system according to the iso-section equation and the symmetry equation, and solving the linear equation system to obtain a correction amount; A distortion correction unit, configured to transform N marking points in the train image taken by the linear array camera based on the correction amount to obtain a train image after distortion correction; Among them, in the solution establishment unit, the same section equation is: ； in, is the first The x-coordinate of the marker point 、 The first Correction amount for each marker point 、 Offset_x is the image offset between the offset camera and the reference camera. is the offset camera train image The x-coordinates of the marker points, is the offset camera train image The correction amount for each marker point; In the solution building unit, the symmetric equation is: ； in, and is the x-coordinate of the marker point of the specific component of the first half of the train in the reference camera image, and for and The amount of correction; and is the x-coordinate of the marker point of the same type of specific component in the second half of the group in the offset camera train image, and for and The amount of correction.

4. The blind correction system for image distortion of a linear array camera of a railway train according to claim 3, characterized in that: In the solution formula establishment unit, assuming that only linear distortion exists, the correction amount of the marked point in the reference camera train image, the correction amount of the marked point in the reference camera train image, and the x-coordinate of the marked point in the offset camera train image are known, and the correction amount of the marked point in each offset camera train image within the reference camera coordinate range is expressed as: ； in, is the offset camera train image The correction amount of the marking point, The marker points are within the coordinate range of the reference camera; and is the first j and j +1 x coordinate of the marker point, and is the first j and j +1 mark point correction, is the offset camera train image The x-coordinates of the marker points are located at and between.

5. A computer storage medium, characterized in that: The method comprises computer instructions, and when the computer instructions are executed on a mobile terminal, the mobile terminal executes a method for blindly correcting image distortion of a linear array camera of a whole railway train as claimed in any one of claims 1 to 2.

6. A computer program product, characterized in that When the computer program product runs on a mobile terminal, the mobile terminal executes a method for blindly correcting image distortion of a linear array camera of a railway train as described in any one of claims 1 to 2.

Images