Orbit region extraction method and device under orthophoto condition

Abstract

The application discloses a method and device for extracting a track region under normal view conditions, which comprises the following steps: acquiring a track infrared image under normal view conditions, wherein the track infrared image is collected by an infrared sensor arranged at the front end of a track train; dividing the track infrared image into a plurality of sub-images along a first direction, wherein the first direction is the extension direction of a track region; in each sub-image, a plurality of pixels in the first direction are accumulated to obtain a plurality of accumulated pixels arranged along a second direction, wherein the second direction is perpendicular to the first direction; determining a track boundary point in each sub-image according to the plurality of accumulated pixels arranged along the second direction; and fitting the track boundary points in the plurality of sub-images to obtain the boundary of the track region in the track infrared image. The application can accurately extract the boundary of the track region in the track infrared image under normal view conditions.

Description

Technical Field

[0001] This invention relates to the field of rail transit technology, and more specifically, to a method and apparatus for extracting track areas under frontal viewing conditions. Background Technology

[0002] With the rapid development of rail transit technology, rail transit has become widespread throughout the country and is gradually becoming the preferred mode of transportation for most people. Rail transit is a mode of transportation with extremely high safety requirements; therefore, the detection of abnormal obstacles within the track area is particularly important during autonomous driving. Track region extraction not only reduces interference factors outside the track area in the image during subsequent obstacle detection but also improves the real-time performance of obstacle detection. Therefore, track region extraction is an essential step to ensure the accuracy of obstacle detection within the track area.

[0003] Most existing track feature extraction techniques are for track region extraction under top-down views. The main method involves extracting track regions based on the different grayscale distributions on the track surface. Track regions in the image have high grayscale values, while non-track regions have lower grayscale values. Dramatic changes in grayscale values ​​can occur in the track regions, which are then used to identify the track areas in the image. However, this method relies on the assumption that the track width remains constant under top-down views, and is not applicable to frontal views of the track taken by sensors positioned at the front of the vehicle. Summary of the Invention

[0004] The summary section introduces a series of simplified concepts, which will be further explained in detail in the detailed description section. The summary section of this invention is not intended to limit the key features and essential technical features of the claimed technical solution, nor is it intended to determine the scope of protection of the claimed technical solution.

[0005] To address the shortcomings of existing technologies, the first aspect of this invention proposes a method for extracting orbital regions under frontal viewing conditions, comprising:

[0006] Acquire an infrared image of the track under direct viewing conditions, wherein the infrared image of the track is acquired by an infrared sensor installed at the front end of the track train;

[0007] The orbital infrared image is divided into multiple sub-images along a first direction, where the first direction is the extension direction of the orbital region;

[0008] In each of the sub-images, multiple pixels in the first direction are accumulated to obtain multiple accumulated pixels arranged along a second direction, which is perpendicular to the first direction;

[0009] Based on the multiple accumulated pixels arranged along the second direction, determine the orbital boundary points in each of the sub-images;

[0010] The boundary points of the orbit in multiple sub-images are fitted to obtain the boundary of the orbit region in the orbit infrared image.

[0011] In some embodiments, determining the orbital boundary points in each of the sub-images based on the plurality of accumulated pixels arranged along the second direction includes:

[0012] The search is performed in the plurality of accumulated pixels using a sliding window, and the pixel sum of at least two accumulated pixels in each sliding window is calculated.

[0013] Determine the target sliding window corresponding to the extreme value of the pixel sum, and determine the track boundary point based on the position of the target sliding window.

[0014] In some embodiments, the method further includes:

[0015] The width of the track region in the current sub-image is determined based on the distance between adjacent track boundary points in the current sub-image;

[0016] The width of the track region in the current sub-image is used to determine the size of the sliding window of the next sub-image above the current sub-image.

[0017] In some embodiments, the extreme values ​​are either maximum or minimum values.

[0018] In some embodiments, dividing the orbital infrared image into multiple sub-images along a first direction includes:

[0019] Valid images are extracted from the orbital infrared image, and the valid images are divided into multiple sub-images along the first direction. The width of the valid images is equal to the width of the orbital infrared image, and the height of the valid images accounts for a preset proportion of the height of the orbital infrared image.

[0020] In some embodiments, before dividing the orbital infrared image into multiple sub-images along a first direction, the method further includes:

[0021] The orbital infrared image is preprocessed to remove noise from it, and the preprocessing includes wavelet denoising.

[0022] In some embodiments, the wavelet denoising process includes:

[0023] Wavelet decomposition was performed on the orbital infrared image to obtain the decomposition coefficients;

[0024] The decomposition coefficients are subjected to thresholding to obtain new decomposition coefficients. The thresholding process uses a local threshold, which is inversely proportional to the decomposition scale of the wavelet decomposition.

[0025] The new decomposition coefficients are used to reconstruct the signal to obtain a denoised infrared image.

[0026] In some embodiments, the threshold function for threshold processing is an adaptive threshold function.

[0027] A second aspect of the present invention provides an apparatus for extracting a track region under frontal viewing conditions. The apparatus includes a memory and a processor. The memory stores a computer program that is executed by the processor. When the computer program is executed by the processor, it performs the track region extraction method under frontal viewing conditions as described above.

[0028] A third aspect of the present invention provides a computer storage medium having a computer program stored thereon, wherein the computer program, when executed, implements the orbital region extraction method under frontal viewing conditions as described above.

[0029] The method and apparatus for extracting the track region under frontal viewing conditions of the present invention divides the track infrared image along the track extension direction into multiple sub-images, identifies track boundary points in each sub-image, and fits the track boundary points in multiple sub-images to obtain the complete track region boundary, which can accurately extract the track region with gradually changing width in the track infrared image under frontal viewing conditions. Attached Figure Description

[0030] The above and other objects, features, and advantages of the present invention will become more apparent from the more detailed description of embodiments thereof in conjunction with the accompanying drawings. The drawings are provided to further illustrate embodiments of the invention and form part of the specification. They are used together with the embodiments to explain the invention and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same parts or steps.

[0031] Figure 1 This is a schematic flowchart of a method for extracting a track region under frontal viewing conditions according to an embodiment of the present invention;

[0032] Figure 2 The above-ground view, visible light front view, and infrared light front view are provided according to an embodiment of the present invention.

[0033] Figure 3 This is a schematic diagram of track region extraction according to an embodiment of the present invention;

[0034] Figure 4 This is a schematic flowchart of a method for extracting a track region under frontal viewing conditions according to an embodiment of the present invention;

[0035] Figure 5 This is a schematic block diagram of a track region extraction device under frontal viewing conditions according to an embodiment of the present invention. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of this application more apparent, exemplary embodiments according to this application will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein. Based on the embodiments of this application described herein, all other embodiments obtained by those skilled in the art without inventive effort should fall within the protection scope of this application.

[0037] The following description provides numerous specific details to offer a more thorough understanding of this application. However, it will be apparent to those skilled in the art that this application can be practiced without one or more of these details. In other instances, certain technical features well-known in the art have not been described to avoid confusion with this application.

[0038] It should be understood that this application can be implemented in various forms and should not be construed as being limited to the embodiments set forth herein. Rather, providing these embodiments will make the disclosure thorough and complete, and will fully convey the scope of this application to those skilled in the art.

[0039] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. When used herein, the singular forms “a,” “an,” and “the” are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the terms “comprising” and / or “including,” when used in this specification, identify the presence of the stated features, integers, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups. When used herein, the term “and / or” includes any and all combinations of the associated listed items.

[0040] To fully understand this application, a detailed structure will be presented in the following description to illustrate the technical solution proposed in this application. Optional embodiments of this application are described in detail below; however, in addition to these detailed descriptions, this application may have other implementation methods.

[0041] The method, apparatus, and computer storage medium for extracting orbital regions under frontal viewing conditions proposed in embodiments of the present invention will now be described with reference to the accompanying drawings. First, see... Figure 1 , Figure 1A schematic flowchart of a method 100 for extracting a track region under frontal viewing conditions according to an embodiment of the present invention is shown. Figure 1 As shown, the orbital region extraction method 100 under frontal viewing conditions according to an embodiment of the present invention includes the following steps:

[0042] In step S110, an infrared image of the track under frontal viewing conditions is acquired, wherein the infrared image of the track is collected by an infrared sensor installed at the front end of the track train;

[0043] In step S120, the orbital infrared image is divided into multiple sub-images along a first direction, where the first direction is the extension direction of the orbital region;

[0044] In step S130, in each of the sub-images, multiple pixels in the first direction are accumulated to obtain multiple accumulated pixels arranged along the second direction, which is perpendicular to the first direction;

[0045] In step S140, the orbital boundary points in each of the sub-images are determined based on the plurality of accumulated pixels arranged along the second direction;

[0046] In step S150, the orbital boundary points in multiple sub-images are fitted to obtain the boundary of the orbital region in the orbital infrared image.

[0047] The track region extraction method 100 under frontal viewing conditions of this invention is used to extract the track region from an infrared image of a track acquired under frontal viewing conditions. In autonomous driving, infrared sensors are usually installed at a certain position on the vehicle body, and the acquired images are usually from a frontal view. However, most existing track extraction methods are based on the assumption that the track width is constant, which is not suitable for situations where the track region gradually narrows under frontal viewing conditions. Therefore, this invention introduces the idea of ​​differentiation into track region extraction, dividing the infrared image of the track into multiple sub-images along the track extension direction. The track width remains approximately constant within any sub-image. Then, the column pixel projection method is used to extract track boundary points in each sub-image. Finally, the track boundary points in all sub-images are fitted to complete the extraction of the track region within the entire infrared image of the track.

[0048] In step S110, an infrared image of the track under direct view is acquired by an infrared sensor located at the front of the train. The track area in the infrared image extends along a first direction, which is the height direction of the infrared image. Acquiring the infrared image of the track by the infrared sensor enables the safe operation of the vehicle in all weather conditions.

[0049] Compared to visible light images, infrared images have the advantage of being unaffected by illumination. However, compared to visible light images, infrared images suffer from problems such as lower signal-to-noise ratio, lower resolution, and poorer visual effects. To improve the accuracy of orbital region extraction, this embodiment of the invention first preprocesses the original infrared image after acquisition to improve image quality.

[0050] In some embodiments, preprocessing includes wavelet denoising. Wavelet denoising, based on the characteristic that signal and noise have different properties at different scales, treats the orbital infrared image as a two-dimensional signal and mainly includes three steps: wavelet decomposition, thresholding, and image reconstruction. Specifically, firstly, wavelet decomposition is performed on the original orbital infrared image f(t) to obtain decomposition coefficients ω; then, thresholding is performed on the decomposition coefficients to obtain new decomposition coefficients. Finally, the new decomposition coefficients were... Signal reconstruction is performed to obtain a denoised orbital infrared image.

[0051] Traditional wavelet denoising typically uses a global threshold in the threshold selection stage, resulting in relatively poor denoising performance. To address this issue, this invention employs a local threshold for thresholding, where the local threshold is inversely proportional to the decomposition scale of the wavelet decomposition. For example, the local threshold δ is:

[0052]

[0053] Where σ is the noise standard deviation, n is the sampled signal length, and N is the decomposition scale. This embodiment of the invention employs a local threshold in the threshold processing stage, and the local threshold is inversely proportional to the wavelet decomposition scale. This overcomes the shortcoming of the global threshold not changing with the wavelet decomposition scale. It incorporates the element of the decomposition scale into the general threshold, and the threshold decreases accordingly as the decomposition scale increases, conforming to the change law of signal-to-noise ratio with increasing decomposition scale during wavelet decomposition.

[0054] Besides the threshold setting, the selection of the threshold function is crucial to the accuracy of the reconstructed image and has a significant impact on the overall wavelet denoising process. Traditional thresholding formulas suffer from edge loss when processing infrared images, while edge information in infrared images is critical for orbit region extraction. Common thresholding functions include soft and hard thresholding functions. While soft thresholding functions are relatively smooth, they can easily cause blurring and distortion in the image. Considering that infrared images have relatively poor clarity compared to visible light, and that subsequent orbit region extraction requires high image clarity, this embodiment of the invention employs a hard thresholding function that better preserves image edge information, and makes corresponding improvements to mitigate issues such as ringing effects.

[0055] Specifically, this embodiment of the invention achieves the effect of protecting image edge information by introducing a matching shrinkage factor into the traditional thresholding formula for threshold correction. While hard thresholding functions can effectively preserve edge information, they suffer from problems such as ringing effects. This embodiment of the invention employs an adaptive thresholding function to improve image clarity while protecting the edge information of infrared orbit images.

[0056] For example, the adaptive semi-hard threshold function used in the embodiments of the present invention is:

[0057]

[0058] Where ω is the original decomposition coefficient, Here, β represents the decomposition coefficients after thresholding, δ is the threshold, and sgn(x) is the sign function. In the above adaptive semi-hard thresholding function, the parameter β is adaptive as α changes, and can be adjusted accordingly to change the decomposition coefficients. The output of this adaptive semi-hard thresholding function is relatively smooth and continuous, which can improve the ringing effect of the hard thresholding function, better preserve the edge information of the image, and maintain high sharpness.

[0059] After completing the thresholding process and obtaining the processed decomposition coefficients Then, the decomposition coefficients after thresholding. Perform inverse wavelet transform to obtain the denoised infrared image. After the wavelet denoising process described above, a clear orbital infrared image with a high signal-to-noise ratio is obtained, which allows for high accuracy in orbital region extraction.

[0060] Traditional track region extraction typically extracts the entire frame of an image. However, in frontal images, the width of the track region gradually decreases along the track's extension direction, resulting in lower accuracy when extracting the track region from the background. Therefore, this embodiment of the invention selects a portion of the effective image I from each frame of infrared images. a Extract the orbital region. Valid image I a The width is the same as the original infrared image, and the height is less than the original infrared image. For example, the lower 40% of the original infrared image can be selected as the effective image, and the orbital region can be extracted from it.

[0061] Previous methods for extracting the track region were mostly performed under top-down conditions, assuming a constant track width. This is not suitable for situations where the track surface gradually narrows under frontal conditions. Therefore, this invention proposes a track region extraction method based on sliding window column pixel projection. This method introduces the idea of ​​differentiation, dividing the track infrared image into multiple sub-images. Since the track region boundary changes little in each sub-image, the track region boundary in each sub-image can be approximated as a point. After determining the track boundary points in multiple sub-images, the track boundary points in multiple sub-images are fitted to obtain the complete track region boundary.

[0062] The orbital infrared image is divided into multiple sub-images along a first direction, which is the extension direction of the orbital region, the direction of the pixel columns in the orbital infrared image, and the height direction of the orbital infrared image. The width of the sub-images is the same as that of the orbital infrared image, and the height is related to the number of divisions. For example, the different sub-images have the same height.

[0063] For example, a sub-image can be extracted from the valid image using a vertical sliding window. Figure 3 As shown, the vertical sliding window starts from the first row of pixels at the bottom of the valid image and slides upwards sequentially. The sliding step size is the same as the height of the vertical sliding window, i.e., l i = i·l+1, i = 0, 1, 2…n, where l i This represents the pixel values ​​of the first row within the (i+1)th vertical sliding window, where l is the height of the vertical sliding window. i The range of values ​​for is l i ≤row, the size of any vertically sliding window is l×col, where row represents the row width and col represents the column width.

[0064] For each sub-image, this embodiment of the invention assumes that the track boundaries on both sides of a single track within the sub-image are parallel, meaning the width of the track region within the sub-image remains constant. Based on this assumption, this embodiment of the invention simplifies each track boundary line within each sub-image to a single track boundary point for extraction, thereby reducing computational complexity. The track region identified in this embodiment of the invention can be the region of one track; in this case, the track boundary points identified within each sub-image are the left and right boundary points of the track region, and the final boundary of the track region is the left and right boundary points of the track region. Alternatively, the track region identified in this embodiment of the invention can be the boundary of two tracks; in this case, the track boundary points identified within each sub-image are the four boundary points of the two track regions, and the final boundary of the track region is the four boundary points of the two track regions.

[0065] Specifically, in each sub-image, multiple pixels along the first direction are accumulated to obtain multiple accumulated pixels arranged along the second direction. The second direction is perpendicular to the first direction, i.e., the direction of the pixel rows in the infrared orbit image, or the width direction of the infrared orbit image. The accumulated pixels s can be represented as: Multiple accumulated pixels are arranged in a one-dimensional horizontal direction to represent the inherent horizontal grayscale distribution of the sub-image.

[0066] Because the image grayscale level changes abruptly in the track region, track boundary points among multiple accumulated pixels can be determined based on the grayscale changes of multiple accumulated pixels. Depending on the track material, the grayscale level of the track region may be higher or lower than that of the background region. Therefore, track boundary points can be determined based on high-grayscale or low-grayscale areas among multiple accumulated pixels. The following description primarily uses the example of the track region having a higher grayscale level than the background region.

[0067] To search for high-grayscale or low-grayscale regions among multiple accumulated pixels, a sliding window is used to search the multiple accumulated pixels, and at least one region is calculated in each sliding window.

[0068] The pixel sum of two accumulated pixels is used; the target sliding window corresponding to the extreme value of the pixel sum is determined, and the orbital boundary points are determined based on the position of the target sliding window. The sliding window can traverse multiple accumulated pixels from beginning to end, and the sliding step size can be one pixel, but is not limited to this. When the sliding window...

[0069] When the sliding step of the aperture is one pixel, the pixel and y(a) can be expressed as:

[0070]

[0071] By sliding the sliding window across multiple accumulated pixels, we can obtain the set of pixel sums Y = {y(1), y(2), ..., y(col-w(i)-1)}, where w(i) represents the width of the sliding window within the i-th sub-image.

[0072] When searching a target sliding window, the extreme values ​​of the pixel sum can be either maximum or minimum. Taking the maximum as an example, if the width of the sliding window is the same as the width of the track region, the pixel sum reaches its maximum when the sliding window overlaps with the track region. In this case, the left and right boundaries of the sliding window can be used as track boundary points. Assuming the sliding window slides from left to right, if the width of the sliding window is greater than the width of the track region, the pixel sum gradually increases when the right boundary of the sliding window enters the track region, reaching its maximum when the sliding window completely covers the track region. As the sliding window continues to slide, if the track region remains within the sliding window's coverage area, the pixel sum remains at its maximum. When the right boundary of the sliding window leaves the track region, the pixel sum gradually decreases. Therefore, two track boundary points can be determined based on the positions of the sliding window when the pixel sum reaches and ends its extreme value. Based on a similar principle, the sliding window can be slid from left to right once to find the pixel and the position of the sliding window when the extreme value is reached, thus determining the right boundary point of the track region; then the sliding window can be slid from right to left once to find the pixel and the position of the sliding window when the extreme value is reached, thus determining the left boundary point of the track region, thereby determining the two track boundary points.

[0073] If the width of the sliding window is less than the width of the track region, the pixel sum reaches its maximum value when the sliding window is fully within the track region. As the sliding window continues to slide, the pixel sum remains at its maximum value if the sliding window remains within the track region. When the right boundary of the sliding window leaves the track region, the pixel sum gradually decreases. Therefore, the two track boundary points can also be determined based on the positions of the sliding window when the pixel sum reaches and ends its extreme value, or by having the sliding window slide twice in two different directions to find the two track boundary points.

[0074] Tests showed that the search effect was better when the width of the sliding window was approximately the same as the width of the track region. However, under normal viewing conditions, the width of the track region was not the same in different sub-images, and the width of the track region in each sub-image decreased from bottom to top. For the first sub-image i1 at the bottom, the preset track width (railwidth) can be used as the width of the sliding window, then w(1) = railwidth. However, this track width is only used in the first sub-image. After determining the track boundary points in the sub-image using column pixel projection, the track width of the first sub-image is determined according to the track boundary points. In the second sub-image, the track width of the first sub-image is used as the width of the sliding window. In the third region, the track width of the second sub-image is used as the width of the sliding window, and so on, until the last sub-image.

[0075] After applying the column pixel projection algorithm described above, for each track boundary, a track boundary point is obtained within each sub-image. The track boundary point within each sub-image can be represented as: (x1, y1), (x2, y2), ..., (x...). n ,y n Then, the orbital boundary points of multiple sub-images are fitted to obtain the boundary of the overall orbital region.

[0076] As is known in practice, the boundary of the orbital region is usually a straight line. Therefore, the boundary of the orbital region can be obtained by fitting the curve using the least squares method. Let the fitted curve be y = kx + b, and the error function can be expressed as:

[0077]

[0078] Solve for parameters k and b such that e 2 At its minimum, the curve at this point represents the boundary line of the orbital region.

[0079] In summary, the orbit region extraction method 100 under frontal viewing conditions of this invention divides the orbit infrared image into multiple sub-images along the orbit extension direction, identifies orbit boundary points in each sub-image, and fits the orbit boundary points in multiple sub-images to obtain complete orbit region boundaries, thus accurately extracting the orbit region with gradually changing width from the orbit infrared image under frontal viewing conditions.

[0080] This invention also provides a device for extracting track regions under frontal viewing conditions, which can be used to implement the track region extraction method 100 under frontal viewing conditions described above. See also Figure 5 , Figure 5 A schematic block diagram of a track area extraction device 500 under frontal viewing conditions according to an embodiment of the present invention is shown. The track area extraction device 500 under frontal viewing conditions according to the embodiment of the present invention can be implemented as a controller of a rail vehicle, a cloud processor, or any electronic device.

[0081] like Figure 5 As shown, the track region extraction device 500 under frontal viewing conditions includes a memory 510, a processor 520, and a computer program stored in the memory 510 and running on the processor 520. When the processor 520 executes the computer program, it can implement the track region extraction method 100 under frontal viewing conditions as described above.

[0082] The memory 510 is a memory for storing processor-executable instructions, such as processor-executable program instructions for implementing corresponding steps in the orbital region extraction method 100 under frontal viewing conditions according to an embodiment of the present invention. The memory 510 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include random access memory (RAM) and / or cache memory. The non-volatile memory may, for example, include read-only memory (ROM), hard disk, flash memory, etc.

[0083] The processor 520 can execute the program instructions stored in the memory 510 to implement the functions (implemented by the processor) in the embodiments of the present invention described herein, and / or other desired functions, such as performing corresponding steps of the orbital region extraction method 100 under frontal viewing conditions according to an embodiment of the present invention. Various application programs and various data, such as various data used and / or generated by the application programs, can also be stored in the computer-readable storage medium.

[0084] Processor 520 may be a central processing unit (CPU), graphics processing unit (GPU), application-specific integrated circuit (ASIC), field-programmable gate array (FPGA), or other form of processing unit with data processing capabilities and / or instruction execution capabilities, and may control other components in the orbital region extraction method apparatus 500 under frontal conditions to perform desired functions. Processor 520 is capable of executing the instructions stored in memory 510 to perform the path planning method described herein. For example, processor 520 may include one or more embedded processors, processor cores, microprocessors, logic circuits, hardware finite state machines (FSMs), digital signal processors (DSPs), or combinations thereof.

[0085] This invention also proposes a computer storage medium storing a computer program that, when executed, implements the orbital region extraction method 100 under frontal viewing conditions according to this invention. The computer storage medium may include, for example, a memory card of a smartphone, a storage component of a tablet computer, a hard disk of a personal computer, a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a portable compact disc read-only memory (CD-ROM), a USB memory, or any combination of the above storage media. The computer-readable storage medium may be any combination of one or more computer-readable storage media.

[0086] The orbit region extraction device and computer storage medium under frontal viewing conditions of this invention are used to implement the orbit region extraction method 100 under frontal viewing conditions described above, and therefore also have similar advantages.

[0087] Although exemplary embodiments have been described herein with reference to the accompanying drawings, it should be understood that the above exemplary embodiments are merely illustrative and are not intended to limit the scope of this application. Various changes and modifications can be made therein by those skilled in the art without departing from the scope and spirit of this application. All such changes and modifications are intended to be included within the scope of this application as claimed in the appended claims.

[0088] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in 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 beyond the scope of this application.

[0089] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed.

[0090] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of this application may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0091] Similarly, it should be understood that, in order to streamline this application and aid in understanding one or more of the various inventive aspects, features of this application may sometimes be grouped together in a single embodiment, figure, or description thereof in the description of exemplary embodiments of this application. However, this approach should not be construed as reflecting an intention that the claimed application requires more features than are expressly recited in each claim. Rather, as reflected in the corresponding claims, its inventive point lies in solving the corresponding technical problem with features fewer than all features of a single disclosed embodiment. Therefore, the claims following the detailed description are hereby expressly incorporated into that detailed description, wherein each claim itself is a separate embodiment of this application.

[0092] Those skilled in the art will understand that, apart from the mutual exclusion of features, all features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all processes or units of any method or apparatus so disclosed can be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) may be replaced by an alternative feature that serves the same, equivalent, or similar purpose.

[0093] Furthermore, those skilled in the art will understand that although some embodiments described herein include certain features but not others included in other embodiments, combinations of features from different embodiments are intended to be within the scope of this application and form different embodiments. For example, in the claims, any one of the claimed embodiments can be used in any combination.

[0094] The various component embodiments of this application can be implemented in hardware, or as software modules running on one or more processors, or a combination thereof. Those skilled in the art will understand that microprocessors or digital signal processors (DSPs) can be used in practice to implement some or all of the functions of some modules according to the embodiments of this application. This application can also be implemented as an apparatus program (e.g., a computer program and computer program product) for performing part or all of the methods described herein. Such an implementation of this application can be stored on a computer-readable medium, or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.

[0095] It should be noted that the above embodiments are illustrative of this application and not limiting of it, and that those skilled in the art can devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses should not be construed as limiting the claims. This application can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third, etc., does not indicate any order. These words can be interpreted as names.

[0096] The above description is merely a specific embodiment or illustration of the embodiments of this application. The scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. The scope of protection of this application shall be determined by the scope of the claims.

Claims

1. A method for extracting orbital regions under frontal viewing conditions, characterized in that, The method includes: Acquire an infrared image of the track under direct viewing conditions, wherein the infrared image of the track is acquired by an infrared sensor installed at the front end of the track train; The orbital infrared image is preprocessed to remove noise from the orbital infrared image, and the preprocessing includes wavelet denoising processing; The wavelet denoising process includes: performing wavelet decomposition on the orbital infrared image to obtain decomposition coefficients; performing thresholding on the decomposition coefficients to obtain new decomposition coefficients; the thresholding process uses a local threshold, which is inversely proportional to the decomposition scale of the wavelet decomposition; and reconstructing the signal from the new decomposition coefficients to obtain the denoised infrared image. The orbital infrared image is divided into multiple sub-images along a first direction, where the first direction is the extension direction of the orbital region; In each of the sub-images, multiple pixels in the first direction are accumulated to obtain multiple accumulated pixels arranged along a second direction, which is perpendicular to the first direction; Based on the multiple accumulated pixels arranged along the second direction, determine the orbital boundary points in each of the sub-images; The boundary points of the orbit in multiple sub-images are fitted to obtain the boundary of the orbit region in the orbit infrared image.

2. The method for extracting the orbital region under frontal viewing conditions according to claim 1, characterized in that, Determining the orbital boundary points in each of the sub-images based on the plurality of accumulated pixels arranged along the second direction includes: The search is performed in the plurality of accumulated pixels using a sliding window, and the pixel sum of at least two accumulated pixels in each sliding window is calculated. Determine the target sliding window corresponding to the extreme value of the pixel sum, and determine the track boundary point based on the position of the target sliding window.

3. The method for extracting the orbital region under frontal viewing conditions according to claim 2, characterized in that, Also includes: The width of the track region in the current sub-image is determined based on the distance between adjacent track boundary points in the current sub-image; The width of the track region in the current sub-image is used to determine the size of the sliding window of the next sub-image above the current sub-image.

4. The method for extracting the orbital region under frontal viewing conditions according to claim 2 or 3, characterized in that, The extreme values ​​are either maximum or minimum values.

5. The method for extracting the orbital region under frontal viewing conditions according to claim 1, characterized in that, The step of dividing the orbital infrared image into multiple sub-images along a first direction includes: Valid images are extracted from the orbital infrared image, and the valid images are divided into multiple sub-images along the first direction. The width of the valid images is equal to the width of the orbital infrared image, and the height of the valid images accounts for a preset proportion of the height of the orbital infrared image.

6. The method for extracting the orbital region under frontal viewing conditions according to claim 1, characterized in that, The threshold function for threshold processing is an adaptive threshold function.

7. A device for extracting a track region under frontal viewing conditions, characterized in that, The device includes a memory and a processor, the memory storing a computer program executed by the processor, the computer program performing the orbital region extraction method under frontal viewing conditions as described in any one of claims 1-6 when executed by the processor.

8. A computer storage medium having a computer program stored thereon, characterized in that, When the computer program is executed, it implements the orbital region extraction method under the frontal viewing condition as described in any one of claims 1-6.

Images