Ultra-wide-angle security distance measuring lens distortion correction method and system

Abstract

The invention relates to the technical field of lens distortion correction, in particular to an ultra-wide-angle security and protection distance measurement lens distortion correction method and system. A distortion correction strategy dynamic configuration module in the system obtains an acquisition picture of an ultra-wide-angle security and protection distance measurement lens aiming at a monitoring area; according to the distance between each pixel point in the collected picture and the center point of the corresponding collected picture, the distortion correction area of the collected picture is dynamically divided, and a corresponding distortion correction strategy is configured for each divided distortion correction area. According to the method, the image collected by the ultra-wide-angle security and protection distance measuring lens can be dynamically divided into a plurality of different distortion correction areas, multi-area differential correction is realized, the distance measuring precision and view integrity of the image collected by the ultra-wide-angle security and protection distance measuring lens are considered, and the problem of dislocation of a correction target and security and protection requirements is solved.

Description

Technical Field

[0001] This invention relates to the field of lens distortion correction technology, specifically to a method and system for correcting distortion in an ultra-wide-angle security rangefinder lens. Background Technology

[0002] Ultra-wide-angle lenses (FoV > 180°) can significantly improve coverage in security monitoring, but the introduced barrel distortion causes severe bending at the image edges. Traditional distortion correction methods use globally uniform parameter models (such as the Brown-Conrady model), which can restore visual straight lines, but still have significant shortcomings. These shortcomings mainly manifest in the problem of balancing ranging accuracy and field of view integrity (strong correction leads to edge image cropping, with a field of view loss rate usually exceeding 30%, while weak correction results in a large ranging error in the central area, usually not less than 3%). Furthermore, there is a mismatch between the correction target and security requirements (optimizing the target focuses on restoring the visual straight line, emphasizing minimizing reprojection error, rather than the distance measurement accuracy required for actual business). Therefore, the industry urgently needs an ultra-wide-angle lens distortion correction method for security ranging that balances ranging accuracy and field of view integrity. Summary of the Invention

[0003] The purpose of this invention is to provide a distortion correction method and system for an ultra-wide-angle security rangefinder lens to solve the problems mentioned in the background art.

[0004] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a distortion correction method for an ultra-wide-angle security rangefinder lens, the method comprising the following steps:

[0005] S1. Obtain the monitoring area corresponding to the ultra-wide-angle security ranging lens, and use the mapping error between pixel coordinates and physical distance within the monitoring area as the optimization target to construct a target optimization model.

[0006] S2. Acquire the image captured by the ultra-wide-angle security rangefinder for the monitoring area. Based on the distance between each pixel in the image and the center point of the corresponding image, dynamically divide the distortion correction area of ​​the image and configure the corresponding distortion correction strategy for each divided distortion correction area.

[0007] S3. Based on the defined distortion correction regions and the corresponding distortion correction strategies, perform distortion correction on the captured image to obtain the corrected image of the captured image.

[0008] S4. Extract semantic information from the corrected image of the acquired image to generate a set of semantic information for the corresponding acquired image; perform semantic feature segmentation on the corrected image of the acquired image based on the set of semantic information to obtain each segmented semantic feature region; and update each distortion correction region by combining the mapping region of each semantic feature region in the acquired image, and repeat step S3 to output the optimized corrected image and the optimized set of semantic information for the image.

[0009] S5. Obtain the minimum bounding rectangle region corresponding to each element in the optimized image semantic information set in the optimized corrected image and mark it. Obtain the mapping region of each minimum bounding rectangle region in the captured image. Bind the center point of each mapping region to the corresponding coordinates based on the preset inverse perspective mapping form and the distance between the corresponding ultra-wide-angle security ranging lens and the corresponding minimum bounding rectangle region. Output the distortion correction information of the captured image.

[0010] This invention dynamically divides the image captured by an ultra-wide-angle security ranging lens into multiple different distortion correction regions and achieves differentiated correction across these regions. For each region (such as the central region, transition region, and edge region), different correction strategies are applied. Strong correction is used in the central region to ensure ranging accuracy and linearity, while weak correction is used in the edge region to retain more field-of-view information. Considering the abrupt changes in distortion correction strategies between the central and edge regions, the concept of a transition region is introduced to achieve adaptive adjustment of the region division and seamless integration of correction strategies, with smoothing applied to the transition region.

[0011] Furthermore, the objective optimization model constructed in S1 is as follows:

[0012]

[0013] Among them, w i The preset weight factor represents the corresponding preset region type obtained based on semantic segmentation; f(i) represents the distance between the coordinates of the i-th pixel in the image within the monitoring area calculated by inverse perspective mapping and the corresponding ultra-wide-angle security ranging lens; ||f(i)-d i || 2 f(i)-d i The corresponding squared Euclidean distance; d i The distance between the actual coordinates of the i-th pixel in the image within the monitoring area and the corresponding ultra-wide-angle security ranging lens is represented by N; N represents the number of pixels in the image within the monitoring area.

[0014] Furthermore, in step S2, the distance between the i-th pixel in the captured image and the corresponding center point of the captured image is denoted as Ai, and the distance coefficient of the i-th pixel in the captured image based on the captured image is obtained. The calculation formula involved is as follows:

[0015] Ri = Ai / max{Ai|i∈[1,N]}

[0016] Where Ri represents the distance coefficient of the i-th pixel in the captured image based on the captured image; max{} represents the maximum value function;

[0017] The distortion correction region types include central region, transition region, and edge region, and different distortion correction region types are bound to different preset distortion correction strategies. The upper limit of the deviation of the preset distortion correction strategy bound to the central region and edge region for the image distortion correction result is fixed. The deviation of the distortion correction strategy for the image distortion correction result represents the distance between the distortion correction result and the actual position in the actual image. The upper limit of the deviation of the distortion correction strategy bound to the transition region for the image distortion correction result changes dynamically with the distance coefficient of the pixel based on the captured image, and the calculation formula involved is as follows:

[0018] kb j =μ i ·kc+(1-μ i )·ke

[0019] Among them, kb j denoted by , represents the distortion correction strategy bound to the j-th pixel within the transition region; kc represents the upper limit of the deviation of the preset distortion correction strategy bound to the center region for the image distortion correction result; ke represents the upper limit of the deviation of the preset distortion correction strategy bound to the edge region for the image distortion correction result; μ i This represents the adjustment factor corresponding to the j-th pixel within the transition region;

[0020]

[0021] Rmax represents the maximum value of the distance coefficients for each pixel within the transition region based on the captured image; R j Rmin represents the distance coefficient of the j-th pixel within the transition region based on the captured image; Rmin represents the minimum value among the distance coefficients of each pixel within the transition region based on the captured image.

[0022] The region consisting of all pixels whose corresponding pixel positions in the captured image have a distance coefficient less than a first preset value based on the captured image is designated as the central region; the region consisting of all pixels whose corresponding pixel positions in the captured image have a distance coefficient greater than or equal to the first preset value and less than or equal to the second preset value based on the captured image is designated as the transition region; and the region consisting of all pixels whose corresponding pixel positions in the captured image have a distance coefficient greater than the second preset value based on the captured image is designated as the edge region.

[0023] In this invention, the distortion correction strategy bound to the transition region is set to change dynamically in order to avoid poor user viewing experience due to abrupt changes in the distortion correction strategy in the corrected image. The purpose is to smooth the distortion correction strategy corresponding to the region between the central region and the edge region, so as to achieve adaptive adjustment of the region division and seamless integration between distortion correction strategies.

[0024] Furthermore, in the process of generating the corresponding semantic information set of the captured image in S4, the captured image is compared with the static image stored in the historical data. All pixels whose absolute value of the difference between the corresponding pixel gray value in the captured image and the corresponding pixel gray value in the static image is greater than a threshold are marked. Adjacent marked pixels are divided into the same semantic information to generate different pixel clusters, and each pixel cluster corresponds to a semantic information.

[0025] Furthermore, during the process of updating the divided distortion correction regions in S4, the most recently acquired set of image semantic information is obtained, as well as the mapping region of the semantic feature region corresponding to each element in the obtained set of image semantic information in the captured image. The mapping region of the semantic feature region corresponding to the g-th element in the obtained set of image semantic information in the captured image is denoted as Qg. If the distortion correction region type to which the pixel in Qg belongs is one, then the distortion correction region to which Qg belongs is not updated. If the distortion correction region type to which the pixel in Qg belongs is multiple, then the distortion correction region closest to the center point of the captured image among the distortion correction regions to which Qg belongs is obtained and denoted as the distortion optimization object region. The update result of the distortion optimization object region is the union region of the original distortion optimization object region and Qg. The update results of each distortion correction region whose distortion correction region type to which the pixel in Qg belongs is not a distortion optimization object region are the remaining regions in the original corresponding distortion correction regions that do not belong to the update result of the distortion optimization object region.

[0026] This invention dynamically updates each distortion correction zone to take into account the security functions of ultra-wide-angle security ranging lenses. Different distortion correction zones employ different distortion correction strategies, each with its own emphasis (e.g., the central zone emphasizes ranging accuracy and linearity, while the edge zone emphasizes the field of view). However, in security applications, when an intrusion target is detected, not only is the distance to the intrusion target required, but the accuracy of the image information is also critical. Furthermore, when the same intrusion target is located in different distortion correction zones, the image information accuracy of some intrusion targets may not meet security requirements. Therefore, a mechanism for dynamically dividing distortion correction zones (not limited to preset distance coefficient index filtering conditions) is needed.

[0027] Furthermore, the distortion correction information of the acquired image output in S5 includes the last optimized correction image, the minimum rectangular region corresponding to each element in the semantic information set of the corresponding optimized correction image, and the ranging result bound to each minimum rectangular marker region.

[0028] A distortion correction system for an ultra-wide-angle security rangefinder lens, the system comprising:

[0029] The target optimization model construction module obtains the monitoring area corresponding to the ultra-wide-angle security ranging lens, takes the mapping error between pixel coordinates and physical distance within the monitoring area as the optimization target, and constructs a target optimization model.

[0030] The distortion correction strategy dynamic configuration module acquires the captured image of the monitoring area by the ultra-wide-angle security ranging lens, dynamically divides the distortion correction area of ​​the captured image according to the distance between each pixel in the captured image and the corresponding center point of the captured image, and configures a corresponding distortion correction strategy for each divided distortion correction area.

[0031] The image correction module performs distortion correction on the captured image according to the divided distortion correction areas and the configured distortion correction strategies, and obtains a corrected image of the captured image.

[0032] The distortion correction region update module extracts semantic information from the corrected image of the acquired frame to generate a set of semantic information for the corresponding acquired frame; it performs semantic feature segmentation on the corrected image of the acquired frame based on the set of semantic information to obtain each segmented semantic feature region; and it updates each segmented distortion correction region by combining the mapping region of each semantic feature region in the acquired frame, outputting an optimized corrected image and an optimized set of semantic information for the acquired frame.

[0033] The distortion correction result output module acquires and marks the minimum bounding rectangle region corresponding to each element in the optimized image semantic information set in the optimized corrected image. It then acquires the mapping region of each minimum bounding rectangle region in the captured image and binds the center point of each mapping region to the corresponding minimum bounding rectangle region based on the coordinates of the preset inverse perspective mapping form and the distance between the corresponding ultra-wide-angle security ranging lens. Finally, it outputs the distortion correction information of the captured image.

[0034] Preferably, the distortion correction region update module includes an image semantic information extraction unit, a semantic feature region segmentation unit, and a correction image optimization unit;

[0035] The image semantic information extraction unit extracts image semantic information from the corrected image of the captured image and generates an image semantic information set for the corresponding captured image.

[0036] The semantic feature region segmentation unit performs semantic feature segmentation on the corrected image of the acquired image based on the set of semantic information of the image, and obtains each segmented semantic feature region.

[0037] The image correction optimization unit combines the mapping area of ​​each semantic feature region in the captured image to update each divided distortion correction region, and outputs the optimized correction image and the optimized image semantic information set.

[0038] Preferably, the distortion correction result output module includes a region mapping analysis unit, a ranging data binding unit, and a distortion correction information output unit;

[0039] The region mapping analysis unit obtains and marks the minimum bounding rectangle region corresponding to each element in the optimized image semantic information set in the optimized corrected image, and obtains the mapping region of each minimum bounding rectangle region in the captured image.

[0040] The ranging data binding unit binds the center point of each mapping area to the corresponding coordinates based on the preset inverse perspective mapping form, the distance between the corresponding ultra-wide-angle security ranging lens, and the corresponding minimum bounding rectangle area.

[0041] The distortion correction information output unit is used to output distortion correction information of the captured image.

[0042] Compared with the prior art, the beneficial effects achieved by the present invention are: the present invention can dynamically divide the image captured by the ultra-wide-angle security ranging lens into multiple different distortion correction areas and realize multi-area differentiated correction, taking into account the ranging accuracy and field of view integrity of the image captured by the ultra-wide-angle security ranging lens, and solving the problem of misalignment between the correction target and security needs. Attached Figure Description

[0043] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0044] Figure 1 This is a schematic diagram of the structure of an ultra-wide-angle security rangefinder lens distortion correction system according to the present invention;

[0045] Figure 2 This is a flowchart illustrating a distortion correction method for an ultra-wide-angle security rangefinder lens according to the present invention. Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] Please see Figures 1-2 The present invention provides a technical solution: such as Figure 1 As shown, this embodiment provides an ultra-wide-angle security rangefinder lens distortion correction system, the system comprising:

[0048] The target optimization model construction module obtains the monitoring area corresponding to the ultra-wide-angle security ranging lens, takes the mapping error between pixel coordinates and physical distance within the monitoring area as the optimization target, and constructs a target optimization model.

[0049] The distortion correction strategy dynamic configuration module acquires the captured image of the monitoring area by the ultra-wide-angle security ranging lens, dynamically divides the distortion correction area of ​​the captured image according to the distance between each pixel in the captured image and the corresponding center point of the captured image, and configures a corresponding distortion correction strategy for each divided distortion correction area.

[0050] The image correction module performs distortion correction on the captured image according to the divided distortion correction areas and the configured distortion correction strategies, and obtains a corrected image of the captured image.

[0051] The distortion correction region update module includes a semantic information extraction unit, a semantic feature region segmentation unit, and a corrected image optimization unit.

[0052] The image semantic information extraction unit extracts image semantic information from the corrected image of the captured image and generates an image semantic information set for the corresponding captured image.

[0053] The semantic feature region segmentation unit performs semantic feature segmentation on the corrected image of the acquired image based on the set of semantic information of the image, and obtains each segmented semantic feature region.

[0054] The image correction optimization unit combines the mapping area of ​​each semantic feature region in the captured image to update each divided distortion correction region, and outputs the optimized correction image and the optimized image semantic information set.

[0055] The distortion correction result output module includes a region mapping analysis unit, a ranging data binding unit, and a distortion correction information output unit.

[0056] The region mapping analysis unit obtains and marks the minimum bounding rectangle region corresponding to each element in the optimized image semantic information set in the optimized corrected image, and obtains the mapping region of each minimum bounding rectangle region in the captured image.

[0057] The ranging data binding unit binds the center point of each mapping area to the corresponding coordinates based on the preset inverse perspective mapping form, the distance between the corresponding ultra-wide-angle security ranging lens, and the corresponding minimum bounding rectangle area.

[0058] The distortion correction information output unit is used to output distortion correction information of the captured image.

[0059] like Figure 2 As shown, this embodiment provides a distortion correction method for an ultra-wide-angle security rangefinder lens, the method comprising the following steps:

[0060] S1. Obtain the monitoring area corresponding to the ultra-wide-angle security ranging lens, and use the mapping error between pixel coordinates and physical distance within the monitoring area as the optimization target to construct a target optimization model.

[0061] It should be specifically noted that the objective optimization model constructed in S1 is as follows:

[0062]

[0063] Among them, w i The preset weight factor represents the corresponding preset region type obtained based on semantic segmentation; f(i) represents the distance between the coordinates of the i-th pixel in the image within the monitoring area calculated by inverse perspective mapping and the corresponding ultra-wide-angle security ranging lens; ||f(i)-d i || 2 f(i)-d i The corresponding squared Euclidean distance; d iThe distance between the actual coordinates of the i-th pixel in the image within the monitoring area and the corresponding ultra-wide-angle security ranging lens is represented by N; N represents the number of pixels in the image within the monitoring area.

[0064] In this embodiment, the preset area types include ground areas and sky areas. The preset weight factor for the sky area is 0, and the preset weight factor for the ground area is 1. The preset area types also include other target areas, and the corresponding preset weight factors have a value range of (0,1).

[0065] S2. Acquire the image captured by the ultra-wide-angle security rangefinder for the monitoring area. Based on the distance between each pixel in the image and the center point of the corresponding image, dynamically divide the distortion correction area of ​​the image and configure the corresponding distortion correction strategy for each divided distortion correction area.

[0066] It should be noted that in S2, the distance between the i-th pixel in the captured image and the corresponding center point of the captured image is denoted as Ai, and the distance coefficient of the i-th pixel in the captured image based on the captured image is obtained. The calculation formula involved is as follows:

[0067] Ri = Ai / max{Ai|i∈[1,N]}

[0068] Where Ri represents the distance coefficient of the i-th pixel in the captured image based on the captured image; max{} represents the maximum value function;

[0069] The distortion correction region types include central region, transition region, and edge region, and different distortion correction region types are bound to different preset distortion correction strategies. The upper limit of the deviation of the preset distortion correction strategy bound to the central region and edge region for the image distortion correction result is fixed. The deviation of the distortion correction strategy for the image distortion correction result represents the distance between the distortion correction result and the actual position in the actual image. The upper limit of the deviation of the distortion correction strategy bound to the transition region for the image distortion correction result changes dynamically with the distance coefficient of the pixel based on the captured image, and the calculation formula involved is as follows:

[0070] kb j =μ i ·kc+(1-μ i )·ke

[0071] Among them, kb j denoted by , represents the distortion correction strategy bound to the j-th pixel within the transition region; kc represents the upper limit of the deviation of the preset distortion correction strategy bound to the center region for the image distortion correction result; ke represents the upper limit of the deviation of the preset distortion correction strategy bound to the edge region for the image distortion correction result; μ iThis represents the adjustment factor corresponding to the j-th pixel within the transition region;

[0072]

[0073] Rmax represents the maximum value of the distance coefficients for each pixel within the transition region based on the captured image; R j Rmin represents the distance coefficient of the j-th pixel within the transition region based on the captured image; Rmin represents the minimum value among the distance coefficients of each pixel within the transition region based on the captured image.

[0074] The region consisting of all pixels whose corresponding pixel positions in the captured image have a distance coefficient less than a first preset value based on the captured image is designated as the central region; the region consisting of all pixels whose corresponding pixel positions in the captured image have a distance coefficient greater than or equal to the first preset value and less than or equal to the second preset value based on the captured image is designated as the transition region; and the region consisting of all pixels whose corresponding pixel positions in the captured image have a distance coefficient greater than the second preset value based on the captured image is designated as the edge region.

[0075] In this embodiment, the image captured by the ultra-wide-angle security ranging lens is dynamically divided into multiple different distortion correction regions (such as the central region, transition region, and edge region), and multi-region differentiated correction is implemented. Different intensity correction strategies are applied to different regions (in terms of requirements: the central region needs high-precision ranging and linearity, while the edge region needs to retain as much field of view information as possible. Therefore, when constructing the distortion correction strategy, the central region focuses on high-precision ranging and needs strong correction; the edge region focuses on preserving the field of view and needs weak correction; and considering the abrupt changes between the distortion correction strategies corresponding to the central region and the edge region, the concept of a transition region is introduced to ensure the connection between different distortion correction strategies, thereby realizing adaptive adjustment of the region division and seamless integration of correction strategies).

[0076] In this embodiment, when dividing the distortion correction area, if Ri < 0.3, the i-th pixel in the captured image is determined to belong to the central area.

[0077] When 0.3 < Ri < 0.7, the i-th pixel in the captured image is determined to belong to the transition region;

[0078] When Ri > 0.7, the i-th pixel in the captured image is determined to belong to the edge region;

[0079] S3. Based on the defined distortion correction regions and the corresponding distortion correction strategies, perform distortion correction on the captured image to obtain the corrected image of the captured image.

[0080] S4. Extract semantic information from the corrected image of the acquired image to generate a set of semantic information for the corresponding acquired image; perform semantic feature segmentation on the corrected image of the acquired image based on the set of semantic information to obtain each segmented semantic feature region; and update each distortion correction region by combining the mapping region of each semantic feature region in the acquired image, and repeat step S3 to output the optimized corrected image and the optimized set of semantic information for the image.

[0081] In the process of generating the corresponding semantic information set of the captured image in S4, the captured image is compared with the static image stored in the historical data. All pixels whose absolute value of the difference between the corresponding pixel gray value in the captured image and the corresponding pixel gray value in the static image is greater than the threshold are marked. Adjacent marked pixels are divided into the same semantic information to generate different pixel clusters. Each pixel cluster corresponds to a semantic information.

[0082] In the process of updating the divided distortion correction regions in S4, the most recently acquired set of image semantic information is obtained, as well as the mapping region of the semantic feature region corresponding to each element in the obtained set of image semantic information in the captured image. The mapping region of the semantic feature region corresponding to the g-th element in the obtained set of image semantic information in the captured image is denoted as Qg. If the distortion correction region type to which the pixel in Qg belongs is one, then the distortion correction region to which Qg belongs is not updated. If the distortion correction region type to which the pixel in Qg belongs is multiple, then the distortion correction region closest to the center point of the captured image in each distortion correction region to which Qg belongs is obtained and denoted as the distortion optimization object region. The update result of the distortion optimization object region is the union region of the original distortion optimization object region and Qg. The update results of each distortion correction region to which the pixel in Qg belongs is not a distortion optimization object region, respectively, are the remaining regions in the original corresponding distortion correction regions that do not belong to the update result of the distortion optimization object region.

[0083] S5. Obtain the minimum bounding rectangle region corresponding to each element in the optimized image semantic information set in the optimized corrected image and mark it. Obtain the mapping region of each minimum bounding rectangle region in the captured image. Bind the center point of each mapping region to the corresponding coordinates based on the preset inverse perspective mapping form and the distance between the corresponding ultra-wide-angle security ranging lens and the corresponding minimum bounding rectangle region. Output the distortion correction information of the captured image.

[0084] The distortion correction information of the acquired image output in S5 includes the last optimized correction image, the minimum rectangular region corresponding to each element in the semantic information set of the corresponding optimized correction image, and the ranging result bound to each minimum rectangular marker region.

[0085] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0086] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A distortion correction method for an ultra-wide-angle security rangefinder lens, characterized in that, The method includes the following steps: S1. Obtain the monitoring area corresponding to the ultra-wide-angle security ranging lens, and use the mapping error between pixel coordinates and physical distance within the monitoring area as the optimization target to construct a target optimization model. S2. Acquire the image captured by the ultra-wide-angle security rangefinder for the monitoring area. Based on the distance between each pixel in the image and the center point of the corresponding image, dynamically divide the distortion correction area of ​​the image and configure the corresponding distortion correction strategy for each divided distortion correction area. S3. Based on the defined distortion correction regions and the corresponding distortion correction strategies, perform distortion correction on the captured image to obtain the corrected image of the captured image. S4. Extract semantic information from the corrected image of the acquired image to generate a set of semantic information for the corresponding acquired image; perform semantic feature segmentation on the corrected image of the acquired image based on the set of semantic information to obtain each segmented semantic feature region; and update each distortion correction region by combining the mapping region of each semantic feature region in the acquired image, and repeat step S3 to output the optimized corrected image and the optimized set of semantic information for the image. S5. Obtain the minimum bounding rectangle region corresponding to each element in the optimized image semantic information set in the optimized corrected image and mark it. Obtain the mapping region of each minimum bounding rectangle region in the captured image. Bind the center point of each mapping region to the corresponding coordinates based on the preset inverse perspective mapping form and the distance between the corresponding ultra-wide-angle security ranging lens and the corresponding minimum bounding rectangle region. Output the distortion correction information of the captured image.

2. The distortion correction method for an ultra-wide-angle security rangefinder lens according to claim 1, characterized in that: The objective optimization model constructed in S1 is as follows: Among them, w i The preset weight factor represents the corresponding preset region type obtained based on semantic segmentation; f(i) represents the distance between the coordinates of the i-th pixel in the image within the monitoring area calculated by inverse perspective mapping and the corresponding ultra-wide-angle security ranging lens; ||f(i)-d i || 2 f(i)-d i The corresponding squared Euclidean distance; d i The distance between the actual coordinates of the i-th pixel in the image within the monitoring area and the corresponding ultra-wide-angle security ranging lens is represented by N; N represents the number of pixels in the image within the monitoring area.

3. The distortion correction method for an ultra-wide-angle security rangefinder lens according to claim 2, characterized in that: In step S2, the distance between the i-th pixel in the captured image and the corresponding center point of the captured image is denoted as Ai, and the distance coefficient of the i-th pixel in the captured image based on the captured image is obtained. The calculation formula involved is as follows: Ri = Ai / max{Ai|i∈[1,N]} Where Ri represents the distance coefficient of the i-th pixel in the captured image based on the captured image; max{} represents the maximum value function; The distortion correction region types include central region, transition region, and edge region, and different distortion correction region types are bound to different preset distortion correction strategies. The upper limit of the deviation of the preset distortion correction strategy bound to the central region and edge region for the image distortion correction result is fixed. The deviation of the distortion correction strategy for the image distortion correction result represents the distance between the distortion correction result and the actual position in the actual image. The upper limit of the deviation of the distortion correction strategy bound to the transition region for the image distortion correction result changes dynamically with the distance coefficient of the pixel based on the captured image, and the calculation formula involved is as follows: kb j =μ i ·kc+(1-μ i )·ke Among them, kb j denoted by , represents the distortion correction strategy bound to the j-th pixel within the transition region; kc represents the upper limit of the deviation of the preset distortion correction strategy bound to the center region for the image distortion correction result; ke represents the upper limit of the deviation of the preset distortion correction strategy bound to the edge region for the image distortion correction result; μ i This represents the adjustment factor corresponding to the j-th pixel within the transition region; Rmax represents the maximum value of the distance coefficients for each pixel within the transition region based on the captured image; R j Rmin represents the distance coefficient of the j-th pixel within the transition region based on the captured image; Rmin represents the minimum value among the distance coefficients of each pixel within the transition region based on the captured image. The region consisting of all pixels whose corresponding pixel positions in the captured image have a distance coefficient less than a first preset value based on the captured image is designated as the central region; the region consisting of all pixels whose corresponding pixel positions in the captured image have a distance coefficient greater than or equal to the first preset value and less than or equal to the second preset value based on the captured image is designated as the transition region; and the region consisting of all pixels whose corresponding pixel positions in the captured image have a distance coefficient greater than the second preset value based on the captured image is designated as the edge region.

4. The distortion correction method for an ultra-wide-angle security rangefinder lens according to claim 1, characterized in that: In the process of generating the corresponding semantic information set of the captured image in S4, the captured image is compared with the static image stored in the historical data. All pixels whose absolute value of the difference between the corresponding pixel gray value in the captured image and the corresponding pixel gray value in the static image is greater than a threshold are marked. Adjacent marked pixels are divided into the same semantic information to generate different pixel clusters. Each pixel cluster corresponds to a semantic information.

5. The distortion correction method for an ultra-wide-angle security rangefinder lens according to claim 1, characterized in that: During the process of updating the various distortion correction regions in S4, the most recently acquired set of image semantic information is obtained, as well as the mapping region of the semantic feature region corresponding to each element in the obtained set of image semantic information in the captured image. The mapping region of the semantic feature region corresponding to the g-th element in the obtained set of image semantic information in the captured image is denoted as Qg. If the distortion correction region type to which the pixel in Qg belongs is one, then the distortion correction region to which Qg belongs is not updated; if the distortion correction region type to which the pixel in Qg belongs is multiple, then the distortion correction region closest to the center point of the captured image among each distortion correction region to which Qg belongs is obtained and denoted as the distortion optimization object region. The update result of the distortion optimization object region is the union region of the original distortion optimization object region and Qg; the update results of each distortion correction region whose distortion correction region type to which the pixel in Qg belongs is not a distortion optimization object region are the remaining regions in the original corresponding distortion correction region that do not belong to the update result of the distortion optimization object region.

6. The distortion correction method for an ultra-wide-angle security rangefinder lens according to claim 1, characterized in that: The distortion correction information of the acquired image output in S5 includes the last optimized correction image, the minimum rectangular region corresponding to each element in the semantic information set of the corresponding optimized correction image, and the ranging result bound to each minimum rectangular marker region.

7. A distortion correction system for an ultra-wide-angle security rangefinder lens, employing the distortion correction method for an ultra-wide-angle security rangefinder lens as described in any one of claims 1-6, characterized in that, The system includes: The target optimization model construction module obtains the monitoring area corresponding to the ultra-wide-angle security ranging lens, takes the mapping error between pixel coordinates and physical distance within the monitoring area as the optimization target, and constructs a target optimization model. The distortion correction strategy dynamic configuration module acquires the captured image of the monitoring area by the ultra-wide-angle security ranging lens, dynamically divides the distortion correction area of ​​the captured image according to the distance between each pixel in the captured image and the corresponding center point of the captured image, and configures a corresponding distortion correction strategy for each divided distortion correction area. The image correction module performs distortion correction on the captured image according to the divided distortion correction areas and the configured distortion correction strategies, and obtains a corrected image of the captured image. The distortion correction region update module extracts semantic information from the corrected image of the acquired frame to generate a set of semantic information for the corresponding acquired frame; it performs semantic feature segmentation on the corrected image of the acquired frame based on the set of semantic information to obtain each segmented semantic feature region; and it updates each segmented distortion correction region by combining the mapping region of each semantic feature region in the acquired frame, outputting an optimized corrected image and an optimized set of semantic information for the acquired frame. The distortion correction result output module acquires and marks the minimum bounding rectangle region corresponding to each element in the optimized image semantic information set in the optimized corrected image. It then acquires the mapping region of each minimum bounding rectangle region in the captured image and binds the center point of each mapping region to the corresponding minimum bounding rectangle region based on the coordinates of the preset inverse perspective mapping form and the distance between the corresponding ultra-wide-angle security ranging lens. Finally, it outputs the distortion correction information of the captured image.

8. The ultra-wide-angle security rangefinder lens distortion correction system according to claim 7, characterized in that: The distortion correction region update module includes a semantic information extraction unit, a semantic feature region segmentation unit, and a correction image optimization unit. The image semantic information extraction unit extracts image semantic information from the corrected image of the captured image and generates an image semantic information set for the corresponding captured image. The semantic feature region segmentation unit performs semantic feature segmentation on the corrected image of the acquired image based on the set of semantic information of the image, and obtains each segmented semantic feature region. The image correction optimization unit combines the mapping area of ​​each semantic feature region in the captured image to update each divided distortion correction region, and outputs the optimized correction image and the optimized image semantic information set.

9. The ultra-wide-angle security rangefinder lens distortion correction system according to claim 7, characterized in that: The distortion correction result output module includes a region mapping analysis unit, a ranging data binding unit, and a distortion correction information output unit. The region mapping analysis unit obtains and marks the minimum bounding rectangle region corresponding to each element in the optimized image semantic information set in the optimized corrected image, and obtains the mapping region of each minimum bounding rectangle region in the captured image. The ranging data binding unit binds the center point of each mapping area to the corresponding coordinates based on the preset inverse perspective mapping form, the distance between the corresponding ultra-wide-angle security ranging lens, and the corresponding minimum bounding rectangle area. The distortion correction information output unit is used to output distortion correction information of the captured image.

Images