Thermal infrared imager smear elimination method, system and device and storage medium

Abstract

The invention relates to a thermal infrared imager smear elimination method and system, and belongs to the technical field of infrared imaging. The method comprises the following steps: acquiring an original image frame set of the thermal infrared imager in a continuous time sequence, and carrying out non-uniformity correction and Gaussian filtering preprocessing on an original image; carrying out inter-frame difference calculation on adjacent frames of images, and extracting a moving target candidate region through thresholding processing and morphological expansion and corrosion operation; calculating a pixel motion vector in the candidate area by adopting an optical flow algorithm, and generating a motion track set through a track association algorithm after abnormal values are filtered; calculating a trajectory space-time consistency score by adopting Kalman filtering or a Hungary algorithm, and separating a real target trajectory from a smear region; a point spread function is constructed for a real target trajectory, and a smear-free target contour image is reconstructed by using Wiener filtering deconvolution. The invention further relates to an electronic device and a computer readable storage medium for implementing the method, the smear problem of the thermal infrared imager is effectively solved, and the image quality is improved.

Description

Technical Field

[0001] The present invention relates to the field of infrared imaging technology, and specifically refers to a method, system, device and storage medium for eliminating smear in an infrared thermal imager. Background Art

[0002] An infrared thermal imager realizes imaging by capturing the infrared radiation on the surface of an object, and is widely used in fields such as security monitoring, industrial inspection, and medical diagnosis. However, when there are fast-moving targets in the observed scene or the device itself vibrates, the infrared thermal imager is prone to generate smear due to relative displacement within the integration time, resulting in blurred target contours and affecting subsequent target recognition and analysis. Traditional smear elimination methods mostly rely on hardware optimization or simple inter-frame averaging, and have limited effects on smear processing in complex motion scenes, and are prone to loss of target details.

[0003] In view of this, it is necessary to improve the above-mentioned defects.

[0004] The information disclosed in this background art section is only intended to enhance the overall understanding of the present invention, and should not be regarded as an admission or any form of implication that this information constitutes prior art already known to those of ordinary skill in the art. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to overcome the above-mentioned defects and provide a method for eliminating smear in an infrared thermal imager.

[0006] To solve the above technical problem, the technical solution provided by the present invention is as follows:

[0007] On the one hand, the present invention provides a method for eliminating smear in an infrared thermal imager, including the following steps: obtaining a set of original image frames of the infrared thermal imager in a continuous time series; performing inter-frame difference calculation on adjacent frame images, and combining morphological processing to extract candidate regions of moving targets; within the candidate regions, calculating a pixel-level motion vector field by an optical flow method to generate a set of motion trajectories of the targets; performing spatio-temporal consistency analysis on the set of motion trajectories to separate real target trajectories that satisfy the motion continuity constraint from smear regions that do not meet the constraint; establishing a motion blur model based on the real target trajectories, and reconstructing a target contour image without smear through deconvolution operation.

[0008] Optionally, the obtaining a set of original image frames of the infrared thermal imager in a continuous time series includes: collecting several consecutive infrared images, and performing non-uniformity correction and Gaussian filtering preprocessing on the original images to remove fixed pattern noise and random noise.

[0009] Optionally, performing inter-frame difference calculation on the adjacent frame images and extracting a candidate region of the moving target in combination with morphological processing includes: calculating an inter-frame difference image, and performing thresholding processing on the difference image to obtain a binary image; performing morphological dilation and erosion operations on the binary image to remove isolated noise points and connect adjacent regions, so as to obtain a candidate region of the moving target.

[0010] Optionally, within the candidate region, calculating a pixel-level motion vector field by an optical flow method to generate a set of motion trajectories of the target includes: using the Lucas-Kanade optical flow algorithm to calculate the motion vectors of pixels within the candidate region; performing outlier filtering on the motion vectors to remove vectors with a norm exceeding a threshold, and generating an effective motion vector field; generating pixel-level motion trajectories based on the motion vectors of adjacent frames through a trajectory association algorithm.

[0011] Optionally, performing spatio-temporal consistency analysis on the set of motion trajectories to separate the real target trajectories that satisfy the motion continuity constraint from the ghosting regions that do not conform to the constraint includes: based on the pre-established spatio-temporal consistency constraint conditions, using the Kalman filter or the Hungarian algorithm to perform time series association on the trajectories, so as to calculate the spatio-temporal consistency score of each trajectory; based on the calculated spatio-temporal consistency score, determining whether the target trajectory is a real target trajectory or a ghosting region.

[0012] Optionally, establishing a motion blur model based on the real target trajectories and reconstructing a target contour image without ghosting through deconvolution operation includes: constructing a point spread function of motion blur for the real target trajectories; performing deconvolution operation using Wiener filtering, so as to reconstruct a target contour image without ghosting.

[0013] On the other hand, the present invention also provides an infrared thermal imager ghosting elimination system, including: an image acquisition module for acquiring a set of original image frames of an infrared thermal imager in a continuous time series; a motion detection module for performing inter-frame difference calculation on adjacent frame images and extracting a candidate region of the moving target in combination with morphological processing; an optical flow calculation module for calculating a pixel-level motion vector field by an optical flow method within the candidate region to generate a set of motion trajectories of the target; a trajectory analysis module for performing spatio-temporal consistency analysis on the set of motion trajectories to separate the real target trajectories that satisfy the motion continuity constraint from the ghosting regions that do not conform to the constraint; an image reconstruction module for establishing a motion blur model based on the real target trajectories and reconstructing a target contour image without ghosting through deconvolution operation.

[0014] Optionally, the image acquisition module further includes: a preprocessing module for acquiring a plurality of consecutive infrared images and performing non-uniformity correction and Gaussian filtering preprocessing on the original images to remove fixed pattern noise and random noise.

[0015] On the other hand, the present invention also provides an electronic device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method for eliminating smear of an infrared thermal imager.

[0016] On the other hand, the present invention also provides a computer-readable storage medium, which stores computer instructions for causing a computer to execute the method for eliminating smear of an infrared thermal imager.

[0017] The present invention quickly locates the moving area through inter-frame difference and morphological processing, accurately calculates the pixel-level motion vector by using the optical flow method, effectively distinguishes the real target from the smear by combining spatio-temporal consistency analysis, and finally reconstructs a clear image through deconvolution with edge-preserving constraints. Compared with the traditional method, the present invention has stronger adaptability to complex motion scenes, can retain the target details while eliminating the smear, and improves the imaging quality of the infrared thermal imager. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a schematic flowchart of a method for eliminating smear of an infrared thermal imager provided by an embodiment of the present invention;

[0019] Figure 2 is provided by an embodiment of the present invention Figure 1 sub-flowchart of S100 therein;

[0020] Figure 3 is provided by an embodiment of the present invention Figure 1 sub-flowchart of S200 therein;

[0021] Figure 4 is provided by an embodiment of the present invention Figure 1 sub-flowchart of S300 therein;

[0022] Figure 5 is provided by an embodiment of the present invention Figure 1 sub-flowchart of S400 therein;

[0023] Figure 6 is provided by an embodiment of the present invention Figure 1 sub-flowchart of S500 therein;

[0024] Figure 7 is a connection block diagram of a system for eliminating smear of an infrared thermal imager provided by an embodiment of the present invention;

[0025] Figure 8 is provided by an embodiment of the present invention Figure 7 connection block diagram of the image acquisition module therein.

[0026] Reference Signs:

[0027] 100, Infrared thermal imager smear elimination system; 110, Image acquisition module; 111, Preprocessing module; 120, Motion detection module; 130, Optical flow calculation module; 140, Trajectory analysis module; 150, Image reconstruction module. Detailed Embodiment

[0028] Here, exemplary embodiments will be described in detail, and examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. On the contrary, they are merely examples of devices consistent with some aspects of the present disclosure as detailed in the appended claims.

[0029] As described above, an infrared thermal imager achieves imaging by capturing the infrared radiation on the surface of an object and is widely used in fields such as security monitoring, industrial inspection, and medical diagnosis. However, when there are fast-moving targets in the observed scene or the device itself vibrates, the infrared thermal imager is prone to smearing due to relative displacement within the integration time, resulting in blurred target contours and affecting subsequent target recognition and analysis. Traditional smear elimination methods mostly rely on hardware optimization or simple inter-frame averaging, and have limited effects on smearing in complex motion scenarios, and are prone to loss of target details.

[0030] In view of this, the present invention provides an infrared thermal imager smear elimination method, system, device, and storage medium to solve the above problems. The present invention solves the problems in the following ways.

[0031] Embodiment 1:

[0032] Please refer to the attached Figure 1 As shown in the figure, the present invention provides an infrared thermal imager smear elimination method, and the method includes the following steps:

[0033] S100. Obtain a set of original image frames of the infrared thermal imager in a continuous time series.

[0034] S200. Perform inter-frame difference calculation on adjacent frame images, and combine morphological processing to extract candidate regions of moving targets.

[0035] S300. In the candidate region, calculate the pixel-level motion vector field by the optical flow method to generate a set of target motion trajectories.

[0036] S400. Perform spatio-temporal consistency analysis on the set of motion trajectories to separate real target trajectories that satisfy the motion continuity constraint from smear regions that do not meet the constraint.

[0037] S500. Establish a motion blur model based on the real target trajectory and reconstruct the target contour image without ghosting through deconvolution operation.

[0038] The present invention quickly locates the moving area through inter-frame difference and morphological processing, accurately calculates the pixel-level motion vector by using the optical flow method, effectively distinguishes the real target from the ghosting by combining spatio-temporal consistency analysis, and finally reconstructs a clear image through deconvolution with edge-preserving constraints. Compared with the traditional method, the present invention has stronger adaptability to complex motion scenes, can retain the target details while eliminating the ghosting, and improves the imaging quality of the infrared thermal imager.

[0039] Embodiment 2:

[0040] Based on the above embodiment, in order to further clearly explain the technical solution therein, the present invention also provides Embodiment 2, as Figures 2 - 6 shown, in Embodiment 2.

[0041] In this Embodiment 2, the S100 may further include:

[0042] S110. Collect a plurality of consecutive infrared images, and perform non-uniformity correction and Gaussian filtering preprocessing on the original images to remove fixed pattern noise and random noise.

[0043] For example: Collect at least N consecutive infrared images at a frame rate f, denoted as , where represents the t-th frame image.

[0044] In this Embodiment 2, the S200 may further include:

[0045] S210. Calculate the difference image between adjacent frames, and perform thresholding on the difference image to obtain a binary image.

[0046] S220. Perform morphological dilation and erosion operations on the binary image to remove isolated noise points and connect adjacent regions, so as to obtain a candidate region of the moving target.

[0047] For example, and continuing with the above example, the calculation of the difference image between adjacent frames is expressed by the following formula:

[0048] (1)

[0049] where ;

[0050] Perform thresholding on the difference image to obtain a binary image , and the threshold T is adaptively determined based on the Otsu algorithm.

[0051] Perform morphological dilation and erosion operations to remove isolated noise points and connect adjacent regions, obtaining candidate regions of moving objects 。 。

[0052] In the second embodiment, the S300 may further include:

[0053] S310. Adopt the Lucas-Kanade optical flow algorithm to calculate the motion vectors of pixels within the candidate region.

[0054] S320. Filter outliers from the motion vectors, removing vectors whose norm exceeds the threshold to generate an effective motion vector field.

[0055] S330. Based on the motion vectors of adjacent frames, generate pixel-level motion trajectories through a trajectory association algorithm.

[0056] Illustrate by way of example and continue with the above example. When adopting the Lucas-Kanade optical flow algorithm to calculate the motion vectors of pixels within the candidate region of the motion vectors , it is necessary to satisfy the optical flow constraint equation: (2)

[0057] where , is the image gradient, is the temporal gradient;

[0058] Filter outliers from the motion vectors, removing vectors whose norm exceeds the threshold to generate an effective motion vector field ;

[0059] Based on the motion vectors of adjacent frames, generate pixel-level motion trajectories , where represents the th trajectory.

[0060] In the second embodiment, the S400 may further include:

[0061] S410. Based on the pre-established spatio-temporal consistency constraint conditions, use the Kalman filter or the Hungarian algorithm to perform time-series association on the trajectories, thereby calculating the spatio-temporal consistency score for each trajectory.

[0062] S420. Based on the calculated spatio-temporal consistency scores, determine whether the target trajectory is a real target trajectory or a ghosting region.

[0063] Illustrate by way of example and continue with the above example. The pre-established spatio-temporal consistency constraint conditions are: the trajectory time length is not less than , the deviation of adjacent frame positions does not exceed , and the rate of change of speed does not exceed .

[0064] The spatio-temporal consistency score of each trajectory is expressed by the following formula:

[0065] (3)

[0066] where is the trajectory length, is the speed consistency, is the acceleration consistency, is the weight coefficient.

[0067] Set a threshold , and judge the trajectory with score as a real target trajectory, otherwise it is a ghosting area.

[0068] In the second embodiment, the S500 may further include:

[0069] S510. For the real target trajectory, construct a point spread function of motion blur.

[0070] S520. Perform deconvolution operation using Wiener filtering to reconstruct the target contour image without ghosting.

[0071] Illustrate by way of example and continuing with the above example, the constructed point spread function of motion blur is expressed by the following formula:

[0072] (4)

[0073] where is the point spread function, is the trajectory length, is the position of the th frame in the trajectory, is the impulse function.

[0074] The process of deconvolution operation is expressed by the following formula:

[0075] (5)

[0076] where represents the reconstructed image, represents the convolution operation, is the regularization parameter, is the gradient operator.

[0077] In this embodiment, by performing non-uniformity correction and Gaussian filtering preprocessing on the original image, fixed pattern noise and random noise are effectively removed, providing a high-quality image basis for subsequent processing. The Otsu algorithm is used to adaptively determine the threshold for inter-frame difference processing, and combined with morphological operations, the candidate region of the moving target is accurately extracted, improving the accuracy of motion detection. The Lucas-Kanade optical flow algorithm is used to calculate the pixel motion vector and filter out outliers, generating an effective motion vector field. The pixel-level motion trajectory is obtained through the trajectory association algorithm, ensuring the reliability of motion information. Based on the spatio-temporal consistency constraint conditions, the Kalman filter or the Hungarian algorithm is used to calculate the spatio-temporal consistency score of the trajectory, accurately separating the real target trajectory from the ghosting area. A point spread function of motion blur is constructed, and Wiener filtering is used for deconvolution operation to realize the reconstruction of the target contour image without ghosting, improving the accuracy of image reconstruction and the overall reliability of the method, and effectively handling the ghosting problem in complex motion scenes.

[0078] Embodiment III:

[0079] Based on the above embodiments, in order to further clearly explain the technical solutions therein, the present invention also provides Embodiment III.

[0080] In Embodiment III, the S200 further includes: S230, calculating the multi-frame cumulative difference image for enhancing the detection sensitivity of slow-moving targets.

[0081] Illustrated by way of example and continuing with the above example, the multi-frame cumulative difference image is represented by the following formula:

[0082] (6)

[0083] Where is the number of cumulative frames,

[0084] In Embodiment III, the S300 further includes: S340, calculating the pixel-level motion vector field using the pyramid optical flow method, and the steps include: constructing an image pyramid and calculating the optical flow layer by layer on the multi-layer pyramid structure to process the pixel-level motion vector of large-displacement moving targets , improving the robustness of optical flow calculation.

[0085] In Embodiment III, the S400 further includes: S430, clustering the trajectories using the DBSCAN clustering algorithm based on the speed, direction, and spatio-temporal distribution of the motion trajectories, clustering the trajectories of the same target into one category, and excluding the interference of isolated noise trajectories on the separation of the ghosting area.

[0086] In Embodiment III, the S500 further includes: S530, the deconvolution operation combines edge-preserving constraints, specifically including:

[0087] During the deconvolution process, an edge detection operator is introduced to protect the edge region of the target contour and avoid edge blurring during the deconvolution process.

[0088] Illustrate with an example and continue with the above example. The above process is expressed by the following correction formula:

[0089] (7)

[0090] Where is the image region, is the edge indicator function, the edge region and the non-edge region .

[0091] In this embodiment, by calculating the multi-frame cumulative difference image, the detection sensitivity of slow-moving targets is enhanced, and the problem of poor detection effect of slow-moving targets by traditional methods is solved; the pyramid optical flow method is adopted to construct an image pyramid and calculate the optical flow layer by layer on a multi-layer structure, effectively processing the pixel-level motion vectors of large-displacement moving targets and improving the robustness of optical flow calculation; the DBSCAN clustering algorithm is introduced to cluster the motion trajectories, clustering the same target trajectory into one category, excluding the interference of isolated noise trajectories on the separation of the smear region, and improving the accuracy of trajectory analysis; in the deconvolution operation, an edge-preserving constraint is combined, and by introducing an edge detection operator to protect the edge region of the target contour, edge blurring during the deconvolution process is avoided, while better retaining the target details when eliminating the smear, further improving the imaging quality of the infrared thermal imager and enhancing the adaptability to complex motion scenes.

[0092] Embodiment 4:

[0093] Based on the same general inventive concept, the present invention also provides an infrared thermal imager smear elimination system 100, which is characterized by including:

[0094] An image acquisition module 110, which is used to obtain a set of original image frames of the infrared thermal imager in a continuous time series.

[0095] Among them, the image acquisition module further includes: 111 A preprocessing module, which is used to collect several consecutive infrared images and perform non-uniformity correction and Gaussian filtering preprocessing on the original images to remove fixed pattern noise and random noise.

[0096] A 120 motion detection module, which is used to perform inter-frame difference calculation on adjacent frame images and extract candidate regions of moving targets in combination with morphological processing.

[0097] A 130 optical flow calculation module, which is used to calculate the pixel-level motion vector field by the optical flow method within the candidate region and generate a set of motion trajectories of the target.

[0098] 140 Trajectory analysis module, which is used to perform spatio-temporal consistency analysis on the set of motion trajectories, and separate the real target trajectories that meet the motion continuity constraints from the ghost regions that do not meet the constraints.

[0099] 150 Image reconstruction module, which is used to establish a motion blur model based on the real target trajectories and reconstruct the target contour image without ghosting through deconvolution operation.

[0100] Embodiment Five:

[0101] Based on the same general inventive concept, the present invention also provides an electronic device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method for eliminating ghosting of an infrared thermal imager.

[0102] Embodiment Six:

[0103] Based on the same general inventive concept, the present invention also provides a computer-readable storage medium, which stores computer instructions for causing a computer to execute the method for eliminating ghosting of an infrared thermal imager.

[0104] It should be understood that in various embodiments of the present invention, the magnitudes of the serial numbers of the above processes do not mean the order of execution is prior or posterior, and the execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0105] In addition, the terms "system" and "network" are often used interchangeably in this article. The term "and / or" in this article is merely a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this article generally represents an "or" relationship between the associated objects before and after.

[0106] It should be understood that in the embodiments of the present invention, "B corresponding to A" means that B is associated with A, and B can be determined according to A. However, it should also be understood that determining B according to A does not mean determining B only according to A, and B can also be determined according to A and / or other information.

[0107] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been generally described according to their functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0108] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0109] In several embodiments provided in the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there can be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed couplings, direct couplings, or communication connections to each other can be indirect couplings or communication connections through some interfaces, devices, or units, and can also be in electrical, mechanical, or other forms of connection.

[0110] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiments of the present invention.

[0111] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0112] Through the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by hardware, or firmware, or a combination thereof. When implemented using software, the above functions can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium. The computer-readable medium includes computer storage media and communication media, where the communication media includes any medium that facilitates the transmission of a computer program from one place to another. The storage media can be any available medium that can be accessed by a computer. By way of example but not limitation: the computer-readable medium can include RAM, ROM, EEPROM, CD-ROM or other optical disc storage, magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and can be accessed by a computer. In addition, any connection can suitably be a computer-readable medium. For example, if the software is transmitted using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL) or wireless technologies such as infrared, radio and microwave from a website, server or other remote source, then the coaxial cable, fiber optic cable, twisted pair, DSL or wireless technologies such as infrared, wireless and microwave are included in the definition of the medium. As used in the present invention, disk and disc include compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk and Blu-ray disc, where disks generally reproduce data magnetically, while discs reproduce data optically with a laser. The above combinations should also be included within the scope of protection of the computer-readable medium.

[0113] In summary, the above are only the preferred embodiments of the technical solution of the present invention, and are not intended to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. An infrared thermal imager smear elimination method, characterized in that, It includes the following steps: Obtain a set of original image frames of an infrared thermal imager with a continuous time series; Perform inter-frame difference calculation on adjacent frame images, and combine morphological processing to extract candidate regions of moving targets; Within the candidate regions, calculate the pixel-level motion vector field by the optical flow method to generate a set of motion trajectories of the targets; Perform spatio-temporal consistency analysis on the set of motion trajectories to separate real target trajectories that satisfy the motion continuity constraint from ghost regions that do not meet the constraint; Establish a motion blur model based on real target trajectories, and reconstruct a target contour image without ghosting through deconvolution operation.

2. The method for eliminating smear of an infrared thermal imager according to claim 1, characterized in that, The obtaining of a set of original image frames of an infrared thermal imager with a continuous time series includes: Collect several consecutive infrared images, and perform non-uniformity correction and Gaussian filtering preprocessing on the original images to remove fixed pattern noise and random noise.

3. The method for eliminating smear of an infrared thermal imager according to claim 1, characterized in that, The performing of inter-frame difference calculation on adjacent frame images and combining morphological processing to extract candidate regions of moving targets includes: Calculate the difference image between adjacent frames, and perform thresholding on the difference image to obtain a binary image; Perform morphological dilation and erosion operations on the binary image to remove isolated noise points and connect adjacent regions, thereby obtaining candidate regions of moving targets.

4. The method for eliminating smear of an infrared thermal imager according to claim 1, wherein, The calculating of the pixel-level motion vector field by the optical flow method within the candidate regions to generate a set of motion trajectories of the targets includes: Adopt the Lucas-Kanade optical flow algorithm to calculate the motion vectors of pixels within the candidate regions; Filter outliers of the motion vectors, eliminate vectors whose norm exceeds the threshold, and generate an effective motion vector field; Based on the motion vectors of adjacent frames, generate pixel-level motion trajectories through a trajectory association algorithm.

5. The method for eliminating smear of an infrared thermal imager according to claim 1, wherein The performing of spatio-temporal consistency analysis on the set of motion trajectories to separate real target trajectories that satisfy the motion continuity constraint from ghost regions that do not meet the constraint includes: Based on pre-established spatio-temporal consistency constraint conditions, use Kalman filtering or the Hungarian algorithm to perform time series association on the trajectories, thereby calculating the spatio-temporal consistency scores of each trajectory; Based on the calculated spatio-temporal consistency scores, determine whether the target trajectory is a real target trajectory or a ghost region.

6. The method for eliminating smear of an infrared thermal imager according to claim 1, characterized in that, The establishing of a motion blur model based on real target trajectories and reconstructing a target contour image without ghosting through deconvolution operation includes: For real target trajectories, construct a point spread function of motion blur; Adopt Wiener filtering for deconvolution operation, thereby reconstructing a target contour image without ghosting.

7. An infrared thermal imager smear elimination system for implementing the method according to any one of claims 1-6, characterized in that, It includes: An image acquisition module for obtaining a set of original image frames of an infrared thermal imager with a continuous time series; A motion detection module for performing inter-frame difference calculation on adjacent frame images and combining morphological processing to extract candidate regions of moving targets; An optical flow calculation module for calculating the pixel-level motion vector field by the optical flow method within the candidate regions to generate a set of motion trajectories of the targets; A trajectory analysis module for performing spatio-temporal consistency analysis on the set of motion trajectories to separate real target trajectories that satisfy the motion continuity constraint from ghost regions that do not meet the constraint; An image reconstruction module for establishing a motion blur model based on real target trajectories and reconstructing a target contour image without ghosting through deconvolution operation.

8. The infrared thermal imager smear elimination system according to claim 7, wherein the image acquisition module further includes: a preprocessing module, configured to acquire a plurality of consecutive infrared images, and perform non-uniformity correction and Gaussian filtering preprocessing on the original images to remove fixed pattern noise and random noise.

9. An electronic device, characterized in that, The electronic device includes: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and when the instructions are executed by the at least one processor, the at least one processor is enabled to execute the infrared thermal imager smear elimination method according to any one of claims 1-6.

10. A computer-readable storage medium, characterized in that, This computer-readable storage medium stores computer instructions for causing a computer to execute the infrared thermal imager smear elimination method according to any one of claims 1-6.

Images