Infrared and visible image fusion method based on woa-dtntp

Abstract

The application discloses an infrared and visible light image fusion method based on WOA-DTNP and belongs to the technical field of image processing. The infrared and visible light source images are first decomposed into low-frequency subbands and high-frequency subbands through non-subsampled contourlet transform; the low-frequency subbands are further decomposed into structure subbands and texture subbands through an adaptive Euler elastic energy regularization method; then, different fusion rules are designed according to the characteristics of different subbands; a discrete-time neural P system is constructed for the high-frequency subbands, and fusion is realized based on pulse matrix comparison; the low-frequency structure subbands are fused by calculating Laplace energy and determining the weight; the low-frequency texture subbands are adaptively optimized by using a whale optimization algorithm; and finally, the fused subbands are reconstructed into a final image through inverse non-subsampled contourlet transform. The application effectively solves the problems of contrast imbalance and detail loss in the traditional method, and significantly improves the visual quality and information retention of the fused image.

Description

Technical Field

[0001] This invention belongs to the field of image processing technology, specifically relating to a method for fusing infrared and visible light images based on WOA-DTNP. Background Technology

[0002] The fusion of infrared and visible light images has been widely used in military, security, and medical diagnostics fields. Infrared images can capture the thermal radiation information of a target, while visible light images provide rich texture and detail information. By fusing these two types of images, a high-quality image that combines thermal radiation characteristics and detail information can be generated, thereby improving the ability to detect, identify, and analyze targets.

[0003] Currently, infrared and visible light image fusion methods are mainly based on deep learning theory. Although these methods have improved the fusion performance to some extent, they also have problems such as contrast imbalance, loss of detail, and high requirements for device computing performance during the fusion process.

[0004] In contrast, multi-scale analysis methods, such as pyramid decomposition, wavelet transform, and non-subsampled contourlet transform, can more effectively preserve the thermal radiation contrast of infrared images and enhance the texture details of visible light images by processing image information in different frequency bands in a layered manner. Furthermore, they have lower computational complexity, making them more suitable for resource-constrained applications. Therefore, multi-scale analysis has significant advantages in improving the quality of fused images and optimizing computational efficiency. However, traditional fusion methods often suffer from information redundancy and detail loss when processing low-frequency subbands, affecting the fusion effect. Summary of the Invention

[0005] To address the aforementioned problems in the existing technology, this invention provides a method for fusion of infrared and visible light images based on WOA-DTNP. The technical problem to be solved by this invention is achieved through the following technical solution: This invention provides a method for fusion of infrared and visible light images based on WOA-DTNP, comprising: The infrared source image is decomposed into an infrared low-frequency subband. and infrared high-frequency subband The visible light image is decomposed into a low-frequency visible light sub-band. and visible light high-frequency subband ; For the infrared low-frequency subband respectively and the aforementioned visible light low-frequency sub-band Further decomposition yields the infrared low-frequency structural subband. Infrared low-frequency textured subband and visible light low-frequency structure subband Visible light low-frequency texture subband ; By constructing a discrete-time neural P system and fusing the infrared high-frequency subband With the visible light high-frequency sub-band To obtain high-frequency fusion subbands; By calculating the Laplace energy, the infrared low-frequency structural subbands are fused. and visible light low-frequency structure subband To obtain a low-frequency structural fusion subband; The whale optimization algorithm is used to determine the fusion weights, and the visible light low-frequency texture subband is then optimized based on these fusion weights. With the infrared low-frequency texture subband Weighted fusion is used to obtain a low-frequency texture fusion sub-band; The low-frequency structure fusion subband and the low-frequency texture fusion subband are fused to obtain a low-frequency fusion subband. The high-frequency fusion subband and the low-frequency fusion subband are then reconstructed using inverse non-subsampled contour wave transform to obtain a fused image.

[0006] In one embodiment of the present invention, the infrared low-frequency sub-bands are respectively... and the aforementioned visible light low-frequency sub-band Further decomposition yields the infrared low-frequency structural subband. Infrared low-frequency textured subband and visible light low-frequency structure subband Visible light low-frequency texture subband The steps include: Based on the adaptive Euler elastic energy regularization formula, the infrared low-frequency subband is respectively... and the aforementioned visible light low-frequency sub-band Further decomposition yields the infrared low-frequency structural subband. Infrared low-frequency textured subband and visible light low-frequency structure subband Visible light low-frequency texture subband ; The adaptive Euler elastic energy regularization formula is expressed as follows: ; In the formula, Indicates infrared low-frequency structure subband Visible light low-frequency structure subband , Indicates infrared low-frequency texture subband Visible light low-frequency texture subband , This represents the gradient of the infrared low-frequency structural subband and the gradient of the visible light low-frequency structural subband. This is a preset constant used to avoid gradient terms. exist Numerical instability occurs when the value is zero. Indicates data fidelity item The weighting coefficients in the text Represents texture item The regularization parameter in the text, Represents the energy function. Represents the integration field. Represents spatial location variables.

[0007] In one embodiment of the present invention, the infrared high-frequency subband is fused by constructing a discrete-time neural P system. With the visible light high-frequency sub-band The steps for obtaining high-frequency fusion subbands include: The infrared high-frequency sub-bands are sequentially... and the visible light high-frequency sub-band As the image to be processed, the following steps are performed: the edge information of the image to be processed is extracted using the Prewitt operator, and the energy distribution map of the image to be processed is calculated using the local energy function. The edge information and the energy distribution map are used as inputs to construct a discrete-time neural P system of the image to be processed. After iteration, the pulse sum matrix of the image to be processed is obtained. By comparing the infrared high-frequency subbands one by one The pulse sum matrix and the visible light high-frequency subband The magnitude of the elements in the pulse summation matrix is ​​used to obtain the high-frequency fusion subband.

[0008] In one embodiment of the present invention, the infrared high-frequency subbands are compared one by one. The pulse sum matrix and the visible light high-frequency subband The steps for obtaining the high-frequency fused subband by determining the size of the elements in the pulse summation matrix include: Compare the infrared high-frequency subbands one by one The pulse sum matrix and the visible light high-frequency subband The size of the elements in the pulse summation matrix; like , then As the first in the high-frequency fusion subband Line 1 The pixel value of the column pixel; otherwise, the pixel value of the column pixel. As the first in the high-frequency fusion subband Line 1 The pixel value of the column pixel; in, Represents the high-frequency subband of visible light pulse sum matrix The middle is located in the first Line 1 Column pixels, Indicates infrared high-frequency subband pulse sum matrix The middle is located in the first Line 1 Column pixels, , These represent the high-frequency subbands of visible light. With infrared high-frequency subband The Middle Line 1 The pixel value of the column pixel.

[0009] In one embodiment of the present invention, the infrared low-frequency structural subband is fused by calculating the Laplace energy. and visible light low-frequency structure subband The steps for obtaining a low-frequency structural fusion subband include: Calculate the infrared low-frequency structural subband Laplace energy and and the visible light low-frequency structure subband Laplace energy and ; According to the Laplace energy and and the Laplace energy and Determine the visible light low-frequency structural subband. The first initial weight map and the infrared low-frequency structure subband The second initial weight map; Based on the first initial weight map and the second initial weight map, the infrared low-frequency structural sub-band is fused. With the visible light low-frequency structure sub-band To obtain a low-frequency structural fusion subband.

[0010] In one embodiment of the invention, based on the Laplace energy and and the Laplace energy and Determine the visible light low-frequency structural subband. The first initial weight map and the infrared low-frequency structure subband The steps for the second initial weight map include: Comparison of infrared low-frequency structural subbands With visible light low-frequency structure subband The sum of the Laplacian energies of each pixel; when At that time, the visible light low-frequency structure subband medium pixel Corresponding first initial weight Set to 1, infrared low-frequency structure subband medium pixel Corresponding second initial weight Conversely, the visible light low-frequency structure subband medium pixel Corresponding first initial weight Set to 0, infrared low-frequency structure subband medium pixel Corresponding second initial weight ; Traversing the infrared low-frequency structure subband and visible light low-frequency structure subband After processing all pixels, the visible light low-frequency structure subband is obtained. The first initial weight map, and the infrared low-frequency structure subband The second initial weight graph.

[0011] In one embodiment of the present invention, the visible light low-frequency structural subbands are calculated according to the following formulas. and the infrared low-frequency structural subband Laplace energy and: ; ; in,

[0012]

[0013] In the formula, , These represent the low-frequency structural subbands of visible light. and infrared low-frequency structure subband The Laplace operator, , Represented by pixels The row and column offsets within the preset window centered on the target. These represent the length and width parameters of the preset window, respectively.

[0014] In one embodiment of the present invention, the infrared low-frequency structural subband is fused based on the first initial weight map and the second initial weight map. With the visible light low-frequency structure sub-band The steps for obtaining a low-frequency structural fusion subband include: Guided filtering is applied to the first initial weight map and the second initial weight map respectively; The visible light low-frequency structure subband The infrared low-frequency structural subband The low-frequency structure fusion subband is obtained by weighting and summing the subband with the first initial weight map after filtering and the second initial weight map after filtering, respectively.

[0015] In one embodiment of the present invention, the whale optimization algorithm is used to determine the fusion weights, and the visible light low-frequency texture subband is adjusted based on the fusion weights. With the infrared low-frequency texture subband The steps for weighted fusion to obtain low-frequency texture fusion subbands include: Based on the visible light low-frequency texture subband With the infrared low-frequency texture subband The fitness function is constructed using the local variance and edge strength. Based on the fitness function, the whale optimization algorithm is used to determine the visible light low-frequency texture subband. Fusion weights and infrared low-frequency textured subband Fusion weights ; Based on the visible light low-frequency texture subband and its fusion weight The infrared low-frequency texture subband and its fusion weight Weighted fusion is performed to obtain a low-frequency texture fusion subband.

[0016] In one embodiment of the present invention, the fitness function is expressed as:

[0017] In the formula, Represents local variance. Indicates edge strength.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention provides an infrared and visible light image fusion method based on WOA-DTNP. After decomposing the infrared source image and the visible light source image into layers, it further uses adaptive Euler elastic energy regularization to fuse the resulting low-frequency subband (infrared low-frequency subband). and visible light low-frequency subband It is further decomposed into low-frequency structure subbands and low-frequency texture subbands to more finely preserve the structural and texture information of the image, thus solving the problem of information redundancy and detail loss that often exist in existing fusion methods when processing low-frequency subbands.

[0019] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0020] Figure 1This is a flowchart of an infrared and visible light image fusion method based on WOA-DTNP provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of an infrared and visible light image fusion method based on WOA-DTNP provided in an embodiment of the present invention; Figure 3a This is a diagram showing the low-frequency sub-band and its further decomposition results provided in an embodiment of the present invention; Figure 3b This is a diagram showing the low-frequency sub-band and its further decomposition results provided in an embodiment of the present invention; Figure 4a This is an infrared source image provided in an embodiment of the present invention; Figure 4b This is a visible light source image provided in an embodiment of the present invention; Figures 4c-4g It uses different infrared and visible light image fusion methods to fuse them. Figure 4a , 4b The resulting fused image; Figure 5a This is another infrared source image provided in the embodiments of the present invention; Figure 5b This is another visible light source image provided in an embodiment of the present invention; Figures 5c-5g It uses different infrared and visible light image fusion methods to fuse them. Figure 5a , 5b The obtained fused image. Detailed Implementation

[0021] The present invention will be further described in detail below with reference to specific embodiments, but the implementation of the present invention is not limited thereto.

[0022] Figure 1 This is a flowchart of an infrared and visible light image fusion method based on WOA-DTNP provided in an embodiment of the present invention. Figure 1 As shown, this embodiment of the invention provides a method for fusing infrared and visible light images based on WOA-DTNP, including: S1. Decompose the infrared source image into infrared low-frequency subbands. and infrared high-frequency subband The visible light image is decomposed into a low-frequency visible light sub-band. and visible light high-frequency subband .

[0023] In this embodiment, non-subsampled contourlet transform (NSCT) can be used to process the infrared source image. and visible light source images Decompose it.

[0024] S2, respectively targeting the infrared low-frequency subband and visible light low-frequency subband Further decomposition yields the infrared low-frequency structural subband. Infrared low-frequency textured subband and visible light low-frequency structure subband Visible light low-frequency texture subband .

[0025] It should be understood that due to information redundancy in the low-frequency subband, which affects the fusion effect, it is necessary to adjust the infrared low-frequency subband. and visible light low-frequency subband Further decomposition is performed to obtain more refined low-frequency background subbands and low-frequency texture subbands.

[0026] In this embodiment, step S2 includes: Based on the adaptive Euler elastic energy regularization formula, the infrared low-frequency subband is respectively... and visible light low-frequency subband Further decomposition yields the infrared low-frequency structural subband. Infrared low-frequency textured subband and visible light low-frequency structure subband Visible light low-frequency texture subband ; The adaptive Euler elastic energy regularization formula is expressed as follows: ; In the formula, Indicates infrared low-frequency structure subband Visible light low-frequency structure subband , Indicates infrared low-frequency texture subband Visible light low-frequency texture subband , This represents the gradient of the infrared low-frequency structural subband and the gradient of the visible light low-frequency structural subband. This is a preset constant used to avoid gradient terms. exist Numerical instability occurs when the value is zero. Indicates data fidelity item The weighting coefficients in the text Represents texture item The regularization parameter in the text, Represents the energy function. Represents the integration field. Represents spatial location variables.

[0027] Specifically, for the infrared low-frequency subband Its adaptive Euler elastic energy regularization formula can be expressed as: ; For the low-frequency subband of visible light Its adaptive Euler elastic energy regularization formula can be expressed as: ; In the formula, Indicates infrared low-frequency structure subband gradient, Represents the low-frequency texture subband of visible light The gradient. Figure 2 This is a further decomposition result diagram of the infrared low-frequency sub-band and the visible light low-frequency sub-band provided in the embodiments of the present invention. Figure 2 It can be seen that after adaptive Euler elasticity regularization decomposition, the structure layer of the low-frequency subband exhibits more significant smoothness characteristics than the original low-frequency subband, while the texture layer can effectively extract the texture information masked by low-frequency energy.

[0028] S3. By constructing a discrete-time neural P system and fusing infrared high-frequency subbands. High-frequency subband of visible light To obtain high-frequency fusion subbands.

[0029] Specifically, step S3 includes: S1, sequentially convert the infrared high-frequency sub-band and visible light high-frequency subband As the image to be processed, the following steps are performed: the edge information of the image to be processed is extracted using the Prewitt operator, and the energy distribution map of the image to be processed is calculated using the local energy function. The edge information and the energy distribution map are used as inputs to construct a discrete-time neural P system of the image to be processed. After iteration, the pulse sum matrix of the image to be processed is obtained. The above process can be represented as follows:

[0030] in, Therefore A square window centered on the center. It is the convolution kernel of the Prewitt compass operator.

[0031]

[0032] Starting from the initial state, the two DTNP systems continue to execute until the maximum number of iterations is reached. S32, By comparing infrared high-frequency subbands one by one The pulse sum matrix and the visible light high-frequency subband The magnitude of the elements in the pulse summation matrix is ​​used to obtain the high-frequency fusion subband.

[0033] In step S32, the infrared high-frequency subbands are compared one by one. The pulse sum matrix and the visible light high-frequency subband The size of the elements in the pulse sum matrix; if , then As the first in the high-frequency fusion subband Line 1 The pixel value of the column pixel; otherwise, the pixel value of the column pixel. As the first in the high-frequency fusion subband Line 1 The pixel values ​​of the column pixels; where, Represents the high-frequency subband of visible light pulse sum matrix The middle is located in the first Line 1 Column pixels, Indicates infrared high-frequency subband pulse sum matrix The middle is located in the first Line 1 Column pixels, , These represent the high-frequency subbands of visible light. With infrared high-frequency subband The Middle Line 1 The pixel value of the column pixel.

[0034] S4. By calculating the Laplace energy, fuse the infrared low-frequency structural subbands. and visible light low-frequency structure subband To obtain a low-frequency structural fusion subband.

[0035] For example, step S4 includes: S41. Calculate the infrared low-frequency structural subband. Laplace energy and and visible light low-frequency structure subband Laplace energy and .

[0036] Calculate the visible light low-frequency structure subbands according to the following formulas. and infrared low-frequency structure subband Laplace energy and: ; ; in,

[0037]

[0038] In the formula, , These represent the low-frequency structural subbands of visible light. and infrared low-frequency structure subband The Laplace operator, , Represented by pixels The row and column offsets within the preset window centered on the target. These represent the length and width parameters of the preset window, respectively.

[0039] S42, based on Laplace energy and and Laplace energy and Determine the low-frequency structure subband of visible light First initial weight map and infrared low-frequency structure subband The second initial weight graph.

[0040] Specifically, comparing infrared low-frequency structural subbands With visible light low-frequency structure subband The sum of the Laplace energy of each pixel; when At that time, the visible light low-frequency structure subband medium pixel Corresponding first initial weight Set to 1, infrared low-frequency structure subband medium pixel Corresponding second initial weight Conversely, the visible light low-frequency structure subband medium pixel Corresponding first initial weight Set to 0, infrared low-frequency structure subband medium pixel Corresponding second initial weight By traversing the infrared low-frequency structural subband and visible light low-frequency structure subband After processing all pixels, the visible light low-frequency structure subband is obtained. The first initial weighting map, and the infrared low-frequency structure subband The second initial weight graph.

[0041] S43. Based on the first initial weight map and the second initial weight map, fuse the infrared low-frequency structural subband. With visible light low-frequency structure subband To obtain a low-frequency structural fusion subband.

[0042] Specifically, firstly, guided filtering is applied to the first and second initial weight maps respectively, and then the visible light low-frequency structure sub-band is... Infrared low-frequency structure subband The low-frequency structure fusion subband is obtained by weighting and summing the subband with the first initial weight map after filtering and the second initial weight map after filtering, respectively.

[0043] S5. Determine the fusion weights using the whale optimization algorithm, and then refine the visible light low-frequency texture subbands based on these fusion weights. With infrared low-frequency textured subband Weighted fusion is used to obtain a low-frequency texture fusion subband.

[0044] In this embodiment, step S5 includes: S51, Visible Light Low-Frequency Texture Subband With infrared low-frequency textured subband The fitness function is constructed using the local variance and edge strength:

[0045] In the formula, Represents local variance. Indicates edge strength.

[0046] S52. Based on the fitness function, the whale optimization algorithm is used to determine the low-frequency texture subband of visible light. Fusion weights and infrared low-frequency textured subband Fusion weights ; S53, Visible Light Low-Frequency Texture Subband and its fusion weight Infrared low-frequency textured subband and its fusion weight Weighted fusion is performed to obtain a low-frequency texture fusion subband.

[0047] S6. Fuse the low-frequency structure fusion subband with the low-frequency texture fusion subband to obtain the low-frequency fusion subband. Further, reconstruct the fused image by combining the high-frequency fusion subband with the low-frequency fusion subband through inverse non-subsampled contour wave transform.

[0048] The following simulation experiment demonstrates the infrared and visible light image fusion method based on WOA-DTNP provided by this invention.

[0049] It should be noted that the fusion evaluation index used in this embodiment is: Q ABThe quality of the fused image is evaluated from the perspectives of edge information preservation, mutual information, feature mutual information, sum of difference correlation, and structural similarity, respectively. All five evaluation indicators are of the type where the larger the value, the better. Existing fusion methods selected include BGR, VSM_WLS, U2Fusion, and DDcGAN, which are used to compare with the infrared and visible light image fusion method provided in this invention.

[0050] (1) Subjective evaluation Using the TNO infrared and visible light image dataset as the research object, Figure 4a This is an infrared source image provided in an embodiment of the present invention. Figure 4b This is a visible light source image provided in an embodiment of the present invention. Figures 4c-4g It uses different infrared and visible light image fusion methods to fuse them. Figure 4a , 4b The obtained fused image, where, Figure 4c , 4d Images 4e, 4f, and 4g respectively employ BGR, VSM_WLS, U2Fusion, DDcGAN, and the infrared and visible light image fusion method provided in this invention. It can be seen that the BGR method has significant shortcomings in overall visual effect, resulting in a noticeable halo effect in the fused image, such as... Figure 4c Halo artifacts around streetlights and pine trees; the DDcGAN method performs poorly in terms of the salience of thermal targets, such as... Figure 4f In the upper left red box, the smoke area is difficult to discern, and the edges of the figures are blurred. Regarding detail preservation, the fusion results of the VSM_WLS and U2Fusion methods show blurred pine leaf outlines and insufficient contrast in the gaps between leaves and the boundaries of tree hollows. In contrast, the fusion method proposed in this invention exhibits significant advantages in several aspects, such as... Figure 4g As shown, it displays a clear leaf outline and a distinct contrast between light and shadow.

[0051] Furthermore, using the FLIR infrared and visible light image dataset as the research object, Figure 5a This is another infrared source image provided in the embodiments of the present invention. Figure 5b This is another visible light source image provided in an embodiment of the present invention. Figures 5c-5g It uses different infrared and visible light image fusion methods to fuse them. Figure 5a , 5b The obtained fused image. It can be seen that... Figure 5e In the U2Fusion method, insufficient detail processing resulted in excessively dark areas in the streetlight region. The BGR and DDcGAN methods performed poorly in preserving thermal targets. Figure 5c The figure at the farthest point in the red box shown is almost invisible. Figure 5fThe red box shows artifacts such as black dots on the portrait; regarding detail preservation, such as... Figure 5d As shown, the fused image obtained using the VSM_WLS method suffers from severe blurring between leaves and the background. (See [link]). Figure 5g The fused image obtained using the method provided by this invention completely preserves the outline features of the leaves.

[0052] (2) Objective evaluation Table 1 Figure 4a , 4b Fusion image evaluation metrics

[0053] Table 2 Figure 5a , 5b Fusion image evaluation metrics

[0054] As shown in Table 1, the method provided by this invention significantly outperforms existing methods in most evaluation metrics, especially in information fusion (MI, FMI, SCD) and structural consistency (SSIM), making it the optimal solution in terms of overall performance; the Q of this invention AB Although slightly lower than BGR and VSM_WLS, the difference is small, and other indicators are significantly superior, indicating that the present invention achieves a better balance between edge preservation and overall performance.

[0055] As can be seen from the above embodiments, the beneficial effects of the present invention are as follows: This invention provides an infrared and visible light image fusion method based on WOA-DTNP. After decomposing the infrared source image and the visible light source image into layers, it further uses adaptive Euler elastic energy regularization to fuse the resulting low-frequency subband (infrared low-frequency subband). and visible light low-frequency subband It is further decomposed into low-frequency structure subbands and low-frequency texture subbands to more finely preserve the structural and texture information of the image, thus solving the problem of information redundancy and detail loss that often exist in existing fusion methods when processing low-frequency subbands.

[0056] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0057] The use of terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples" indicates that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. In addition, those skilled in the art can combine and integrate the different embodiments or examples described in this specification.

[0058] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A method for fusing infrared and visible light images based on WOA-DTNP, characterized in that, include: The infrared source image is decomposed into an infrared low-frequency subband. and infrared high-frequency subband The visible light image is decomposed into a low-frequency visible light sub-band. and visible light high-frequency subband ; For the infrared low-frequency subband respectively and the aforementioned visible light low-frequency sub-band Further decomposition yields the infrared low-frequency structural subband. Infrared low-frequency textured subband and visible light low-frequency structure subband Visible light low-frequency texture subband ; By constructing a discrete-time neural P system and fusing the infrared high-frequency subband With the visible light high-frequency sub-band To obtain high-frequency fusion subbands; By calculating the Laplace energy, the infrared low-frequency structural subbands are fused. and visible light low-frequency structure subband To obtain low-frequency structural fusion subbands; The whale optimization algorithm is used to determine the fusion weights, and the visible light low-frequency texture subband is then optimized based on these fusion weights. With the infrared low-frequency texture subband Weighted fusion is used to obtain a low-frequency texture fusion sub-band; The low-frequency structure fusion subband and the low-frequency texture fusion subband are fused to obtain a low-frequency fusion subband. The high-frequency fusion subband and the low-frequency fusion subband are then reconstructed using inverse non-subsampled contour wave transform to obtain a fused image.

2. The infrared and visible light image fusion method based on WOA-DTNP according to claim 1, characterized in that, For the infrared low-frequency subband respectively and the aforementioned visible light low-frequency sub-band Further decomposition yields the infrared low-frequency structural subband. Infrared low-frequency textured subband and visible light low-frequency structure subband Visible light low-frequency texture subband The steps include: Based on the adaptive Euler elastic energy regularization formula, the infrared low-frequency subband is respectively... and the aforementioned visible light low-frequency sub-band Further decomposition yields the infrared low-frequency structural subband. Infrared low-frequency textured subband and visible light low-frequency structure subband Visible light low-frequency texture subband ; The adaptive Euler elastic energy regularization formula is expressed as follows: ； In the formula, Indicates infrared low-frequency structure subband Visible light low-frequency structure subband , Indicates infrared low-frequency texture subband Visible light low-frequency texture subband , This represents the gradient of the infrared low-frequency structural subband and the gradient of the visible light low-frequency structural subband. This is a preset constant used to avoid gradient terms. exist Numerical instability occurs when the value is zero. Indicates data fidelity item The weighting coefficients in the text Represents texture item The regularization parameter in the text, Represents the energy function. Represents the integration field. Represents spatial location variables.

3. The infrared and visible light image fusion method based on WOA-DTNP according to claim 1, characterized in that, By constructing a discrete-time neural P system and fusing the infrared high-frequency subband With the visible light high-frequency sub-band The steps for obtaining high-frequency fusion subbands include: The infrared high-frequency sub-bands are sequentially... and the visible light high-frequency sub-band As the image to be processed, the following steps are performed: the edge information of the image to be processed is extracted using the Prewitt operator, and the energy distribution map of the image to be processed is calculated using the local energy function. The edge information and the energy distribution map are used as inputs to construct a discrete-time neural P system of the image to be processed. After iteration, the pulse sum matrix of the image to be processed is obtained. By comparing the infrared high-frequency subbands one by one The pulse sum matrix and the visible light high-frequency subband The magnitude of the elements in the pulse summation matrix is ​​used to obtain the high-frequency fusion subband.

4. The infrared and visible light image fusion method based on WOA-DTNP according to claim 3, characterized in that, By comparing the infrared high-frequency subbands one by one The pulse sum matrix and the visible light high-frequency subband The steps for obtaining the high-frequency fused subband by determining the size of the elements in the pulse summation matrix include: Compare the infrared high-frequency subbands one by one The pulse sum matrix and the visible light high-frequency subband The size of the elements in the pulse summation matrix; like , then As the first in the high-frequency fusion subband Line 1 The pixel value of the column pixel; otherwise, the pixel value of the column pixel. As the first in the high-frequency fusion subband Line 1 The pixel value of the column pixel; in, Represents the high-frequency subband of visible light pulse sum matrix Located in the middle Line 1 Column pixels, Indicates infrared high-frequency subband pulse sum matrix The middle is located in the first Line 1 Column pixels, , These represent the high-frequency subbands of visible light. With infrared high-frequency subband The Middle Line 1 The pixel value of the column pixel.

5. The infrared and visible light image fusion method based on WOA-DTNP according to claim 1, characterized in that, By calculating the Laplace energy, the infrared low-frequency structural subbands are fused. and visible light low-frequency structure subband The steps for obtaining a low-frequency structural fusion subband include: Calculate the infrared low-frequency structural subband Laplace energy and and the visible light low-frequency structure subband Laplace energy and ; According to the Laplace energy and and the Laplace energy and Determine the visible light low-frequency structural subband. The first initial weight map and the infrared low-frequency structure subband The second initial weight map; Based on the first initial weight map and the second initial weight map, the infrared low-frequency structural sub-band is fused. With the visible light low-frequency structure sub-band To obtain a low-frequency structural fusion subband.

6. The infrared and visible light image fusion method based on WOA-DTNP according to claim 5, characterized in that, According to the Laplace energy and and the Laplace energy and Determine the visible light low-frequency structural subband. The first initial weight map and the infrared low-frequency structure subband The steps for the second initial weight map include: Comparison of infrared low-frequency structural subbands With visible light low-frequency structure subband The sum of the Laplacian energies of each pixel; when At that time, the visible light low-frequency structure subband medium pixel Corresponding first initial weight Set to 1, infrared low-frequency structure subband medium pixel Corresponding second initial weight Conversely, the visible light low-frequency structure subband medium pixel Corresponding first initial weight Set to 0, infrared low-frequency structure subband medium pixel Corresponding second initial weight ; Traversing the infrared low-frequency structure subband and visible light low-frequency structure subband After processing all pixels, the visible light low-frequency structure subband is obtained. The first initial weight map, and the infrared low-frequency structure subband The second initial weight graph.

7. The infrared and visible light image fusion method based on WOA-DTNP according to claim 6, characterized in that, The visible light low-frequency structural subbands are calculated according to the following formulas respectively. and the infrared low-frequency structural subband Laplace energy and: ； ； in, In the formula, , These represent the low-frequency structural subbands of visible light. and infrared low-frequency structure subband The Laplace operator, , Represented by pixels The row and column offsets within the preset window centered on the target. These represent the length and width parameters of the preset window, respectively.

8. The infrared and visible light image fusion method based on WOA-DTNP according to claim 7, characterized in that, Based on the first initial weight map and the second initial weight map, the infrared low-frequency structural sub-band is fused. With the visible light low-frequency structure sub-band The steps for obtaining a low-frequency structural fusion subband include: Guided filtering is applied to the first initial weight map and the second initial weight map respectively; The visible light low-frequency structure subband The infrared low-frequency structural subband The low-frequency structure fusion subband is obtained by weighting and summing the subband with the first initial weight map after filtering and the second initial weight map after filtering, respectively.

9. The infrared and visible light image fusion method based on WOA-DTNP according to claim 1, characterized in that, The whale optimization algorithm is used to determine the fusion weights, and the visible light low-frequency texture subband is then optimized based on these fusion weights. With the infrared low-frequency texture subband The steps for weighted fusion to obtain low-frequency texture fusion subbands include: Based on the visible light low-frequency texture subband With the infrared low-frequency texture subband The fitness function is constructed using the local variance and edge strength. Based on the fitness function, the whale optimization algorithm is used to determine the visible light low-frequency texture subband. Fusion weights and infrared low-frequency textured subband Fusion weights ; Based on the visible light low-frequency texture subband and its fusion weight The infrared low-frequency texture subband and its fusion weight Weighted fusion is performed to obtain a low-frequency texture fusion subband.

10. The infrared and visible light image fusion method based on WOA-DTNP according to claim 9, characterized in that, The fitness function is expressed as follows: In the formula, Represents local variance. Indicates edge strength.

Images