An infrared image and visible light image fusion method

Abstract

The application discloses an infrared image and visible light image fusion method, and relates to the technical field of image fusion, which comprises the following steps: calculating the abnormality degree of the infrared image gray and the abnormality degree of the visible light image gray; calculating the abnormality degree of the infrared image color and the abnormality degree of the visible light image color; acquiring the abnormality degree of the infrared image gray, the abnormality degree of the visible light image gray, the abnormality degree of the infrared image color and the abnormality degree of the visible light image color, calculating the image gray weight and the color weight, and further calculating the image fusion weight; detecting the feature points in the infrared image and the visible light image, aligning the two images in the same space reference system, and calculating the fusion parameters of each pixel point after alignment. In the fusion process, the detail information in the infrared image and the visible light image is reserved, the respective shortcomings can be made up, and the information complementation is realized.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of image fusion, and particularly relates to an infrared image and visible light image fusion method. BACKGROUND

[0002] With the rapid development of computer science and technology, optics, photography and other fields, image fusion technology has been widely concerned and applied. In particular, in the late 1980s, image fusion technology was widely used in the processing of multispectral remote sensing images, and remarkable fusion effects were achieved. For example, in 1985, researchers successfully fused the multispectral remote sensing images obtained by the TM thematic imager of the Landsat satellite with the high-resolution images of the SPOT satellite, generating multispectral images with higher quality. After entering the 1990s, with the launch of multiple remote sensing radar satellites and the in-depth research of various countries on image fusion, the application range of image fusion technology gradually expanded to visible light images, infrared images and medical images and other fields. In the military field, image fusion technology is widely used in battlefield cooperative guidance, target reconnaissance, weapon thermal collimator, ship navigation and other scenes, significantly improving the accuracy and efficiency of military operations. At the same time, in the medical field, image fusion technology is also used for auxiliary diagnosis and treatment, and through the fusion of image information of different modalities, the disease detection rate and positioning accuracy are improved. As an important branch of image fusion, infrared image and visible light image fusion technology has unique technical advantages and application value. Infrared images mainly capture the infrared radiation emitted by objects to obtain image information, so they can provide better target detection and recognition ability at night or in bad weather conditions. However, the resolution of infrared images is usually low, and the color information is limited. In contrast, visible light images can provide rich color and texture information, making the scene more realistic and vivid. However, the imaging quality of visible light images will decrease significantly at night or in bad weather conditions.

[0003] In a Chinese invention application with the application publication number CN118657671A, an infrared image and visible light image fusion method, device and storage medium are disclosed, which comprises acquiring real-time images of infrared images and real-time images of visible light images; inputting the acquired real-time images into a pre-trained dual-modal feature extraction model for feature extraction to obtain feature maps of infrared images and feature maps of visible light images; inputting the real-time images and feature maps of visible light images into a pre-trained semantic information model for semantic information fusion to obtain feature maps of visible light images containing semantic information; and fusing the feature maps of visible light images containing semantic information and the feature maps of infrared images to obtain a fusion image.

[0004] In the above invention application, the feature map of the visible light image containing semantic information and the feature map of the infrared image are fused to obtain a fused image, but the fusion of the images is based on semantics, the semantic information in many images is not clear, there may be ambiguity, ambiguity or polysemy, and the semantic information often has strong subjective color, different observers may have different semantic understanding for the same image, and such subjectivity may cause inconsistency and unpredictability of the fusion result in the image fusion process.

[0005] To this end, the present application provides an infrared image and visible light image fusion method. SUMMARY

[0006] (I) Technical problems solved

[0007] In view of the deficiencies in the prior art, the present application provides an infrared image and visible light image fusion method, the present application obtains the infrared image gray abnormality degree Yht, the visible light image gray abnormality degree Ykt, the infrared image color abnormality degree Sht and the visible light image color abnormality degree Skt, calculates the image gray weight x1, x2 and the color weight y1, y2, and further calculates the image fusion weight Z1 and Z2, detects the feature points in the infrared image and the visible light image, aligns the matching in the two images in the same spatial reference system, and calculates the fusion parameters of each pixel point after alignment, which helps to retain the detail information in the infrared image and the visible light image in the fusion process, can make up for the respective shortcomings, realizes the complementation of information, and has important significance for improving the precision of image fusion, retaining detail information, enhancing robustness and optimizing visual effect, thereby solving the technical problems recorded in the background art.

[0008] (II) Technical solutions

[0009] To achieve the above object, the present application is realized by the following technical solutions: an infrared image and visible light image fusion method, comprising the following steps:

[0010] The infrared image and the visible light image are imported into OpenCV, the mean and variance of the gray images of the infrared image and the visible light image are calculated, the infrared image gray abnormality degree Yht and the visible light image gray abnormality degree Ykt are calculated according to the infrared image gray mean Hj, the infrared image gray variance Hf, the visible light image gray mean Kj and the visible light image gray variance Kf;

[0011] The RGB values of each pixel point in the infrared image and the visible light image are extracted, the RGB variances of the pixel points in the infrared image and the visible light image are calculated, the infrared R variance Rh, the infrared G variance Gh, the infrared B variance Bh, the visible light R variance Rk, the visible light G variance Gk and the visible light B variance Bk are calculated according to the infrared R variance Rh, the infrared G variance Gh, the infrared B variance Bh, the visible light R variance Rk, the visible light G variance Gk and the visible light B variance Bk, the color abnormality degree Sht of the infrared image and the color abnormality degree Skt of the visible light image are calculated;

[0012] The infrared image gray abnormality degree Yht, the visible light image gray abnormality degree Ykt, the infrared image color abnormality degree Sht and the visible light image color abnormality degree Skt are obtained, the image gray weight x1, x2 and the color weight y1, y2 are calculated, and the image fusion weight Z1 and Z2 are further calculated, the feature points are detected in the infrared image and the visible light image, the two images are matched and aligned in the same spatial reference system, and the fusion parameters of each pixel point after alignment are calculated.

[0013] Further, the infrared image and the visible light image are imported into the OpenCV and NumPy libraries, the cv2.cvtColor function of OpenCV is used to convert the image into a gray image, and the np.mean and np.var functions of NumPy are used to calculate the mean and variance of the gray image of the infrared image and the visible light image, which are denoted as the infrared image gray mean Hj, the infrared image gray variance Hf, the visible light image gray mean Kj and the visible light image gray variance Kf.

[0014] Further, the infrared image gray mean Hj, the infrared image gray variance Hf, the visible light image gray mean Kj and the visible light image gray variance Kf are obtained, and the infrared image gray abnormality degree Yht and the visible light image gray abnormality degree Ykt are calculated:

[0015]

[0016] wherein, is the expected normal gray mean of the infrared image, is the expected normal gray mean of the visible light image, and is the mean of historical data, is a small constant, which is used to avoid the case that the denominator is zero, and ϵ can be 1e−5 or 1e−10.

[0017] Further, the RGB values of each pixel point in the infrared image and the visible light image are extracted using OpenCV, and the RGB variances of the pixel points in the infrared image and the visible light image are calculated, which are denoted as the infrared R variance Rh, the infrared G variance Gh, the infrared B variance Bh, the visible light R variance Rk, the visible light G variance Gk and the visible light B variance Bk.

[0018] In computer graphics, the RGB color model is one of the most commonly used color representation methods. By adjusting the values of the three components R, G, and B, almost all colors visible to the human eye can be mixed. The value range of each component is usually 0 to 255 (8-bit color depth), or higher, such as 0 to 1023 (10-bit color depth), to provide richer color levels. R (Red) represents the intensity of the red component. G (Green): represents the intensity of the green component.

[0019] B (Blue): represents the intensity of the blue component.

[0020] Further, the infrared R variance Rh, the infrared G variance Gh, the infrared B variance Bh, the visible light R variance Rk, the visible light G variance Gk, and the visible light B variance Bk are obtained, and the infrared image color abnormality degree Sht and the visible light image color abnormality degree Skt are calculated:

[0021]

[0022] wherein, is the average value of the variances of the R, G, and B channels of the historical infrared images, used to normalize the variance values of each channel, reflects the deviation of each channel variance relative to the average variance, takes into account the differences between the variances of each channel to capture the unevenness of the color distribution.

[0023] Further, the infrared image gray abnormality degree Yht and the visible light image gray abnormality degree Ykt are obtained, and the image gray weight x1 and x2 are calculated:

[0024]

[0025] wherein, is the average value of all historical infrared image gray abnormality degrees Yht, is the average value of all historical visible light image gray abnormality degrees Ykt.

[0026] Further, the infrared image color abnormality degree Sht and the visible light image color abnormality degree Skt are obtained, and the image color weight y1 and y2 are calculated:

[0027]

[0028] wherein, is the average value of all historical infrared image color abnormality degrees Sht, is the average value of all historical visible light image color abnormality degrees Skt.

[0029] Furthermore, the image grayscale weights x1, x2 and color weights y1, y2 are obtained, and the image fusion weights Z1 and Z2 are calculated:

[0030]

[0031] The calculation formulas of the corresponding image fusion weights Z1 and Z2 are as above.

[0032] Furthermore, feature points are detected in the infrared image and the visible light image, descriptors are calculated for each detected feature point, the feature descriptors in the two images are matched, the geometric transformation relationship between the two images is estimated based on the matched feature point pairs, and one image is transformed according to the obtained geometric transformation parameters so that the two images are aligned in the same spatial reference system.

[0033] Furthermore, the fusion parameters of each pixel of the infrared image are calculated based on the image fusion weight Z1, and the fusion parameters of each pixel of the visible light image are calculated based on the image fusion weight Z2. Finally, the fusion parameters of the infrared image and the visible light image of each pixel after alignment are summed to obtain the fusion parameters of each pixel after alignment.

[0034] For example, if the R value of an infrared image pixel is 100 and the R value of a visible light image pixel is 200, the infrared image fusion parameter is 100Z1, the visible light image fusion parameter is 200Z2, and the alignment fusion parameter is 100Z1+200Z2.

[0035] (3) Beneficial effects

[0036] The present invention provides a method for fusing infrared images and visible light images, which has the following beneficial effects:

[0037] 1. Import the infrared image and the visible light image into OpenCV, calculate the mean and variance of the grayscale images of the infrared image and the visible light image, and calculate the infrared image grayscale anomaly degree Yht and the visible light image grayscale anomaly degree Ykt based on the infrared image grayscale mean Hj, the infrared image grayscale variance Hf, the visible light image grayscale mean Kj and the visible light image grayscale variance Kf. This can objectively evaluate the quality of the infrared image and the visible light image, providing a basis for subsequent image processing and analysis.

[0038] 2. Extract the RGB value of each pixel in the infrared image and the visible light image, calculate the RGB variance of the pixel in the infrared image and the visible light image, and calculate the color anomaly degree Sht of the infrared image and the color anomaly degree Skt of the visible light image based on the infrared R variance Rh, infrared G variance Gh, infrared B variance Bh, visible light R variance Rk, visible light G variance Gk and visible light B variance Bk. This can evaluate the color consistency of the two images and provide an important reference for the formulation of the fusion strategy.

[0039] 3、Obtain the infrared image gray abnormal degree Yht, the visible light image gray abnormal degree Ykt, the infrared image color abnormal degree Sht and the visible light image color abnormal degree Skt, calculate the image gray weight x1, x2 and the color weight y1, y2, and further calculate the image fusion weight Z1 and Z2, detect the feature points in the infrared image and the visible light image, align the two images in the same spatial reference system, calculate the fusion parameters of each pixel point after alignment, which is helpful to retain the detail information in the infrared image and the visible light image in the fusion process, can make up for the respective shortcomings, realize the complement of information, and has important significance for improving the precision of image fusion, retaining detail information, enhancing robustness and optimizing visual effect. BRIEF DESCRIPTION OF DRAWINGS

[0040] Figure 1 It is a flowchart of the infrared image and visible light image fusion method. DETAILED DESCRIPTION

[0041] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0042] Please refer to Figure 1 The present application provides an infrared image and visible light image fusion method, comprising the following steps:

[0043] Step one, import the infrared image and the visible light image into OpenCV, calculate the mean and variance of the gray images of the infrared image and the visible light image, and calculate the infrared image gray abnormal degree Yht and the visible light image gray abnormal degree Ykt according to the infrared image gray mean Hj, the infrared image gray variance Hf, the visible light image gray mean Kj and the visible light image gray variance Kf.

[0044] The step one includes the following contents:

[0045] Step 101, import the infrared image and the visible light image into OpenCV (opencv-python) and NumPy library, convert the image into a gray image using the cv2.cvtColor function of OpenCV, and calculate the mean and variance of the gray images of the infrared image and the visible light image using the np.mean and np.var functions of NumPy, respectively, and mark them as the infrared image gray mean Hj, the infrared image gray variance Hf, the visible light image gray mean Kj and the visible light image gray variance Kf.

[0046] Step 102, obtain the infrared image gray mean Hj, infrared image gray variance Hf, visible light image gray mean Kj and visible light image gray variance Kf, calculate the infrared image gray abnormal degree Yht and the visible light image gray abnormal degree Ykt:

[0047]

[0048] Wherein, is the expected normal gray mean of the infrared image, is the expected normal gray mean of the visible light image, is the mean of the historical data, is a small constant, used to avoid the case of zero denominator, in practical application, ϵ can be taken as 1e−5 or 1e−10.

[0049] In use, combined with the contents in steps 101 and 102:

[0050] The infrared image and the visible light image are imported into OpenCV, the mean and variance of the gray image of the infrared image and the visible light image are calculated, the infrared image gray abnormal degree Yht and the visible light image gray abnormal degree Ykt are calculated according to the infrared image gray mean Hj, the infrared image gray variance Hf, the visible light image gray mean Kj and the visible light image gray variance Kf, and the quality of the infrared image and the visible light image can be objectively evaluated, providing a basis for subsequent image processing and analysis.

[0051] Step two, extract the RGB value of each pixel point in the infrared image and the visible light image, calculate the RGB variance of the pixel points in the infrared image and the visible light image, calculate the infrared image color abnormal degree Sht and the visible light image color abnormal degree Skt according to the infrared R variance Rh, the infrared G variance Gh, the infrared B variance Bh, the visible light R variance Rk, the visible light G variance Gk and the visible light B variance Bk.

[0052] The step two includes the following contents:

[0053] Step 201, use OpenCV to extract the RGB value of each pixel point in the infrared image and the visible light image, calculate the RGB variance of the pixel points in the infrared image and the visible light image, and record the infrared R variance Rh, the infrared G variance Gh, the infrared B variance Bh, the visible light R variance Rk, the visible light G variance Gk and the visible light B variance Bk.

[0054] In computer graphics, RGB color model is one of the most commonly used color representation methods. By adjusting the values of R, G, B three components, almost all colors visible to human eyes can be mixed. The value range of each component is usually 0 to 255 (8-bit color depth), or higher, such as 0 to 1023 (10-bit color depth), to provide more rich color levels. R (Red) represents the intensity of the red component. G (Green): represents the intensity of the green component.

[0055] B (Blue): represents the intensity of the blue component.

[0056] Step 202, obtain infrared R variance Rh, infrared G variance Gh, infrared B variance Bh, visible light R variance Rk, visible light G variance Gk and visible light B variance Bk, calculate the color abnormality degree Sht of infrared image and the color abnormality degree Skt of visible light image:

[0057]

[0058] Wherein, is the average value of the variance of the R, G, B three channels of the historical infrared image, which is used to standardize the variance value of each channel, reflects the deviation of each channel variance relative to the average variance, takes into account the difference between the variances of each channel to capture the imbalance of color distribution.

[0059] In use, in combination with the contents in steps 201 and 202:

[0060] The RGB values of each pixel point in the infrared image and the visible light image are extracted, the RGB variances of the pixel points in the infrared image and the visible light image are calculated, and the color abnormality degree Sht of the infrared image and the color abnormality degree Skt of the visible light image are calculated according to the infrared R variance Rh, the infrared G variance Gh, the infrared B variance Bh, the visible light R variance Rk, the visible light G variance Gk and the visible light B variance Bk. The performance of the two images in color consistency can be evaluated, which provides an important reference for the development of fusion strategy.

[0061] Step three, obtain the infrared image gray abnormality degree Yht, the visible light image gray abnormality degree Ykt, the infrared image color abnormality degree Sht and the visible light image color abnormality degree Skt, calculate the image gray weight x1, x2 and the color weight y1, y2, and further calculate the image fusion weight Z1 and Z2. In the infrared image and the visible light image, the feature points are detected, the two images are matched and aligned in the same spatial reference system, and the fusion parameters of each pixel point after alignment are calculated.

[0062] The step three includes the following contents:

[0063] Step 301, obtain the infrared image gray abnormality degree Yht and the visible light image gray abnormality degree Ykt, and calculate the image gray weight x1 and x2:

[0064]

[0065] wherein, is the mean value of all historical infrared image gray abnormality degrees Yht, is the mean value of all historical visible light image gray abnormality degrees Ykt.

[0066] Step 302, obtain the infrared image color abnormality degree Sht and the visible light image color abnormality degree Skt, and calculate the image color weight y1 and y2:

[0067]

[0068] wherein, is the mean value of all historical infrared image color abnormality degrees Sht, is the mean value of all historical visible light image color abnormality degrees Skt.

[0069] Step 303, obtain the image gray weight x1, x2 and the color weight y1, y2, and calculate the image fusion weight Z1 and Z2:

[0070]

[0071] The calculation formula of the corresponding image fusion weight Z1 and Z2 is as above.

[0072] Step 304, detect feature points in the infrared image and the visible light image, common feature detection operators include Harris, FAST, SIFT, SURF, ORB, etc., calculate the descriptor for each detected feature point, match the feature descriptors in the two images, common matching strategies include nearest neighbor matching, ratio test, etc., estimate the geometric transformation relationship between the two images based on the matched feature points, which is usually an affine transformation or a projection transformation, and according to the obtained geometric transformation parameters, transform one image to make the two images aligned in the same spatial reference system.

[0073] Step 305, calculate the fusion parameter of each pixel point of the infrared image according to the image fusion weight Z1, calculate the fusion parameter of each pixel point of the visible light image according to the image fusion weight Z2, and finally sum the infrared image fusion parameter and the visible light image fusion parameter of each pixel point after alignment to obtain the fusion parameter of each pixel point after alignment.

[0074] For example, the R value of the infrared image pixel point is 100, the R value of the visible light image pixel point is 200, the infrared image fusion parameter is 100Z1, the visible light image fusion parameter is 200Z2, and the alignment fusion parameter is 100Z1+200Z2.

[0075] In use, the content in steps 301 to 305 is combined:

[0076] The infrared image gray abnormality degree Yht, the visible light image gray abnormality degree Ykt, the infrared image color abnormality degree Sht, and the visible light image color abnormality degree Skt are obtained, the image gray weight x1, x2 and the color weight y1, y2 are calculated, and the image fusion weight Z1 and Z2 are further calculated. The feature points are detected in the infrared image and the visible light image, the two images are matched and aligned under the same spatial reference system, and the fusion parameter of each pixel point after alignment is calculated, which helps to retain the detail information in the infrared image and the visible light image in the fusion process, and is of great significance for improving the precision of image fusion, retaining detail information, enhancing robustness, and optimizing visual effect.

[0077] The above embodiments can be realized wholly or partially by software, hardware, firmware, or any other combination. When realized by software, the above embodiments can be realized in the form of a computer program product wholly or partially. Those skilled in the art can realize that the units and algorithm steps of the examples described in connection with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized by hardware or software depends on the specific application and design constraints of the technical solutions.

[0078] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, i.e., they can be located in one place or distributed on multiple network units. Part or all of the units can be selected to achieve the purpose of the embodiments according to actual needs.

[0079] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application.

Claims

1. A method for fusing infrared images and visible light images, characterized by: The steps include: Import the infrared image and the visible light image into OpenCV, calculate the mean and variance of the grayscale images of the infrared image and the visible light image, and calculate the infrared image grayscale anomaly degree Yht and the visible light image grayscale anomaly degree Ykt based on the infrared image grayscale mean Hj, the infrared image grayscale variance Hf, the visible light image grayscale mean Kj and the visible light image grayscale variance Kf; Extract the RGB value of each pixel in the infrared image and the visible light image, calculate the RGB variance of the pixel in the infrared image and the visible light image, and calculate the color anomaly degree Sht of the infrared image and the color anomaly degree Skt of the visible light image based on the infrared R variance Rh, infrared G variance Gh, infrared B variance Bh, visible light R variance Rk, visible light G variance Gk and visible light B variance Bk; Obtain the infrared image grayscale anomaly degree Yht, the visible light image grayscale anomaly degree Ykt, the infrared image color anomaly degree Sht, and the visible light image color anomaly degree Skt, calculate the image grayscale weights x1, x2 and color weights y1, y2, and further calculate the image fusion weights Z1 and Z2, detect feature points in the infrared image and visible light image, match the two images and align them in the same spatial reference system, and calculate the fusion parameters of each pixel after alignment.

2. The method for fusing infrared images and visible light images according to claim 1, characterized in that: The infrared image and visible light image are imported into OpenCV and NumPy libraries. The cv2.cvtColor function of OpenCV is used to convert the image into grayscale image. The np.mean and np.var functions of NumPy are used to calculate the mean and variance of the grayscale images of the infrared image and visible light image, respectively, which are recorded as the infrared image grayscale mean Hj, the infrared image grayscale variance Hf, the visible light image grayscale mean Kj, and the visible light image grayscale variance Kf.

3. The method for fusing infrared images and visible light images according to claim 2, characterized in that: Obtain the infrared image grayscale mean Hj, infrared image grayscale variance Hf, visible light image grayscale mean Kj and visible light image grayscale variance Kf, and calculate the infrared image grayscale anomaly Yht and visible light image grayscale anomaly Ykt: in, is the expected normal grayscale mean of the infrared image, is the expected normal grayscale mean of the visible light image, is the mean of the historical data, is a small constant used to avoid the denominator being zero, and ϵ can be 1e−5 or 1e−10.

4. The method for fusing infrared images and visible light images according to claim 1, wherein: Use OpenCV to extract the RGB value of each pixel in the infrared image and the visible light image, and calculate the RGB variance of the pixel in the infrared image and the visible light image, which are recorded as infrared R variance Rh, infrared G variance Gh, infrared B variance Bh, visible light R variance Rk, visible light G variance Gk and visible light B variance Bk.

5. The method for fusing infrared images and visible light images according to claim 4, characterized in that: Obtain the infrared R variance Rh, infrared G variance Gh, infrared B variance Bh, visible light R variance Rk, visible light G variance Gk, and visible light B variance Bk, and calculate the infrared image color anomaly degree Sht and the visible light image color anomaly degree Skt: in, is the average value of the variance of the three channels R, G, and B of the historical infrared image, It is the average value of the variance of the three channels R, G, and B of the historical visible light image, which is used to standardize the variance value of each channel. It reflects the degree of deviation of the variance of each channel relative to the average variance. The difference between the variances of each channel is considered to capture the unevenness of color distribution.

6. The method for fusing infrared images and visible light images according to claim 3, characterized in that: Obtain the infrared image grayscale anomaly degree Yht and the visible light image grayscale anomaly degree Ykt, and calculate the image grayscale weights x1 and x2: in, is the mean value of the grayscale abnormality Yht of all historical infrared images, is the mean of the grayscale anomaly levels Ykt of all historical visible light images.

7. The method for fusing infrared images and visible light images according to claim 5, characterized in that: Obtain the color anomaly degree Sht of the infrared image and the color anomaly degree Skt of the visible light image, and calculate the image color weights y1 and y2: in, is the mean value of the color anomaly degree Sht of all historical infrared images, is the mean of the color anomaly levels Skt of all historical visible light images.

8. The method for fusing infrared images and visible light images according to claim 7, characterized in that: Get the image grayscale weights x1, x2 and color weights y1, y2, and calculate the image fusion weights Z1 and Z2: The calculation formulas of the corresponding image fusion weights Z1 and Z2 are as above.

9. The method for fusing infrared images and visible light images according to claim 1, characterized in that: Feature points are detected in infrared and visible light images, descriptors are calculated for each detected feature point, the feature descriptors in the two images are matched, the geometric transformation relationship between the two images is estimated based on the matched feature point pairs, and one image is transformed according to the obtained geometric transformation parameters so that the two images are aligned in the same spatial reference frame.

10. The method for fusing infrared images and visible light images according to claim 8, characterized in that: The fusion parameters of each pixel in the infrared image are calculated based on the image fusion weight Z1, and the fusion parameters of each pixel in the visible light image are calculated based on the image fusion weight Z2. Finally, the fusion parameters of the infrared image and the visible light image of each pixel after alignment are summed to obtain the fusion parameters of each pixel after alignment.

Images