CN-121458559-B - Method and device for enhancing infrared remote sensing image, computer equipment and storage medium

CN121458559BCN 121458559 BCN121458559 BCN 121458559BCN-121458559-B

Abstract

The application provides an enhancement method, device, computer equipment and storage medium of an infrared remote sensing image, wherein the method comprises the steps of carrying out global alignment on acquired continuous multi-frame infrared remote sensing images to generate multi-frame registered infrared remote sensing images, fusing the multi-frame registered infrared remote sensing images to generate a global enhancement image, after initial position information of a target object in the global enhancement image is determined, intercepting an initial local image of the target object from the infrared remote sensing image according to the initial position information of the target object, determining the target position information of the target object in the infrared remote sensing image by utilizing a Kalman filter, the initial local image and the position information of the target object at the previous moment, intercepting a target local image corresponding to the target object from the infrared remote sensing image according to the target position information of the target object, and generating the target enhancement image based on the global enhancement image and the target local image corresponding to the target object.

Inventors

LI LU
LI CHAO
CHEN YUE
ZHANG CHAOZI

Assignees

之江实验室

Dates

Publication Date: 20260508
Application Date: 20260106

Claims (10)

1. A method for enhancing an infrared remote sensing image, the method comprising: acquiring continuous multi-frame infrared remote sensing images; the method comprises the steps of carrying out global alignment on a plurality of frames of infrared remote sensing images to generate a plurality of frames of registered infrared remote sensing images, determining the image signal-to-noise ratio of the registered infrared remote sensing images according to the image mean value and the image variance of each frame of registered infrared remote sensing images, determining the image gradient energy of the registered infrared remote sensing images according to the pixel number of gradient images corresponding to the registered infrared remote sensing images and gradient information in a plurality of directions, determining the second-order gradient value of the registered infrared remote sensing images by utilizing a Laplace operator, determining the image ambiguity of the registered infrared remote sensing images according to the second-order gradient value, determining the image ambiguity of the registered infrared remote sensing images based on the image signal-to-noise ratio, the image gradient energy and the image ambiguity, determining the pixel information of the same pixel positions in the registered infrared remote sensing images based on the weighting information corresponding to the plurality of frames of registered remote sensing images, summing the pixel information corresponding to the pixel information of the same pixel positions in the registered infrared remote sensing images, and generating global enhanced fused pixel information based on the pixel position information corresponding to each pixel position; After initial position information of a target object in the global enhanced image is determined, intercepting an initial local image of the target object from the infrared remote sensing image according to the initial position information of the target object for each frame of the infrared remote sensing image, and determining the target position information of the target object in the infrared remote sensing image by utilizing a Kalman filter, the initial local image and the position information of the target object at the previous moment corresponding to the infrared remote sensing image, wherein the target object is an object with a pixel size smaller than a preset size in the infrared remote sensing image; And according to the target position information of the target object, intercepting a target local image corresponding to the target object from the infrared remote sensing image, and generating a target enhanced image based on the global enhanced image and the target local image corresponding to the target object.
2. The method according to claim 1, wherein the method further comprises: respectively carrying out image preprocessing on the multi-frame infrared remote sensing image to generate a processed infrared remote sensing image, wherein the image preprocessing comprises at least one of non-uniformity correction processing, dead pixel restoration processing and background preliminary inhibition processing; global alignment is carried out on the multi-frame infrared remote sensing images to generate multi-frame registered infrared remote sensing images, and the method comprises the following steps: and carrying out global alignment on the multi-frame processed infrared remote sensing image to generate multi-frame registered infrared remote sensing images.
3. The method of claim 1, wherein globally aligning the plurality of frames of infrared remote sensing images to generate a plurality of frames of registered infrared remote sensing images comprises: respectively extracting initial gradient images of each frame of infrared remote sensing image; processing the initial gradient image of the infrared remote sensing image by utilizing an optical flow pyramid calculation model aiming at each frame of infrared remote sensing image to generate a plurality of pyramid gradient images corresponding to the infrared remote sensing image, and uniformly sampling a target number of control points from each pyramid gradient image; Determining affine transformation matrix parameters between a second infrared remote sensing image and a first infrared remote sensing image in the multi-frame infrared remote sensing image based on a plurality of control points respectively included in the multi-frame infrared remote sensing image, wherein the first infrared remote sensing image is a reference infrared remote sensing image in the multi-frame infrared remote sensing image, and the second infrared remote sensing image is other infrared remote sensing images except the first infrared remote sensing image; And carrying out geometric correction and resampling on the second infrared remote sensing image by utilizing the affine transformation matrix parameters to generate a registered second infrared remote sensing image, wherein the first infrared remote sensing image and the registered second infrared remote sensing image form the multi-frame registered infrared remote sensing image.
4. The method according to claim 1, wherein determining the target position information of the target object in the infrared remote sensing image using a kalman filter, the initial partial image, and the position information of the target object at a previous time corresponding to the infrared remote sensing image includes: Correcting the energy centroid coordinates in the neighborhood of the position information of the target object at the previous moment to generate corrected position information; Tracking and correcting the corrected position information by using the Kalman filter and the determined state vector at the current moment to generate predicted position information of the target object at the current moment; Acquiring detection position information of the target object detected at the current moment, correcting energy centroid coordinates in a neighborhood of the detection position information based on pixel information of the initial partial image, and generating corrected detection position information; and determining target position information of the target object in the infrared remote sensing image based on the predicted position information and the corrected detection position information.
5. The method of claim 4, wherein the determining target location information for the target object in the infrared remote sensing image based on the predicted location information and the corrected detected location information comprises: Determining a position offset value between the predicted position information and the corrected detected position information; when the position offset value is smaller than a preset offset value, determining the predicted position information as target position information of the target object; And when the position offset value is greater than or equal to the preset offset value, determining average value position information between the predicted position information and the corrected detection position information, and determining the average value position information as target position information of the target object.
6. The method according to any one of claims 1-5, wherein generating a target enhanced image based on the global enhanced image and the target local image corresponding to the target object comprises: fusing a plurality of frames of target local images corresponding to the target object to generate a fused local image; Determining an adaptive fusion weight corresponding to the fusion local image, wherein the adaptive fusion weight is positively correlated with the signal-to-noise ratio of the fusion local image; and embedding the fusion local image into the global enhanced image based on the self-adaptive fusion weight to generate a target enhanced image.
7. The method of claim 6, wherein determining the adaptive fusion weights corresponding to the fused partial images comprises: determining a pixel mean value based on pixel information of a plurality of target pixel points belonging to the target object in the fused partial image; determining a pixel standard deviation based on pixel information of other pixel points except the target pixel point in the fused partial image; Calculating a local signal to noise ratio according to the pixel mean value and the pixel standard deviation; and generating self-adaptive fusion weights corresponding to the fusion local images according to the local signal-to-noise ratio and the set weight extreme value.
8. An apparatus for enhancing an infrared remote sensing image, the apparatus comprising: the acquisition module is used for acquiring continuous multi-frame infrared remote sensing images; the first generation module is used for carrying out global alignment on the multi-frame infrared remote sensing images to generate multi-frame registered infrared remote sensing images; determining the image gradient energy of the registered infrared remote sensing image according to the number of pixels of the gradient image corresponding to the registered infrared remote sensing image and gradient information in a plurality of directions, determining a second-order gradient value of the registered infrared remote sensing image by utilizing a Laplacian operator, determining the image ambiguity of the registered infrared remote sensing image according to the second-order gradient value, determining the weight information of the registered infrared remote sensing image based on the image signal-to-noise ratio, the image gradient energy and the image ambiguity, summing the pixel information of the same pixel positions in the infrared remote sensing images after a plurality of frames of registration based on the weight information respectively corresponding to the infrared remote sensing images after the plurality of frames of registration, obtaining the fused pixel information corresponding to the pixel positions, and generating a global enhanced image based on the fused pixel information respectively corresponding to each pixel position; The determining module is used for intercepting an initial local image of the target object from the infrared remote sensing image according to the initial position information of the target object for each frame of the infrared remote sensing image after determining the initial position information of the target object in the global enhanced image, and determining the target position information of the target object in the infrared remote sensing image by utilizing a Kalman filter, the initial local image and the position information of the target object at the previous moment corresponding to the infrared remote sensing image, wherein the target object is an object with a pixel size smaller than a preset size in the infrared remote sensing image; and the second generation module is used for intercepting a target local image corresponding to the target object from the multi-frame infrared remote sensing image according to the target position information of the target object, and generating a target enhanced image based on the global enhanced image and the target local image corresponding to the target object.
9. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the steps of the method of any of claims 1-7.
10. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor performs the steps of the method according to any of claims 1-7.

Description

Method and device for enhancing infrared remote sensing image, computer equipment and storage medium Technical Field The present application relates to the field of image processing technologies, and in particular, to an infrared remote sensing image enhancement method, an infrared remote sensing image enhancement device, a computer device, and a storage medium. Background The infrared remote sensing image is widely applied in the fields of military reconnaissance, environmental monitoring and the like, but often faces the problems of low signal-to-noise ratio, poor contrast, difficult identification of weak and small targets and the like due to the limitation of a sensor and complex background interference, and a plurality of signal-to-noise ratio enhancement methods are provided by researchers for improving the target detection and identification performance in order to improve the image quality and support the rear-end task. In the related technology, in order to fully mine the statistical characteristics or structural components of the image, the signal to noise ratio of the target is improved by utilizing the intrinsic characteristics and structural information of the infrared image, for example, the difference between a small target and a background is utilized, and a multi-scale gray difference operator and a local image entropy are combined to construct a weighted local image entropy, so that background clutter is effectively suppressed and a weak and small target is enhanced. In the other method, an image decomposition technology is adopted, the image is decomposed into different components, then the contrast enhancement is carried out by combining global and local mapping, and the detail components are enhanced by assisting an adaptive algorithm, so that the whole contrast is improved, and the edge is sharpened and the noise is suppressed. However, the enhancement algorithm takes a single infrared remote sensing image as an information source, and the signal-to-noise ratio enhancement effect is severely limited by the global signal-to-noise ratio and the imaging quality of the image, so that the enhancement effect is unstable. Disclosure of Invention In view of the foregoing, the present application provides a method, apparatus, computer device and storage medium for enhancing infrared remote sensing images. Specifically, the application is realized by the following technical scheme: In a first aspect, an embodiment of the present application provides a method for enhancing an infrared remote sensing image, including: acquiring continuous multi-frame infrared remote sensing images; the multi-frame infrared remote sensing images are subjected to global alignment to generate multi-frame registered infrared remote sensing images; After initial position information of a target object in the global enhanced image is determined, intercepting an initial local image of the target object from the infrared remote sensing image according to the initial position information of the target object for each frame of the infrared remote sensing image, and determining the target position information of the target object in the infrared remote sensing image by utilizing a Kalman filter, the initial local image and the position information of the target object at the previous moment corresponding to the infrared remote sensing image, wherein the target object is an object with a pixel size smaller than a preset size in the infrared remote sensing image; And according to the target position information of the target object, intercepting a target local image corresponding to the target object from the infrared remote sensing image, and generating a target enhanced image based on the global enhanced image and the target local image corresponding to the target object. In an alternative embodiment, the method further comprises: respectively carrying out image preprocessing on the multi-frame infrared remote sensing image to generate a processed infrared remote sensing image, wherein the image preprocessing comprises at least one of non-uniformity correction processing, dead pixel restoration processing and background preliminary inhibition processing; global alignment is carried out on the multi-frame infrared remote sensing images to generate multi-frame registered infrared remote sensing images, and the method comprises the following steps: and carrying out global alignment on the multi-frame processed infrared remote sensing image to generate multi-frame registered infrared remote sensing images. In an optional implementation manner, the performing global alignment on the multi-frame infrared remote sensing image to generate a multi-frame registered infrared remote sensing image includes: respectively extracting initial gradient images of each frame of infrared remote sensing image; processing the initial gradient image of the infrared remote sensing image by utilizing an optical flow pyramid calculation model aiming a