Remote-sensing image color balancing method and apparatus, electronic device and medium

Abstract

The present invention relates to technical fields such as computer vision and image processing. Provided are a remote-sensing image color balancing method and apparatus, an electronic device and a medium. The specific implementation scheme comprises: acquiring reference instance images respectively corresponding to a plurality of reference objects, and an image to be color-balanced; for each reference instance image, determining, from the image to be color-balanced, at least one target instance image which has the same reference object as the reference instance image; for each reference instance image, performing control point matching between the reference instance image and the corresponding at least one target instance image to obtain at least one control point pair; on the basis of color information of the at least one control point pair corresponding to each of the plurality of reference instance images, determining a plurality of color mapping relationships, which represent a correspondence, before and after color transformation, of color information of each pixel point in the image to be color-balanced; and on the basis of the plurality of color mapping relationships, performing color balancing processing on the image to be color-balanced, so as to obtain a target image.

Description

Remote sensing image color uniformity method, device, electronic equipment and medium Technical Field

[0001] The present invention relates to technical fields such as computer vision and image processing, and in particular to a remote sensing image color uniformity method, device, electronic equipment and storage medium. Background Art

[0002] In the application of remote sensing images, due to differences in imaging conditions such as imaging time, lighting, shooting angle, and ground object type, large color differences can occur between images. In order to ensure the overall color consistency of the image, it is usually necessary to perform color uniformity processing on the remote sensing image. Summary of the Invention

[0003] The present invention provides a remote sensing image color uniformity method, device, electronic equipment, storage medium and computer program product.

[0004] According to one aspect of the present invention, a remote sensing image color equalization method is provided, comprising: obtaining reference instance images corresponding to a plurality of reference objects and an image to be color equalized; for each reference instance image, determining at least one target instance image in the image to be color equalized that is identical to the reference object in the reference instance image; for each reference instance image, performing control point matching between the reference instance image and the corresponding at least one target instance image to obtain at least one control point pair; determining a plurality of color mapping relationships based on color information of the at least one control point pair corresponding to each of the plurality of reference instance images, the plurality of color mapping relationships representing the corresponding relationship between the color information of each pixel point in the image to be color equalized before and after color change; and performing color equalization processing on the image to be color equalized based on the plurality of color mapping relationships to obtain a target image.

[0005] According to an embodiment of the present invention, for each reference instance image, control point matching is performed between the reference instance image and the corresponding at least one target instance image to obtain at least one control point pair, including: determining at least one reference control point in the reference instance image; for each reference control point, determining the color deviation between the reference control point and each pixel point in the corresponding at least one target instance image, and based on the color deviation, determining the target control point corresponding to the reference control point from a plurality of pixel points contained in the corresponding at least one target instance image, wherein the reference control point and the corresponding target control point constitute a control point pair.

[0006] According to an embodiment of the present invention, the color information of each control point pair includes the color information of the reference control point in the control point pair in the corresponding reference instance image and the color information of the target control point in the control point pair in the corresponding at least one target instance image; based on the color information of at least one control point pair corresponding to each of the multiple reference instance images, determining multiple color mapping relationships includes: based on at least one control point pair corresponding to each of the multiple reference instance images, obtaining a reference control point set including multiple reference control points, a target control point set including multiple target control points, and a correspondence between each reference control point and each target control point; using the Delaunay triangulation algorithm, constructing a reference triangular mesh and a target triangular mesh based on the reference control point set and the target control point set respectively; and determining multiple color mapping relationships based on the reference triangular mesh, the target triangular mesh, the correspondence between each reference control point and each target control point, the color information of each reference control point in the corresponding reference instance image, and the color information of each target control point in the corresponding at least one target instance image.

[0007] According to an embodiment of the present invention, based on the reference triangular mesh, the target triangular mesh, the correspondence between each reference control point and each target control point, the color information of each reference control point in the corresponding reference instance image, and the color information of each target control point in the corresponding at least one target instance image, determining multiple color mapping relationships includes: for any target tetrahedron in the target triangular mesh, based on the correspondence between each reference control point and each target control point, determining a reference tetrahedron corresponding to the target tetrahedron in the reference triangular mesh; determining a color mapping relationship between the target tetrahedron and the corresponding reference tetrahedron according to the color information corresponding to each of the four target control points constituting the target tetrahedron, the color information corresponding to each of the four reference control points constituting the corresponding reference tetrahedron, and the correspondence between each reference control point and each target control point; repeating the above operations until the color mapping relationships between the multiple target tetrahedrons in the target triangular mesh and the corresponding reference tetrahedrons in the reference triangular mesh are obtained; and using the color mapping relationships between the multiple target tetrahedrons in the target triangular mesh and the corresponding reference tetrahedrons in the reference triangular mesh as the multiple color mapping relationships.

[0008] According to an embodiment of the present invention, performing color uniformity processing on an image to be uniformly colored based on multiple color mapping relationships to obtain a target image includes: for any pixel point in the image to be uniformly colored, based on the color information of the pixel point, determining a target color mapping relationship that matches the color information of the pixel point from multiple color mapping relationships; mapping the color information of the pixel point based on the target color mapping relationship; and repeatedly performing the above operations until the color information of all pixels in the image to be uniformly colored is mapped to obtain the target image.

[0009] According to an embodiment of the present invention, based on the color information of the pixel point, determining a target color mapping relationship that matches the color information of the pixel point from multiple color mapping relationships includes: based on the color information of the pixel point, determining a target tetrahedron that matches the color information of the pixel point in the target triangular mesh; obtaining the target color mapping relationship based on the target tetrahedron that matches the color information of the pixel point and the correspondence between multiple target tetrahedrons and multiple color mapping relationships in the target triangular mesh.

[0010] According to an embodiment of the present invention, based on the color deviation degree, determining the target control point corresponding to the reference control point from multiple pixel points contained in the corresponding at least one target instance image includes: in response to determining that the number of pixel points whose color deviation degree meets the preset threshold in the corresponding at least one target instance image is greater than or equal to the quantity threshold, clustering the pixel points whose color deviation degree meets the preset threshold according to the color information of each pixel point to obtain at least one first pixel point cluster; taking the first pixel point cluster with the largest number of pixels in the at least one first pixel point cluster as the target pixel point cluster; and determining the cluster center of the target pixel point cluster as the target control point corresponding to the reference control point.

[0011] According to an embodiment of the present invention, determining at least one reference control point in a reference instance image includes: clustering the pixels in the reference instance image based on the color information of each pixel to obtain at least one second pixel cluster; extracting one pixel from each of the at least one second pixel clusters to obtain at least one reference control point.

[0012] According to one aspect of the present invention, a remote sensing image color equalization device is provided, comprising: an acquisition module for acquiring reference instance images corresponding to a plurality of reference objects and an image to be color equalized; a first determination module for determining, for each reference instance image, at least one target instance image in the image to be color equalized that is identical to the reference object in the reference instance image; a matching module for performing control point matching between the reference instance image and the corresponding at least one target instance image for each reference instance image to obtain at least one control point pair; a second determination module for determining, based on color information of at least one control point pair corresponding to each of the plurality of reference instance images, a plurality of color mapping relationships, the plurality of color mapping relationships representing the corresponding relationship between the color information of each pixel point in the image to be color equalized before and after color change; and a color equalization module for performing color equalization processing on the image to be color equalized based on the plurality of color mapping relationships to obtain a target image.

[0013] According to another aspect of the present invention, an electronic device is provided, comprising: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors execute the method described above.

[0014] According to another aspect of the present invention, a computer-readable storage medium is provided, on which executable instructions are stored. When the instructions are executed by a processor, the processor is enabled to implement the method described above.

[0015] According to another aspect of the present invention, a computer program product is provided, comprising a computer program, which implements the method described above when executed by a processor. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort.

[0017] FIG1 is a flow chart of a remote sensing image color uniformity method according to an embodiment of the present invention;

[0018] FIG2 schematically shows a process of determining a target instance image that is identical to a reference object in a reference instance image in an image to be color-homogenized;

[0019] FIG3 schematically shows a process diagram for determining a color mapping relationship between a target tetrahedron and a corresponding reference tetrahedron;

[0020] FIG4 is a block diagram of a remote sensing image color uniformity device according to an embodiment of the present invention;

[0021] FIG5 is a block diagram of an electronic device suitable for implementing a remote sensing image color uniformity method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] The following will be combined with the embodiments of the present invention and the accompanying drawings to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of the present invention.

[0023] It should be noted that the sequence numbers of the operations in the following method are only used to indicate the operation for the purpose of description and should not be regarded as indicating the order in which the operations should be performed. Unless explicitly stated, the method does not need to be performed in the order shown.

[0024] In addition, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0025] In the technical solution of the present invention, the collection, storage, use, processing, transmission, provision, disclosure and application of the data involved (for example, including but not limited to user personal information) shall comply with the provisions of relevant laws and regulations and shall not violate public order and good morals.

[0026] In the technical solution of the present invention, before obtaining or collecting relevant data, the authorization or consent of the data owner is obtained.

[0027] To facilitate understanding of the technical solutions and technical effects of the embodiments of the present invention, the terms involved in the embodiments of the present invention are explained below.

[0028] Instance-level Image Retrieval (IIR): Given a query image of a specific instance, the goal is to find images containing the same instance from candidate images. The candidate images may have been taken under different conditions, such as imaging distance, viewing angle, background, lighting, and weather conditions.

[0029] RGB color space: In RGB color space, each color is obtained by varying and superimposing the three color channels: red (R), green (G), and blue (B). In other words, any color can be represented by a three-dimensional color vector [R, G, B].

[0030] HSV color space: In the HSV (Hue, Saturation, Value) color space, any color can be obtained by varying and superimposing the three channels of hue (Hue), saturation (Saturation), and value (Value). In other words, any color can be represented by a three-dimensional color vector [H, S, V].

[0031] Lab color space: In the Lab color space, each color is represented by a lightness component (L), a color component ranging from green to red (a, hereafter referred to as the "first color component"), and a color component ranging from blue to yellow (b, hereafter referred to as the "second color component"). In other words, any color can be represented by a three-dimensional color vector [L, a, b].

[0032] Delaunay triangulation algorithm: The definition of the triangulation algorithm is to assume that V is a finite set of points in the two-dimensional real number field, an edge e is a closed line segment formed by the points in the point set as endpoints, and E is the set of e. Then a triangulation T = (V, E) of the point set V is a planar graph G that satisfies the following conditions: (1) Except for the endpoints, the edges in the planar graph do not contain any points in the point set; (2) there are no intersecting edges; (3) all faces in the planar graph are triangular faces, and the union of all triangular faces is the convex hull of the scattered point set V. The definition of a Delaunay edge is to assume that an edge e in E (with two endpoints a and b) is called a Delaunay edge if it satisfies the following conditions: there exists a circle passing through points a and b, and the circle does not contain any other points in the point set V (note that it is the circle, and at most three points on the circle are cocircular). This property is also called the empty circle property. The Delaunay triangulation algorithm is defined as follows: if a triangulation T of a point set V contains only Delaunay edges, then the triangulation is called a Delaunay triangulation.

[0033] FIG1 is a flow chart of a remote sensing image color uniformity method according to an embodiment of the present invention.

[0034] As shown in FIG1 , the remote sensing image color uniformity method 100 includes operations S110 to S150 .

[0035] In operation S110 , a plurality of reference instance images corresponding to respective reference objects and an image to be color-uniformed are acquired.

[0036] In operation S120 , for each reference instance image, at least one target instance image having the same object as the reference object in the reference instance image is determined in the image to be uniformed.

[0037] In operation S130 , for each reference instance image, control point matching is performed between the reference instance image and at least one corresponding target instance image to obtain at least one control point pair.

[0038] In operation S140 , a plurality of color mapping relationships are determined based on color information of at least one control point pair corresponding to each of the plurality of reference instance images.

[0039] In operation S150 , the image to be uniformly colored is subjected to color uniformity processing based on the multiple color mapping relationships to obtain a target image.

[0040] The image to be color-homogenized can be, for example, a remote sensing image acquired by an image acquisition device (including, but not limited to, satellite remote sensing equipment, unmanned aerial vehicle remote sensing equipment, etc.). The image to be color-homogenized can include data from multiple bands, such as R (red) band data, G (green) band data, and B (blue) band data.

[0041] A reference object is an independent object, such as a house or lake. A reference instance image is an instance image that contains the reference object. Each reference instance image contains a reference object, and the reference objects in each reference instance image are different.

[0042] Each reference example image can be, for example, an example image pre-extracted from different remote sensing images. The reference objects in these reference example images cover all objects to be color-leveled in the image to be color-leveled. From these reference example images, a reference example image matching each object to be color-leveled can be selected to serve as a "color template." By mapping the colors of the selected reference example image to the image to be color-leveled, color-leveling can be achieved for all objects to be color-leveled in the image to be color-leveled.

[0043] Figure 2 is a schematic diagram illustrating the process of identifying a target instance image in an image to be color-leveled that is identical to a reference object in a reference instance image. It should be noted that, for ease of explanation, in this embodiment of the present invention, instance images with identical objects are identified with the same symbol. Furthermore, the number of reference instance images and the number of objects to be color-leveled in the image to be color-leveled shown in Figure 2 are merely exemplary, and embodiments of the present invention are not limited thereto and may include a greater number of reference instance images and objects to be color-leveled.

[0044] As shown in FIG2 , multiple reference example images 210 include, for example, reference example images corresponding to five reference objects, such as reference example image 211 corresponding to reference object 1, reference example image 212 corresponding to reference object 2, ..., and reference example image 215 corresponding to reference object 5. Reference objects 1 through 5 are not shown in FIG2 . The reference objects (e.g., reference objects 1 through 5) in multiple reference example images 210 cover all objects to be color-leveled in image 220 to be color-leveled.

[0045] For example, a target instance image identical to reference object 1 in reference instance image 211 is identified in image 220 to be color-leveled. For example, based on an instance retrieval algorithm or a pre-trained instance retrieval deep learning model, multiple candidate instance images identical to reference object 1 in reference instance image 211 can be identified from image 220 to be color-leveled. These multiple candidate instance images are then used as multiple target instance images (as shown in the dashed boxes in FIG. 2 ). Each target instance image (e.g., target instance image 221) and reference instance image 211 share the same reference object, e.g., the same reference object 1.

[0046] Similarly, for other reference instance images in the plurality of reference instance images 210, such as reference instance images 212 through 215, based on an instance retrieval algorithm or a pre-trained instance retrieval deep learning model, if at least one candidate instance image identical to the reference object in each reference instance image can be identified from the image to be homogenized 220, then that at least one candidate instance image is used as the at least one target instance image corresponding to that reference instance image. For example, at least one target instance image corresponding to each of reference instance images 212 through 214 can be matched in the image to be homogenized 220. If no candidate instance image identical to the reference object in the reference instance image is matched in the image to be homogenized 220, then that reference instance image is deleted. For example, if no target instance image corresponding to reference instance image 215 is matched in the image to be homogenized 220, then reference instance image 215 can be deleted from the plurality of reference instance images 210.

[0047] In an embodiment of the present invention, a suitable instance retrieval algorithm or a pre-trained instance retrieval deep learning model can be selected according to actual needs, and the present invention is not limited to this.

[0048] Through the above method, instance-level correspondences can be established between multiple reference instance images and the target instance images in the image to be color-leveled. In other words, correspondences are established between the reference objects in each reference instance image and the respective objects to be color-leveled in the image to be color-leveled. Because each reference instance image and the corresponding at least one target instance image contain multiple pixels, and these pixels have corresponding color information in the corresponding instance image, based on the above instance-level correspondences, a pixel-level color mapping relationship can be established between each reference instance image and the corresponding at least one target instance image.

[0049] For each reference example image, control point matching is performed between the reference example image and at least one corresponding target example image to obtain at least one control point pair. Then, based on the color information of the at least one control point pair corresponding to each of the multiple reference example images, multiple color mapping relationships are determined. These multiple color mapping relationships represent the corresponding relationship between the color information of each pixel in the image to be color-leveled before and after the color change.

[0050] In an embodiment of the present invention, the color information of at least one control point pair corresponding to each reference instance image can be used to characterize the color correspondence between each reference instance image and the corresponding at least one target instance image. Based on the above color correspondence, a pixel-level color mapping relationship can be constructed between each reference instance image and the corresponding at least one target instance image. In other words, based on the color information of at least one control point pair corresponding to each of the multiple reference instance images, a color mapping relationship for each color throughout the entire color space can be constructed between the multiple reference instance images and the image to be color-leveled, that is, multiple pixel-level color mapping relationships can be constructed between the multiple reference instance images and the image to be color-leveled. Thus, multiple color mapping relationships can be used to map the color information of each pixel in the image to be color-leveled, thereby achieving color-leveling processing of the image to be color-leveled.

[0051] According to the technical solution of an embodiment of the present invention, when performing color grading on an image to be grading using multiple reference example images as color templates, an instance-level correspondence is established between the multiple reference example images and the target example image in the image to be grading. Based on this correspondence, control points are matched between each reference example image and at least one corresponding target example image, and multiple pixel-level color mapping relationships are constructed based on the color information of each control point pair. The image to be grading is then grading using these multiple color mapping relationships. This grading solution significantly reduces the computational effort involved in the image grading process while ensuring grading accuracy and quality, thereby improving grading efficiency.

[0052] According to an embodiment of the present invention, for each reference instance image, performing control point matching between the reference instance image and at least one corresponding target instance image to obtain at least one control point pair includes the following operations.

[0053] First, at least one reference control point is determined in a reference instance image.

[0054] In one example, a clustering algorithm (e.g., including but not limited to a k-mean clustering algorithm and a k-nearest neighbor clustering algorithm) can be used to cluster the pixels in the reference instance image based on the color information of each pixel to obtain at least one second pixel cluster. Each second pixel cluster corresponds to a color category. Then, one pixel is extracted from each of the at least one second pixel clusters to obtain at least one reference control point. The color information of the at least one reference control point can be used to characterize the color information of all pixels in the reference instance image.

[0055] Next, for each reference control point, the color deviation between the reference control point and each pixel point in the corresponding at least one target instance image is determined, and based on the color deviation, the target control point corresponding to the reference control point is determined from the multiple pixel points contained in the corresponding at least one target instance image, where the reference control point and the corresponding target control point constitute a control point pair.

[0056] In one example, the reference instance image and the corresponding at least one target instance image can be converted from the RGB color space to the HSV color space respectively, and the color deviation between each reference control point and each pixel in the corresponding at least one target instance image can be calculated in the HSV color space.

[0057] Since the RGB color space and the HSV color space essentially represent the same color using different color models, for any pixel in each instance image, the color information of the pixel can be converted from the RGB color space to the HSV color space.

[0058] In the HSV color space, color deviation can include, for example, hue deviation, saturation deviation, and lightness deviation. The following formulas (1) to (3) can be used to determine the hue deviation, saturation deviation, and lightness deviation. Δh = min(|h1-h2|, 360-|h1-h2|) / 180 (1) Δs = |s1-s2| (2) Δv = |v1-v2| / 255 (3)

[0059] In formulas (1) to (3), Δh represents the hue deviation between the reference control point and any pixel in the corresponding at least one target instance image, Δs represents the saturation deviation between the reference control point and any pixel in the corresponding at least one target instance image, Δv represents the brightness deviation between the reference control point and any pixel in the corresponding at least one target instance image, h1 and h2 represent the hue of the reference control point and the hue of any pixel in the corresponding at least one target instance image, s1 and s2 represent the saturation of the reference control point and the saturation of any pixel in the corresponding at least one target instance image, and v1 and v2 represent the brightness of the reference control point and the brightness of any pixel in the corresponding at least one target instance image, respectively.

[0060] In another example, the reference instance image and the corresponding at least one target instance image can be converted from the RGB color space to the Lab color space respectively, and the color deviation between each reference control point and each pixel point in the corresponding at least one target instance image can be calculated in the Lab color space.

[0061] In the Lab color space, the color deviation degree may include, for example, the brightness component deviation degree and the color component deviation degree. The brightness component deviation degree and the color component deviation degree are calculated using the following formulas (4) to (5), respectively. ΔL = |L1-L2| (4)

[0062] In formulas (4) to (5), ΔL represents the luminance component deviation between the reference control point and any pixel in the corresponding at least one target instance image, Δc represents the color component deviation between the reference control point and any pixel in the corresponding at least one target instance image, L1 and L2 represent the luminance component of the reference control point and the luminance component of any pixel in the corresponding at least one target instance image, respectively, a1 and b1 represent the first color component and the second color component of the reference control point, respectively, and a2 and b2 represent the first color component and the second color component of any pixel in the corresponding at least one target instance image, respectively.

[0063] It should be noted that, since at least one target instance image matching each reference instance image is identified from the image to be color-leveled, converting each target instance image from the RGB color space to the HSV color space (or Lab color space) is equivalent to converting the image to be color-leveled from the RGB color space to the HSV color space (or Lab color space). In some embodiments, the image to be color-leveled may also be directly converted from the RGB color space to the HSV color space (or Lab color space), which is not limited in the present invention.

[0064] In an embodiment of the present invention, determining a target control point corresponding to a reference control point from a plurality of pixel points included in at least one target instance image based on the color deviation may include the following operations.

[0065] When it is determined that the number of pixels corresponding to at least one target instance image whose color deviation satisfies a preset threshold is greater than or equal to the number threshold, clustering is performed on the pixels whose color deviation satisfies the preset threshold based on the color information of each pixel to obtain at least one first pixel cluster. Thereafter, the first pixel cluster with the largest number of pixels in the at least one first pixel cluster is determined as the target pixel cluster, and the cluster center of the target pixel cluster is determined as the target control point corresponding to the reference control point.

[0066] In an embodiment of the present invention, one of the following methods may be used to determine the pixel points corresponding to at least one target instance image whose color deviation satisfies a preset threshold.

[0067] Method 1: When calculating the color deviation between each reference control point and each pixel in the corresponding at least one target instance image in the HSV color space, for each pixel in the corresponding at least one target instance image, determine whether the hue deviation between the pixel and the reference control point satisfies a first preset threshold, whether the saturation deviation satisfies a second preset threshold, and whether the lightness deviation satisfies a third preset threshold. If it is determined that the hue deviation satisfies the first preset threshold, the saturation deviation satisfies the second preset threshold, and the lightness deviation satisfies the third preset threshold—in other words, if the hue deviation, saturation deviation, and lightness deviation all satisfy the corresponding preset thresholds—then the pixel is determined to have a color deviation that satisfies the preset thresholds; otherwise, the pixel is discarded.

[0068] Method 2: When calculating the color deviation between each reference control point and each pixel in the corresponding at least one target instance image in the Lab color space, for each pixel in the corresponding at least one target instance image, determine whether the brightness component deviation between the pixel and the reference control point satisfies a fourth preset threshold, and whether the color component deviation satisfies a fifth preset threshold. If it is determined that the brightness component deviation satisfies the fourth preset threshold and the color component deviation satisfies the fifth preset threshold, then the pixel is determined to be a pixel whose color deviation satisfies the preset threshold; otherwise, the pixel is discarded.

[0069] After all pixels corresponding to at least one target instance image whose color deviation satisfies a preset threshold are screened using the above method, a determination is made as to whether the number of screened pixels is greater than or equal to a threshold. If the number of pixels whose color deviation satisfies the preset threshold is greater than or equal to the threshold, clustering is performed on these pixels based on the color information of each screened pixel to obtain at least one first pixel cluster. The first pixel cluster with the largest number of pixels in the at least one first pixel cluster is then determined as the target pixel cluster, and the cluster center of the target pixel cluster is determined as the target control point corresponding to the reference control point. If the number of pixels whose color deviation satisfies the preset threshold is less than the threshold, this indicates that the number of pixels corresponding to the reference control point in the at least one target instance image is relatively small, indicating that there may be an error in the matching process between the reference instance image and the at least one target instance image. In this case, a target instance image with the same reference object as the reference instance image can be re-identified in the image to be color-matched, so that control point pairs can be matched between the reference instance image and the new target instance image until at least one control point pair corresponding to the reference instance image is obtained. In some embodiments, if at least one corresponding control point pair cannot be obtained by matching using a reference instance image, the reference instance image is deleted from the plurality of reference instance images.

[0070] In an embodiment of the present invention, during the process of generating target control points, pixel points are filtered by using a quantity threshold, so that the pixel point clusters that ultimately participate in generating the target control points are themselves clustered with pixel point clusters of more colors, which is conducive to improving the accuracy of generating target control points.

[0071] In the embodiment of the present invention, the first to fifth preset thresholds and the quantity threshold can be set according to actual needs, and the present invention does not limit this.

[0072] According to an embodiment of the present invention, the color information of each control point pair includes the color information of the reference control point in the control point pair in the corresponding reference instance image and the color information of the target control point in the control point pair in the corresponding at least one target instance image. In an embodiment of the present invention, the color information of any pixel in the target pixel cluster obtained during the generation of the target control point in the corresponding at least one target instance image can be used as the color information of the target control point in the corresponding at least one target instance image. In an embodiment of the present invention, the color information of the target control point in the corresponding at least one target instance image is defined in the same manner and will not be further described.

[0073] Since the color information of at least one control point pair corresponding to each reference instance image can be used to characterize the color correspondence between each reference instance image and the corresponding at least one target instance image, based on the color information of at least one control point pair corresponding to each of the multiple reference instance images, a color mapping relationship for each color throughout the entire color space can be constructed between the multiple reference instance images and the image to be uniformly colored, that is, multiple pixel-level color mapping relationships can be constructed between the multiple reference instance images and the image to be uniformly colored.

[0074] Since the Lab color space is larger than the RGB color space, this means that any pixel point (any color) in the RGB color space can find a corresponding mapping point in the Lab color space. In addition, in the Lab color space, the brightness channel and the color channel can be independent of each other and do not interfere with each other, which facilitates better processing of color components. Therefore, in some embodiments, the color information of each control point pair can be converted from the RGB color space to the Lab color space, and in the Lab color space, multiple color mapping relationships are constructed based on the color correspondence between the reference control point and the target control point in each control point pair. Then, the image to be uniformly colored is converted from the RGB color space to the Lab color space, and the color information of each pixel point in the image to be uniformly colored is mapped using multiple color mapping relationships in the Lab color space to obtain the target image. Thereafter, the target image is converted from the Lab color space to the RGB color space. Performing uniform color processing on the image to be uniformly colored in the above manner is conducive to obtaining higher color change accuracy and better uniform color effect.

[0075] The following describes the technical solution of the present invention using the example of constructing multiple color mapping relationships in the Lab color space and utilizing these multiple color mapping relationships to perform color grading on a color-homogeneous image. It should be understood that the embodiments of the present invention are not limited to this example. Multiple color mapping relationships may also be constructed in other suitable color spaces and utilized to perform color grading on a color-homogeneous image, depending on actual circumstances. The present invention is not limited to this example.

[0076] Determining multiple color mapping relationships based on color information of at least one control point pair corresponding to each of multiple reference instance images may include the following operations, for example.

[0077] First, the color information of at least one control point pair corresponding to each of the plurality of reference instance images is converted from the RGB color space to the Lab color space.

[0078] Next, in the Lab color space, based on at least one control point pair corresponding to each of the multiple reference instance images, a reference control point set including multiple reference control points, a target control point set including multiple target control points, and a correspondence between each reference control point and each target control point are obtained.

[0079] According to an embodiment of the present invention, each control point pair includes a reference control point and a target control point. The reference control point and the target control point have a corresponding relationship. Based on at least one control point pair corresponding to each of the multiple reference instance images, a reference control point set including multiple reference control points, a target control point set including multiple target control points, and a corresponding relationship between each reference control point in the reference control point set and each target control point in the target control point set can be extracted.

[0080] Next, the Delaunay triangulation algorithm is used to construct the reference triangular mesh and the target triangular mesh based on the reference control point set and the target control point set, respectively.

[0081] Next, multiple color mapping relationships are determined based on the reference triangular mesh, the target triangular mesh, the correspondence between each reference control point and each target control point, the color information of each reference control point in the corresponding reference instance image, and the color information of each target control point in the corresponding at least one target instance image.

[0082] Fig. 3 schematically shows a process diagram of determining a color mapping relationship between a target tetrahedron and a corresponding reference tetrahedron.

[0083] As shown in FIG3 , a target triangular mesh 31 and a reference triangular mesh 32 can be constructed based on a target control point set and a reference control point set using the Delaunay triangulation algorithm.

[0084] For any target tetrahedron in the target triangular mesh 31 , a reference tetrahedron corresponding to the target tetrahedron is determined in the reference triangular mesh 32 based on the correspondence between each reference control point and each target control point.

[0085] For example, target tetrahedron 31_1 is any target tetrahedron in target triangular mesh 31. The four target control points constituting target tetrahedron 31_1 are target control point 311A, target control point 311B, target control point 311C, and target control point 311D. Based on the four target control points constituting target tetrahedron 31_1 (i.e., target control points 311A-311D), the correspondence between each reference control point and each target control point is queried to obtain the reference control point corresponding to each target control point. For example, target control point 311A ​​corresponds to reference control point 321A, target control point 311B corresponds to reference control point 321B, target control point 311C corresponds to reference control point 321C, and target control point 311D corresponds to reference control point 321D. Then, based on the reference control points (eg, reference control points 321A to 321D) obtained by the query, a reference tetrahedron 32_1 corresponding to the target tetrahedron 31_1 is determined in the reference triangular mesh 32 .

[0086] Next, the color mapping relationship between the target tetrahedron 31_1 and the reference tetrahedron 32_1 is determined based on the color information corresponding to the four target control points constituting the target tetrahedron 31_1, the color information corresponding to the four reference control points constituting the reference tetrahedron 32_1, and the correspondence between each reference control point and each target control point.

[0087] In the Lab color space, the color information of each control point refers to, for example, the brightness component L, the first color component a, and the second color component b of the control point in the Lab color space. The color information of each control point can be expressed as [L, a, b]. Assume that the color information corresponding to the four target control points 311A ​​to 311D constituting the target tetrahedron 31_1 is [L, a, b]. 11 ,a 11 ,b 11 ]、[L 12 ,a 12 ,b 12 ]、[L 13 ,a 13 ,b 13 ]、[L 14 ,a 14 ,b 14 ], the color information corresponding to the four reference control points 321A to 321D constituting the reference tetrahedron 32_1 are [L 21 ,a 21 ,b 21 ]、[L 22 ,a 22 ,b 22 ]、[L 23 ,a 23 ,b 23 ]、[L 24 ,a 24 ,b 24 ].

[0088] Based on the correspondence between each reference control point and each target control point, it can be seen that target control points 311A-311D correspond to reference control points 321A-321D, respectively. Based on this correspondence, the color information of each of the four target control points 311A-311D and the color information of each of the reference control points 321A-321D can be used to construct a color mapping relationship between target tetrahedron 31_1 and reference tetrahedron 32_1. In other words, based on the color mapping relationship between the four target control points that constitute target tetrahedron 31_1 and the four reference control points that constitute reference tetrahedron 32_1, a color mapping relationship corresponding to each color throughout the entire color space can be constructed between target tetrahedron 31_1 and reference tetrahedron 32_1.

[0089] In an embodiment of the present invention, for each of the target control points 311A ​​to 311D, a color mapping relationship as shown in the following formula (6) can be constructed based on the color information of the target control point and the color information of the corresponding reference control point, that is, the color information of a control point pair.

[0090] In formula (6), L 1k 、a 1k 、b 1k They represent the brightness component, the first color component and the second color component of the kth target control point in the Lab color space, L 2k 、a 2k 、b 2k They represent the brightness component, the first color component, and the second color component of the reference control point corresponding to the kth target control point in the Lab color space, respectively. represents the color mapping coefficient matrix, λ 11 ~λ 33 represents the color mapping coefficients, Represents a constant matrix, μ1~μ3 are all constants, k=1,2,3,4.

[0091] Based on the above-mentioned target control points 311A~311D and reference control points 321A~321D, four control point pairs can be obtained. By applying the color information of the four control point pairs to formula (6) respectively, the color mapping relationship between the target tetrahedron 31_1 and the reference tetrahedron 32_1 can be constructed. By solving the color mapping relationship between the above-mentioned target tetrahedron 31_1 and the reference tetrahedron 32_1, the above-mentioned color mapping coefficient matrix and constant matrix can be determined. Thereafter, based on the color mapping coefficient matrix and the constant matrix, the color mapping relationship between the target tetrahedron 31_1 and the reference tetrahedron 32_1 can be determined. In an embodiment of the present invention, there is a color mapping relationship between a target tetrahedron and a corresponding reference tetrahedron.

[0092] By repeating the above operations, the color mapping relationships between the target tetrahedrons in the target triangular mesh 31 and the corresponding reference tetrahedrons in the reference triangular mesh 32 can be obtained. Then, the color mapping relationships between the target tetrahedrons in the target triangular mesh 31 and the corresponding reference tetrahedrons in the reference triangular mesh 32 are used as the above-mentioned multiple color mapping relationships.

[0093] After obtaining the above-mentioned multiple color mapping relationships, the image to be color-leveled can be converted from the RGB color space to the Lab color space. Then, in the Lab color space, the image to be color-leveled is color-leveled based on the multiple color mapping relationships to obtain the target image.

[0094] According to an embodiment of the present invention, in the Lab color space, for any pixel in the image to be uniformly colored, a target color mapping relationship that matches the color information of the pixel can be determined from multiple color mapping relationships based on the color information of the pixel.

[0095] In an embodiment of the present invention, for example, in Lab color space, for any pixel in the image to be color-leveled, a target tetrahedron matching the color information of the pixel is determined in the target triangular mesh based on the color information of the pixel. Then, based on the target tetrahedron matching the color information of the pixel and the correspondence between multiple target tetrahedrons and multiple color mapping relationships in the target triangular mesh, a target color mapping relationship is obtained.

[0096] For example, in the Lab color space, the color information of the pixel point and the color information of each target control point in the triangular mesh can be regarded as their respective position coordinates in the Lab color space. Based on the color information of the pixel point and the color information of each target control point in the triangular mesh, the positional relationship between the pixel point and each target tetrahedron in the target triangular mesh can be determined. Then, based on the positional relationship, the target tetrahedron that matches the color information of the pixel point is determined, and based on the target tetrahedron that matches the color information of the pixel point, the correspondence between multiple target tetrahedrons and multiple color mapping relationships in the target triangular mesh is queried to obtain the above-mentioned target color mapping relationship.

[0097] For example, if it is determined that the pixel point is located within a target tetrahedron in the target triangular mesh, the color mapping relationship corresponding to the target tetrahedron is determined as the target color mapping relationship. If it is determined that the pixel point is located on a common edge (or common point) of multiple target tetrahedrons in the target triangular mesh, then one target tetrahedron can be arbitrarily selected from the multiple target tetrahedrons, and based on the selected target tetrahedron, the correspondence between the multiple target tetrahedrons in the target triangular mesh and the multiple color mapping relationships is queried to obtain the above-mentioned target color mapping relationship.

[0098] Next, in the Lab color space, the color information of the pixel points is mapped based on the target color mapping relationship.

[0099] In an embodiment of the present invention, for each pixel in the image to be color-matched, when changing the color of the pixel in the Lab color space, only the first color component a and the second color component b of the pixel can be changed, without modifying the brightness component L. Therefore, when mapping the color information of the pixel using the target color mapping relationship, the brightness component L of the pixel can be kept unchanged, and only the first color component a and the second color component b can be mapped. Thus, the adjusted first color component a' and the second color component b' can be obtained. Subsequently, the adjusted first color component a' and the second color component b' are merged with the original brightness component L to obtain the color information after the color change (for example, it can be expressed as [L, a', b']), and the color information after the color change is assigned to the pixel. The above operation is repeated until the color information of all pixels in the image to be color-matched is mapped, and the target image is obtained. Subsequently, the target image is converted from the Lab color space to the RGB color space. Based on the above method, the color matching process of the image to be color-matched can be completed.

[0100] FIG4 is a block diagram of a remote sensing image color uniformity device according to an embodiment of the present invention.

[0101] As shown in FIG. 4 , the remote sensing image color uniformity device 400 includes an acquisition module 410 , a first determination module 420 , a matching module 430 , a second determination module 440 and a color uniformity module 450 .

[0102] The acquisition module 410 is used to acquire reference instance images corresponding to a plurality of reference objects and an image to be color-homogenized.

[0103] The first determining module 420 is configured to determine, for each reference instance image, at least one target instance image in the image to be uniformed that is identical to the reference object in the reference instance image.

[0104] The matching module 430 is configured to perform control point matching between each reference instance image and at least one corresponding target instance image to obtain at least one control point pair.

[0105] The second determination module 440 is used to determine multiple color mapping relationships based on the color information of at least one control point pair corresponding to each of the multiple reference example images. The multiple color mapping relationships represent the corresponding relationship between the color information of each pixel point in the image to be uniformed before and after the color change.

[0106] The color uniformity module 450 is used to perform color uniformity processing on the image to be uniformed based on multiple color mapping relationships to obtain a target image.

[0107] According to an embodiment of the present invention, matching module 430 includes: a first determining unit and a second determining unit. The first determining unit is configured to determine at least one reference control point in a reference instance image; the second determining unit is configured to determine, for each reference control point, a color deviation between the reference control point and each pixel in the corresponding at least one target instance image, and based on the color deviation, determine a target control point corresponding to the reference control point from a plurality of pixels in the corresponding at least one target instance image, wherein the reference control point and the corresponding target control point constitute a control point pair.

[0108] According to an embodiment of the present invention, the color information of each control point pair includes the color information of the reference control point in the control point pair in the corresponding reference example image and the color information of the target control point in the control point pair in the corresponding at least one target example image. The second determination module 440 includes: a third determination unit, a construction unit, and a fourth determination unit. The third determination unit is configured to obtain, based on at least one control point pair corresponding to each of the multiple reference example images, a reference control point set including multiple reference control points, a target control point set including multiple target control points, and corresponding relationships between each reference control point and each target control point. The construction unit is configured to construct, using a Delaunay triangulation algorithm, a reference triangular mesh and a target triangular mesh based on the reference control point set and the target control point set, respectively. The fourth determination unit is configured to determine multiple color mapping relationships based on the reference triangular mesh, the target triangular mesh, the corresponding relationships between each reference control point and each target control point, the color information of each reference control point in the corresponding reference example image, and the color information of each target control point in the corresponding at least one target example image.

[0109] According to an embodiment of the present invention, the fourth determining unit includes: a first determining subunit, a second determining subunit, a third determining subunit, and a fourth determining subunit. The first determining subunit is used to determine, for any target tetrahedron in the target triangular mesh, a reference tetrahedron corresponding to the target tetrahedron in the reference triangular mesh based on the correspondence between each reference control point and each target control point; the second determining subunit is used to determine the color mapping relationship between the target tetrahedron and the corresponding reference tetrahedron based on the color information corresponding to each of the four target control points constituting the target tetrahedron, the color information corresponding to each of the four reference control points constituting the corresponding reference tetrahedron, and the correspondence between each reference control point and each target control point; the third determining subunit is used to repeatedly perform the above operations until the color mapping relationships between multiple target tetrahedrons in the target triangular mesh and the corresponding reference tetrahedrons in the reference triangular mesh are obtained; the fourth determining subunit is used to use the color mapping relationships between multiple target tetrahedrons in the target triangular mesh and the corresponding reference tetrahedrons in the reference triangular mesh as multiple color mapping relationships.

[0110] According to an embodiment of the present invention, the color leveling module 450 includes: a matching unit, a first mapping unit, and a second mapping unit. The matching unit is configured to determine, for any pixel in the image to be leveled, a target color mapping relationship that matches the color information of the pixel from multiple color mapping relationships based on the color information of the pixel; the first mapping unit is configured to map the color information of the pixel based on the target color mapping relationship; and the second mapping unit is configured to repeatedly perform the above operations until the color information of all pixels in the image to be leveled is mapped to obtain the target image.

[0111] According to an embodiment of the present invention, a matching unit includes: a first matching subunit and a second matching subunit. The first matching subunit is configured to determine, based on the color information of the pixel, a target tetrahedron that matches the color information of the pixel in the target triangular mesh; and the second matching subunit is configured to obtain a target color mapping relationship based on the target tetrahedron that matches the color information of the pixel and the correspondence between multiple target tetrahedrons and multiple color mapping relationships in the target triangular mesh.

[0112] According to an embodiment of the present invention, the second determination unit includes: a first clustering subunit, a fifth determination subunit, and a sixth determination subunit. The first clustering subunit is configured to, in response to determining that the number of pixel points corresponding to at least one target instance image whose color deviation satisfies a preset threshold is greater than or equal to the number threshold, cluster the pixel points whose color deviation satisfies the preset threshold according to the color information of each pixel point to obtain at least one first pixel point cluster; the fifth determination subunit is configured to select the first pixel point cluster with the largest number of pixels in the at least one first pixel point cluster as the target pixel point cluster; and the sixth determination subunit is configured to determine the cluster center of the target pixel point cluster as the target control point corresponding to the reference control point.

[0113] According to an embodiment of the present invention, the first determination unit includes: a second clustering subunit and an extraction subunit. The second clustering subunit is configured to cluster pixels in the reference example image based on color information of each pixel to obtain at least one second pixel cluster; and the extraction subunit is configured to extract one pixel from each of the at least one second pixel clusters to obtain at least one reference control point.

[0114] It should be noted that the implementation methods, technical problems solved, functions implemented, and technical effects achieved of each module in the device part embodiment are the same or similar to the implementation methods, technical problems solved, functions implemented, and technical effects achieved of each corresponding step in the method part embodiment, and will not be repeated here.

[0115] FIG5 schematically shows a block diagram of an electronic device suitable for implementing a remote sensing image color uniformity method according to an embodiment of the present invention.

[0116] As shown in Figure 5, the electronic device 500 according to an embodiment of the present invention includes a processor 501, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 502 or the program loaded from the storage part 508 into the random access memory (RAM) 503. The processor 501 may, for example, include a general-purpose microprocessor (such as a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (such as an application-specific integrated circuit (ASIC)), etc. The processor 501 may also include an onboard memory for caching purposes. The processor 501 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present invention.

[0117] The RAM 503 stores various programs and data required for the operation of the electronic device 500. The processor 501, ROM 502, and RAM 503 are connected to each other via a bus 504. The processor 501 executes the programs in the ROM 502 and / or RAM 503 to perform the various operations of the method flow according to the embodiment of the present invention. It should be noted that the programs may also be stored in one or more memories other than the ROM 502 and RAM 503. The processor 501 may also execute the programs stored in the one or more memories to perform the various operations of the method flow according to the embodiment of the present invention.

[0118] According to an embodiment of the present invention, electronic device 500 may further include an input / output (I / O) interface 505, which is also connected to bus 504. Electronic device 500 may further include one or more of the following components connected to I / O interface 505: an input portion 506 including a keyboard, a mouse, etc.; an output portion 507 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and speakers; a storage portion 508 including a hard disk; and a communication portion 509 including a network interface card such as a LAN card or a modem. Communication portion 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to I / O interface 505 as needed. Removable media 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed in drive 510 as needed, so that computer programs read from the removable media can be installed into storage portion 508 as needed.

[0119] The present invention also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments, or may exist independently and not incorporated into the device / apparatus / system. The computer-readable storage medium carries one or more programs, which, when executed, implement the method according to the embodiments of the present invention.

[0120] According to an embodiment of the present invention, a computer-readable storage medium may be a non-volatile computer-readable storage medium, such as, but not limited to, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in combination with an instruction execution system, apparatus, or device. For example, according to an embodiment of the present invention, a computer-readable storage medium may include the ROM 502 and / or RAM 503 described above and / or one or more memories other than ROM 502 and RAM 503.

[0121] Embodiments of the present invention also include a computer program product comprising a computer program containing program code for executing the method shown in the flowchart. When the computer program product is executed in a computer system, the program code causes the computer system to implement the remote sensing image color uniformity method provided by the embodiments of the present invention.

[0122] The computer program executes the above functions defined in the system / device of the embodiment of the present invention when the computer program is executed by the processor 501. According to the embodiment of the present invention, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0123] In one embodiment, the computer program may be stored on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may be transmitted and distributed in the form of a signal on a network medium, downloaded and installed via the communication portion 509, and / or installed from a removable medium 511. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to wireless, wired, or any suitable combination thereof.

[0124] In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 509 and / or installed from the removable medium 511. When the computer program is executed by the processor 501, the above-described functions defined in the system of the embodiment of the present invention are performed. According to the embodiment of the present invention, the systems, devices, means, modules, units, etc. described above can be implemented by computer program modules.

[0125] According to an embodiment of the present invention, the program code for executing the computer program provided by the embodiment of the present invention can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, Python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (e.g., using an Internet service provider to connect via the Internet).

[0126] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the above-mentioned module, program segment, or a part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0127] It will be understood by those skilled in the art that the features described in the various embodiments of the present invention may be combined and / or coupled in various ways, even if such combinations or couplings are not explicitly described in the present invention. In particular, the features described in the various embodiments of the present invention may be combined and / or coupled in various ways without departing from the spirit and teachings of the present invention. All such combinations and / or couplings fall within the scope of the present invention.

[0128] The above describes embodiments of the present invention. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of the present invention. Although each embodiment has been described separately above, this does not mean that the measures in each embodiment cannot be advantageously used in combination. Without departing from the scope of the present invention, those skilled in the art may make various substitutions and modifications, which should all fall within the scope of the present invention.

Claims

1. A remote sensing image color uniformity method, characterized in that: include: Acquire reference instance images and images to be color-homogenized corresponding to a plurality of reference objects; For each reference instance image, determining in the image to be uniformly colored at least one target instance image having the same reference object as the reference object in the reference instance image; For each reference instance image, performing control point matching between the reference instance image and at least one corresponding target instance image to obtain at least one control point pair; Based on the color information of at least one control point pair corresponding to each of the multiple reference example images, a plurality of color mapping relationships are determined, wherein the plurality of color mapping relationships represent the corresponding relationship between the color information of each pixel point in the image to be color-uniformed before and after the color change; The image to be color-homogenized is subjected to color-homogenization processing based on the multiple color mapping relationships to obtain a target image.

2. The method according to claim 1, characterized in that For each reference instance image, performing control point matching between the reference instance image and at least one corresponding target instance image to obtain at least one control point pair includes: determining at least one reference control point in the reference instance image; For each reference control point, determine the color deviation between the reference control point and each pixel in the corresponding at least one target instance image, and based on the color deviation, determine the target control point corresponding to the reference control point from a plurality of pixels contained in the corresponding at least one target instance image, wherein the reference control point and the corresponding target control point constitute a control point pair.

3. The method according to claim 2, characterized in that The color information of each control point pair includes the color information of the reference control point in the control point pair in the corresponding reference instance image and the color information of the target control point in the control point pair in the corresponding at least one target instance image; The determining of the plurality of color mapping relationships based on the color information of at least one control point pair corresponding to each of the plurality of reference instance images comprises: Based on at least one control point pair corresponding to each of the multiple reference instance images, a reference control point set including multiple reference control points, a target control point set including multiple target control points, and a correspondence relationship between each reference control point and each target control point is obtained; Using a Delaunay triangulation algorithm, based on the reference control point set and the target control point set, respectively, a reference triangular mesh and a target triangular mesh are constructed; and The multiple color mapping relationships are determined based on the reference triangular mesh, the target triangular mesh, the correspondence between each reference control point and each target control point, the color information of each reference control point in the corresponding reference instance image, and the color information of each target control point in the corresponding at least one target instance image.

4. The method according to claim 3, characterized in that The determining of the plurality of color mapping relationships based on the reference triangular mesh, the target triangular mesh, the correspondence between each reference control point and each target control point, the color information of each reference control point in the corresponding reference instance image, and the color information of each target control point in the corresponding at least one target instance image comprises: For any target tetrahedron in the target triangular mesh, based on the correspondence between each reference control point and each target control point, determine a reference tetrahedron corresponding to the target tetrahedron in the reference triangular mesh; Determine a color mapping relationship between the target tetrahedron and the corresponding reference tetrahedron according to color information corresponding to each of four target control points constituting the target tetrahedron, color information corresponding to each of four reference control points constituting the corresponding reference tetrahedron, and a correspondence between each reference control point and each target control point; Repeat the above operation until a color mapping relationship between a plurality of target tetrahedrons in the target triangular mesh and corresponding reference tetrahedrons in the reference triangular mesh is obtained; The color mapping relationships between the multiple target tetrahedrons in the target triangular mesh and the corresponding reference tetrahedrons in the reference triangular mesh are used as the multiple color mapping relationships.

5. The method according to claim 4, characterized in that The performing color uniformity processing on the image to be color uniformed based on the multiple color mapping relationships to obtain a target image comprises: For any pixel point in the image to be color-homogenized, based on the color information of the pixel point, determine a target color mapping relationship that matches the color information of the pixel point from the multiple color mapping relationships; Mapping the color information of the pixel points based on the target color mapping relationship; The above operation is repeatedly performed until the color information of all pixels in the image to be uniformly colored is mapped to obtain the target image.

6. The method according to claim 5, characterized in that The determining, based on the color information of the pixel point, from the multiple color mapping relationships, a target color mapping relationship that matches the color information of the pixel point comprises: Based on the color information of the pixel point, determining a target tetrahedron matching the color information of the pixel point in the target triangular mesh; The target color mapping relationship is obtained based on a target tetrahedron that matches the color information of the pixel point and a correspondence between a plurality of target tetrahedrons in the target triangular mesh and the plurality of color mapping relationships.

7. The method according to any one of claims 2 to 6, characterized in that The step of determining, based on the color deviation, a target control point corresponding to the reference control point from a plurality of pixel points included in the corresponding at least one target instance image comprises: In response to determining that the number of pixels whose color deviation satisfies a preset threshold in the corresponding at least one target instance image is greater than or equal to the number threshold, clustering the pixels whose color deviation satisfies the preset threshold according to color information of each pixel to obtain at least one first pixel cluster; Taking the first pixel point cluster with the largest number of pixels in at least one first pixel point cluster as the target pixel point cluster; The cluster center of the target pixel point cluster is determined as the target control point corresponding to the reference control point.

8. The method according to any one of claims 2 to 6, characterized in that Determining at least one reference control point in the reference instance image comprises: Clustering the pixels in the reference example image according to the color information of each pixel to obtain at least one second pixel cluster; A pixel point is extracted from each of the at least one second pixel point clusters to obtain the at least one reference control point.

9. A remote sensing image color homogenization device, characterized in that: include: An acquisition module is used to acquire reference instance images corresponding to a plurality of reference objects and an image to be uniformly colored; A first determination module is used to determine, for each reference instance image, at least one target instance image in the image to be uniformly colored that is identical to the reference object in the reference instance image; A matching module, configured to perform control point matching between each reference instance image and at least one corresponding target instance image to obtain at least one control point pair; A second determination module is used to determine a plurality of color mapping relationships based on the color information of at least one control point pair corresponding to each of the plurality of reference example images, wherein the plurality of color mapping relationships represent the corresponding relationship between the color information of each pixel point in the image to be color-uniformed before and after the color change; The color uniformity module is used to perform color uniformity processing on the image to be uniformed based on the multiple color mapping relationships to obtain a target image.

10. An electronic device, characterized in that: include: one or more processors; a memory for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors are enabled to execute the method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that: Executable instructions are stored thereon, and when the instructions are executed by a processor, the processor is caused to execute the method according to any one of claims 1 to 8.

12. A computer program product, characterized in that The invention comprises a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 8.

Images