Rapid color cast detection method and device for space remote sensing image and storage medium

Abstract

The invention provides a spaceflight remote sensing image-oriented color cast rapid detection method and device and a storage medium, and the method comprises the steps: converting the color space of a spaceflight remote sensing image into a Lab color space, and obtaining a conversion image; based on a preset channel threshold, creating a pixel mask, and performing pixel screening of an effective area on the converted image by using the pixel mask to obtain a screened image; the average chromaticity D and the chromaticity center distance M of the screened image on the a-b chromaticity plane are calculated, a first color cast coefficient and a second color cast coefficient are determined based on the ratio relation and the difference relation of the average chromaticity D and the chromaticity center distance M, and the color cast degree of the space remote sensing image is judged through the first color cast coefficient and the second color cast coefficient. The method is simple and efficient, the quality inspection task of massive space remote sensing images can be efficiently met, and the quality inspection efficiency and quality are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of aerospace remote sensing image quality inspection technology, and more specifically to a method, device and storage medium for rapid color deviation detection of aerospace remote sensing images. Background Technology

[0002] Due to complex factors such as fluctuations in space payload conditions and near-space electromagnetic interference, color distortion is a common phenomenon in space remote sensing images, affecting the quality and effectiveness of image products.

[0003] Existing methods for detecting color casts in aerospace remote sensing images can be mainly divided into three categories: statistical analysis-based methods, color model-based methods, and machine learning-based methods.

[0004] Statistical analysis-based methods primarily determine color cast by analyzing the pixel statistical characteristics of remote sensing images. These methods mainly include histogram analysis and analysis of mean and variance (ANOVA). Histogram analysis calculates the histograms for each color channel of the remote sensing image and checks for abnormal peaks or asymmetrical distributions to determine color cast. ANOVA calculates and compares the mean and variance of each color channel in the remote sensing image to determine color cast; significant differences indicate a potential color cast.

[0005] Color model-based methods mainly include the gray world assumption and the perfect reflectance assumption. The gray world assumption posits that in natural scenes, lighting effects cause remotely sensed images to appear grayish. Based on this assumption, the average value of each color channel in the remotely sensed image is calculated, and the actual average value of each color channel is compared with the target average value to calculate the deviation for each channel. A large deviation indicates a significant color cast in the remotely sensed image. The perfect reflectance assumption states that the reflection of light of all colors should be uniform. The average value of each color channel in the image and the proportion of each color channel relative to a certain channel are calculated. If a certain proportion deviates significantly from 1, it may indicate a color cast in the remotely sensed image.

[0006] Machine learning-based methods mainly include supervised learning, unsupervised learning, and deep learning. Specifically, analysis and discrimination are performed through feature learning, classification, and regression. Feature learning extracts features such as color, texture, and shape from spaceborne remote sensing images. Classification models are used to categorize remote sensing images into different color deviation categories, while regression models predict the specific values ​​of color deviations. The model parameters are optimized through the training process to accurately identify color deviations, thereby improving the color accuracy of the images.

[0007] Existing methods for detecting color cast in aerospace remote sensing images have been widely applied, but they all have certain limitations. Statistical analysis-based methods heavily rely on low-level pixel statistical features, leading to poor performance in non-uniform lighting, complex scenes, and bright backgrounds; for example, these methods tend to fail when the background of the remote sensing image is snow. Color model-based methods emphasize the theoretical assumptions of the color model and are only applicable when the assumption of "neutral colors" holds true. For instance, in marine observation images, blue accounts for a large proportion, making it difficult to determine the true color using the gray-scale world rule. Therefore, relying solely on the average chromaticity or the chromaticity of the maximum brightness values ​​in a remote sensing image to measure the degree of color cast is somewhat one-sided. Machine learning-based methods require large training datasets, feature extraction, and model training, resulting in high computational complexity and time costs.

[0008] Therefore, how to provide a simple, efficient, and effective design scheme for a rapid color distortion detection method, equipment, and storage medium that can improve the quality inspection efficiency of massive aerospace remote sensing images is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0009] In view of the above problems, this invention is proposed to provide a method, device and storage medium for rapid color deviation detection of aerospace remote sensing images that overcomes or at least partially solves the above problems. It is simple and efficient, and can efficiently meet the quality inspection tasks of massive aerospace remote sensing images, effectively improving the efficiency and quality of quality inspection. It has a wide range of applications in fields such as agricultural and forestry protection and land monitoring.

[0010] To achieve the above objectives, the present invention adopts the following technical solution:

[0011] In a first aspect, embodiments of the present invention provide a method for rapid color cast detection in aerospace remote sensing images, comprising the following steps: S1: Convert the color space of the aerospace remote sensing image to Lab color space to obtain the converted image; S2: Based on a preset channel threshold, create a pixel mask, and use the pixel mask to filter the effective region of the converted image to obtain the filtered image; S3: Calculate the average chromaticity D and chromaticity center distance M of the filtered image on the ab chromaticity plane, and determine the first color cast coefficient and the second color cast coefficient based on the ratio and difference between the average chromaticity D and the chromaticity center distance M, and use the first color cast coefficient and the second color cast coefficient to determine the degree of color cast of the aerospace remote sensing image.

[0012] Preferably, the space remote sensing image in S1 is a browsing image generated after compression of the original space remote sensing image. The file size of the browsing image is smaller than that of the original space remote sensing image, and it retains the overall color characteristics of the original space remote sensing image.

[0013] Preferably, step S1 further includes a size normalization preprocessing step for the spaceborne remote sensing image: Identify the dimensions of the aerospace remote sensing image and obtain the side with the largest dimension; While scaling the edges to a preset size, the aspect ratio of the aerospace remote sensing image remains unchanged, resulting in a preprocessed aerospace remote sensing image that participates in the color space conversion step.

[0014] Preferably, step S2 includes the following steps: Based on the preset channel threshold of the brightness L channel, a pixel mask is created. The pixel mask is then used to filter the a and b channels of the aerospace remote sensing image, retaining only the a and b channel information of the valid pixel positions identified by the pixel mask, in order to obtain the filtered image.

[0015] Preferably, the average chromaticity D is calculated using the following formula:

[0016]

[0017]

[0018] in, These represent the chromaticity offset of pixels in aerospace remote sensing images along the a-axis and b-axis, respectively. W , H These represent the width and height of the spaceborne remote sensing image, respectively.

[0019] Preferably, the chromaticity center distance M is calculated using the following formula:

[0020]

[0021]

[0022]

[0023]

[0024] in, These represent the chromaticity offset of pixels in aerospace remote sensing images along the a-axis and b-axis, respectively. W , H These represent the width and height of the spaceborne remote sensing image, respectively.

[0025] Preferably, in S3: The first color deviation coefficient K is determined based on the ratio of the average chromaticity D to the chromaticity center distance M. The second color cast coefficient G is determined based on the difference between the average chromaticity D and the chromaticity center distance M. The degree of color cast in the aerospace remote sensing image is negatively correlated with the first color cast coefficient K and positively correlated with the second color cast coefficient G.

[0026] Preferably, the step of determining the degree of color cast in the spaceborne remote sensing image using the first color cast coefficient and the second color cast coefficient includes: In response to determining that the first color cast coefficient K is less than a first threshold and the second color cast coefficient G is greater than a second threshold, it is determined that the aerospace remote sensing image has a color cast.

[0027] In a second aspect, embodiments of the present invention provide an electronic device, comprising: One or more processors; Memory; And one or more computer programs, wherein the one or more computer programs are stored in the memory and configured to be executed by the one or more processors, the programs including instructions for performing the color cast detection method for spaceborne remote sensing imagery.

[0028] Thirdly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, characterized in that the computer program, when executed by a processor, implements the method for rapid color distortion detection of aerospace remote sensing images.

[0029] The beneficial effects of the above-described technical solutions provided in the embodiments of the present invention include at least the following: This invention realizes a color cast detection technology that is highly applicable to aerospace remote sensing images. It maps the color distribution characteristics of aerospace remote sensing images onto the equivalent circle parameter of the a-b color coordinate plane, which can better reflect the overall color cast of remote sensing images. It avoids the limitations of simply relying on pixel statistical features for analysis, and has low requirements for hardware computing power, effectively saving the time cost of collecting and organizing datasets and training models.

[0030] This invention introduces two types of color cast coefficients in the calculation and discrimination of color cast coefficients, and innovatively constructs rapid color cast detection conditions for aerospace remote sensing images, which effectively improves the accuracy of the detection results.

[0031] This invention features intuitive principles, high computational efficiency, low computational cost, and wide application scenarios, and can meet the real-time needs of rapid quality inspection of large-scale aerospace remote sensing image data. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0033] Figure 1 This is a flowchart of a method for rapid color shift detection of aerospace remote sensing images provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of the equivalent circle of the a-b chromaticity coordinate plane provided in the embodiments of the present invention. Detailed Implementation

[0034] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0035] This invention discloses a rapid color cast detection method for aerospace remote sensing images, comprising the following steps: S1: Convert the color space of the aerospace remote sensing image to Lab color space to obtain the converted image; S2: Based on the preset channel threshold, create a pixel mask, and use the pixel mask to filter the effective area of ​​the converted image to obtain the filtered image; S3: Calculate the average chromaticity D and chromaticity center distance M of the filtered image on the ab chromaticity plane, and determine the first and second color cast coefficients based on the ratio and difference between the average chromaticity D and the chromaticity center distance M. Use the first and second color cast coefficients to determine the degree of color cast of the aerospace remote sensing image.

[0036] It should be noted that the Lab color space is designed based on the human visual perception model, which can more accurately reflect color differences. The Lab color space includes luminance (L) and two color components (a and b). The a channel represents the color change from green to red, and the b channel represents the color change from blue to yellow. This separation method gives the Lab space great advantages in color adjustment, color matching, and image processing, and it is suitable for precise color matching and analysis.

[0037] Based on the concept of the equivalent circle, this invention comprehensively considers the relative relationship and absolute difference between the average chromaticity D and the chromaticity center distance M of the image, thereby introducing two color cast coefficients and innovatively constructing color cast discrimination conditions for aerospace remote sensing images, which can significantly reduce the probability of misjudgment and missed detection during image quality inspection.

[0038] In one embodiment, commonly used aerospace remote sensing images are typically in TIFF format. Because this format combines original resolution image data with geospatial information, it occupies a large amount of space; for example, a single standard scene image typically ranges in size from several hundred MB to tens of GB. The aerospace remote sensing image in embodiment S1 is a view image generated by compressing the original aerospace remote sensing image. The view image's file size is smaller than the original aerospace remote sensing image's file size, and it retains the overall color characteristics of the original aerospace remote sensing image.

[0039] Spaceborne remote sensing image browsing maps are image files generated from original high-resolution images through efficient compression. They are commonly in PNG format, with file sizes ranging from tens of KB to several MB. Although spatial resolution is lost, the overall information of the original remote sensing image is preserved. Since color cast is a characteristic of spaceborne remote sensing images as a whole, and considering the efficiency of image file storage, transmission, retrieval, and computation, this embodiment uses spaceborne remote sensing image browsing maps as the basic input for analysis, processing, and computational analysis.

[0040] In one embodiment, the multi-source aerospace remote sensing image browsing maps have different resolutions and sizes. To facilitate unified calculation and analysis, the browsing maps need to undergo size standardization preprocessing. This embodiment S1 also includes a size standardization preprocessing step for the aerospace remote sensing images: Identify the dimensions of aerospace remote sensing images and obtain the side with the largest dimension; While scaling the edges to a preset size, the aspect ratio of the aerospace remote sensing image remains unchanged, resulting in a preprocessed aerospace remote sensing image that participates in the color space conversion step.

[0041] A preprocessing method that unifies the maximum side length and maintains the aspect ratio is adopted. This method fixes the image size by keeping the size of the maximum side of all images consistent. On this basis, it maintains the original proportion of the input browsing image without stretching or compressing the image, effectively avoiding problems such as browsing image distortion, excessive scaling and loss of image information.

[0042] In one embodiment, the color space of aerospace remote sensing images is typically RGB. However, RGB color space has drawbacks such as being unintuitive, non-uniform, and device-dependent. Converting aerospace remote sensing image browsing maps from RGB to Lab color space involves three steps: The first step is to convert from RGB color space to sRGB color space. sRGB color space is a standard RGB color space that ensures consistency and compatibility of image colors across different devices, allowing images to be displayed more accurately on various screens.

[0043] Step 2 converts the sRGB space to the XYZ space through linear transformation. The XYZ space is based on the human eye's perception of color and does not depend on any specific display device or light source. The representation of colors in this space is more objective and accurate.

[0044] Step 3 converts the XYZ space to the Lab space.

[0045] In one embodiment, S2 includes the following steps: Based on the preset channel threshold of the brightness L channel, a pixel mask is created. The pixel mask is then used to filter the a and b channels of the aerospace remote sensing image, retaining only the a and b channel information of the valid pixel positions identified by the pixel mask, in order to obtain the filtered image.

[0046] In this embodiment, the aerospace remote sensing image converted to the Lab color space is separated into three components (L, a, and b channels). To ensure the accuracy of subsequent processing, a pixel mask is created to filter and mark those pixels with practical significance. Specifically, pixels with an L (luminance) channel value greater than 0 are selected, as these pixels typically represent valid areas in the image, while pixels with an L value of 0 often correspond to background or invalid areas, which is particularly common in aerospace remote sensing images, such as clouds, sensor noise, or pure black backgrounds. This mask is used to filter the a and b channels, retaining only the color information corresponding to valid pixels. This filters valid pixels and excludes background and noise, ensuring that subsequent image calculations and analyses are based only on real and meaningful pixel data, thus improving the reliability and accuracy of the results.

[0047] In one embodiment, the average chromaticity D is calculated using the following formula:

[0048]

[0049]

[0050] in, These represent the chromaticity offset of pixels in aerospace remote sensing images along the a-axis and b-axis, respectively. W , H These represent the width and height of the spaceborne remote sensing image, respectively. W Indicates the number of horizontal pixels in an image. HThis indicates the number of vertical pixels in the image.

[0051] In one embodiment, the chromaticity center distance M is calculated using the following formula:

[0052]

[0053]

[0054]

[0055]

[0056] in, These represent the chromaticity offset of pixels in aerospace remote sensing images along the a-axis and b-axis, respectively. W , H These represent the width and height of the spaceborne remote sensing image, respectively. W Indicates the number of horizontal pixels in an image. H This indicates the number of vertical pixels in the image.

[0057] It should be noted that, as Figure 2 As shown, on the ab chromaticity plane, the radius of the equivalent circle is... The chromaticity distribution is characterized by its discreteness, reflecting the distribution range of pixels in the image on the ab chromaticity plane. The distance from the center of the equivalent circle to the origin of the neutral axis of the ab chromaticity plane (a=0, b=0) is... It reflects the overall direction of color shift.

[0058] In one embodiment, S3: The first color deviation coefficient K is determined based on the ratio of average chromaticity D to chromaticity center distance M: ; The second color cast coefficient G is determined based on the relationship between the average chromaticity D and the chromaticity center distance M: ; The degree of color cast in aerospace remote sensing images is negatively correlated with the first color cast coefficient K and positively correlated with the second color cast coefficient G.

[0059] In this embodiment, the step of determining the degree of color cast in a space remote sensing image using the first color cast coefficient and the second color cast coefficient includes: In response to determining that the first color cast coefficient K is less than the first threshold and the second color cast coefficient G is greater than the second threshold, it is determined that the space remote sensing image has a color cast.

[0060] It should be noted that in a two-dimensional histogram constructed using the ab chromaticity coordinate plane, if the image's chromaticity distribution exhibits a single peak or relatively concentrated characteristics, it indicates that the image's chromaticity values ​​are concentrated within a specific range. This concentration may suggest that the image's colors are biased towards a particular hue. Conversely, a larger average chromaticity value usually indicates the presence of a color cast, and the larger the average chromaticity value, the more severe the color cast.

[0061] Conversely, in a two-dimensional histogram constructed from the ab chromaticity coordinate plane, if the image's chromaticity distribution exhibits obvious multi-peak values ​​and is relatively dispersed, it indicates that the image's chromaticity values ​​are distributed across multiple different ranges. This multi-peak and dispersion suggests that the image has rich colors, or that the chromaticity values ​​are evenly distributed across different hues. In this case, the image's color cast is significantly reduced, or there may even be no color cast at all.

[0062] The overall color cast of aerospace remote sensing images can be determined by the position of the equivalent circle on the AB chromaticity plane. >0 indicates a reddish tint; otherwise, a greenish tint. >0 indicates a yellowish tint; otherwise, a bluish tint.

[0063] If the average chromaticity D of an image is significantly greater than the chromaticity center distance M, it indicates that the overall image shift is stronger than natural dispersion, resulting in severe color cast. Comparing the magnitudes of D and M can be done by comparing their relative values ​​or their absolute differences. Specifically, this involves comparing and subtracting them to obtain the first color cast coefficient K and the second color cast coefficient G. To improve the accuracy of the discrimination results, it is necessary to consider both the relative magnitude and the absolute difference between the image's average chromaticity D and the chromaticity center distance M. Therefore, a color cast discrimination condition for aerospace remote sensing imagery is constructed: the smaller the first color cast coefficient K and the larger the second color cast coefficient G, the more severe the image color cast.

[0064] Based on the same inventive concept, embodiments of the present invention also provide an electronic device, including: one or more processors; a memory; and one or more computer programs, wherein the one or more computer programs are stored in the memory and configured to be executed by the one or more processors, the programs including instructions for executing the rapid color cast detection method for spaceborne remote sensing images. Since the principle by which the electronic device solves the problem is similar to the aforementioned front-end architecture, the implementation of this electronic device can refer to the implementation of the aforementioned method, and repeated details will not be elaborated further.

[0065] Based on the same inventive concept, embodiments of the present invention also provide a computer-readable storage medium storing a computer program, characterized in that, when the computer program is executed by a processor, it implements the method for rapid color distortion detection of aerospace remote sensing images. Since the principle by which the computer-readable storage medium solves the problem is similar to the aforementioned front-end architecture, the implementation of this computer-readable storage medium can refer to the implementation of the aforementioned method, and repeated details will not be elaborated further.

[0066] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0067] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A rapid color cast detection method for aerospace remote sensing images, characterized in that, Includes the following steps: S1: Convert the color space of the aerospace remote sensing image to Lab color space to obtain the converted image; S2: Based on a preset channel threshold, create a pixel mask, and use the pixel mask to filter the effective region of the converted image to obtain the filtered image; S3: Calculate the average chromaticity D and chromaticity center distance M of the filtered image on the ab chromaticity plane, and determine the first color cast coefficient and the second color cast coefficient based on the ratio and difference between the average chromaticity D and the chromaticity center distance M, and use the first color cast coefficient and the second color cast coefficient to determine the degree of color cast of the aerospace remote sensing image.

2. The method for rapid color cast detection in aerospace remote sensing images as described in claim 1, characterized in that, The space remote sensing image of S1 is a browsing image generated after the original space remote sensing image is compressed. The file data size of the browsing image is smaller than that of the original space remote sensing image, and the overall color characteristics of the original space remote sensing image are preserved.

3. The method for rapid color cast detection in aerospace remote sensing images as described in claim 1, characterized in that, S1 also includes a size normalization preprocessing step for space remote sensing images: Identify the dimensions of the aerospace remote sensing image and obtain the side with the largest dimension; While scaling the edges to a preset size, the aspect ratio of the aerospace remote sensing image remains unchanged, resulting in a preprocessed aerospace remote sensing image that participates in the color space conversion step.

4. The method for rapid color cast detection in aerospace remote sensing images as described in claim 1, characterized in that, S2 includes the following steps: Based on the preset channel threshold of the brightness L channel, a pixel mask is created. The pixel mask is then used to filter the a and b channels of the aerospace remote sensing image, retaining only the a and b channel information of the valid pixel positions identified by the pixel mask, in order to obtain the filtered image.

5. The method for rapid color cast detection in aerospace remote sensing images as described in claim 1, characterized in that, The average chromaticity D is calculated using the following formula: ； ； ； in, These represent the chromaticity offset of pixels in aerospace remote sensing images along the a-axis and b-axis, respectively. W , H These represent the width and height of the spaceborne remote sensing image, respectively.

6. The method for rapid color cast detection in aerospace remote sensing images as described in claim 1, characterized in that, The chromaticity center distance M is calculated using the following formula: ； ； ； ； ； in, These represent the chromaticity offset of pixels in aerospace remote sensing images along the a-axis and b-axis, respectively. W , H These represent the width and height of the spaceborne remote sensing image, respectively.

7. The method for rapid color cast detection in aerospace remote sensing images as described in claim 1, characterized in that, In S3: The first color deviation coefficient K is determined based on the ratio of the average chromaticity D to the chromaticity center distance M. The second color cast coefficient G is determined based on the difference between the average chromaticity D and the chromaticity center distance M. The degree of color cast in the aerospace remote sensing image is negatively correlated with the first color cast coefficient K and positively correlated with the second color cast coefficient G.

8. The method for rapid color cast detection in aerospace remote sensing images as described in claim 7, characterized in that, The steps for determining the degree of color cast in the spaceborne remote sensing image using a first color cast coefficient and a second color cast coefficient include: In response to determining that the first color cast coefficient K is less than a first threshold and the second color cast coefficient G is greater than a second threshold, it is determined that the aerospace remote sensing image has a color cast.

9. An electronic device, comprising: One or more processors; Memory; And one or more computer programs, wherein the one or more computer programs are stored in the memory and configured to be executed by the one or more processors, the programs including instructions for performing the rapid color distortion detection method for spaceborne remote sensing imagery as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the rapid color distortion detection method for aerospace remote sensing images as described in any one of claims 1 to 8.

Images