A method and device for acquiring equivalent low-orbit satellite remote sensing images

Abstract

The present invention relates to a method for acquiring an equivalent low-orbit satellite remote sensing image, comprising: S1: downsampling a drone aerial image according to a resolution difference parameter to obtain a downsampled aerial image; S2: performing color conversion on the downsampled aerial image according to a color difference parameter to obtain an equivalent remote sensing image of a low-orbit satellite. The method for acquiring an equivalent low-orbit satellite remote sensing image of the present invention accurately adjusts the resolution of the drone aerial image through the resolution difference parameter so that it is consistent with the low-orbit satellite remote sensing image in terms of spatial resolution. At the same time, the hue, saturation and brightness of the drone aerial image are adjusted through the color difference parameter to ensure that the color features are consistent with the low-orbit satellite remote sensing image, thereby achieving a high degree of simulation of the conversion of the drone aerial image to the low-orbit satellite remote sensing image, thereby solving the problem of poor conversion effect of the drone aerial image in complex environments.

Description

Technical Field

[0001] The present invention relates to the field of satellite remote sensing image processing, and in particular to a method, device, electronic equipment and computer storage medium for acquiring an equivalent low-orbit satellite remote sensing image. Background Art

[0002] In recent years, drone aerial images have become a high-resolution, real-time updated, low-cost source of surface information, and are widely used in visual navigation, target detection, remote sensing mapping, object recognition, disaster prediction, and landform change detection. Based on this, since the remote sensing images of low-orbit satellites are easily blocked by meteorological conditions, such as cloud cover that produces partial blind spots, the existing technology combines drone aerial images to complement the drone aerial images with the remote sensing images of low-orbit satellites, so that the complementary aerial images can be used to supplement more blind area terrain information in the original low-orbit satellite remote sensing images.

[0003] However, if low-orbit satellite remote sensing images need to be used for deep learning, for example, if deep learning models need to be used to identify the enemy's geographic information, weapons and equipment, atmospheric environment, remote sensing information and other fields in low-orbit satellite remote sensing images, there are significant differences in resolution, perspective, lighting, color, terrain and other aspects between drone aerial images and low-orbit satellite remote sensing images, resulting in the inability to directly use training data for deep learning. In addition, because actual business applications mainly use low-orbit satellite remote sensing images, inconsistent training data can easily lead to poor application results.

[0004] Based on this, in the existing heterogeneous remote sensing image simulation conversion technology, deep learning is generally used to extract features and transfer styles of drone aerial images, so that drone aerial images are close to or equivalent to remote sensing images from low-orbit satellites, and more high-quality data sources are obtained for application in the field of heterogeneous remote sensing image simulation conversion. However, the deep learning model for equivalent conversion of remote sensing images requires a large amount of training data, and has limited generalization ability in scenes with large differences, resulting in poor conversion effects of drone aerial images in complex environments. Summary of the invention

[0005] Based on this, the object of the present invention is to provide a method for acquiring equivalent low-orbit satellite remote sensing images.

[0006] A method for acquiring an equivalent low-orbit satellite remote sensing image comprises the following steps:

[0007] S1: downsampling the aerial image according to the resolution difference parameter to obtain a downsampled aerial image;

[0008] Wherein, the resolution difference parameter is the difference in pixel density;

[0009] S2: Perform color conversion on the downsampled aerial image according to the color difference parameter to obtain an equivalent remote sensing image of the low-orbit satellite;

[0010] The color difference parameters include a hue parameter ΔH, a saturation parameter ΔS, and a brightness parameter ΔV.

[0011] The method for acquiring equivalent low-orbit satellite remote sensing images described in the present invention accurately adjusts the resolution of drone aerial images through resolution difference parameters so that the spatial resolution is consistent with that of low-orbit satellite remote sensing images. At the same time, the hue, saturation and brightness of the drone aerial images are adjusted through color difference parameters to ensure that the color features are consistent with the low-orbit satellite remote sensing images, thereby achieving a high degree of simulation of the conversion of drone aerial images into low-orbit satellite remote sensing images, thereby solving the problem of poor conversion effect of drone aerial images in complex environments.

[0012] Furthermore, the steps of obtaining the resolution difference parameter are as follows:

[0013] The steps for obtaining the resolution difference parameter are as follows:

[0014] R1: Obtain the aerial image and its corresponding camera and other sensor parameters;

[0015] The sensor parameters of the camera device include internal reference information, external reference information and geographic location information at the time of imaging;

[0016] R2: According to the orbital parameters of the low-orbit satellite, the remote sensing images passing overhead, the sensor parameters such as the shooting device, the direction and scale of the aerial images are registered to obtain the registered aerial images;

[0017] R3: Roughly match the registered aerial image with the low-orbit satellite remote sensing image to obtain several low-orbit satellite remote sensing sub-images;

[0018] R4: Calculate the pixel density of the registered aerial image and the current low-orbit satellite remote sensing sub-image respectively;

[0019] R5: Calculate the difference between the pixel density of the aerial image and the pixel density of the current low-orbit satellite remote sensing sub-image to obtain the resolution difference parameter;

[0020] The resolution difference parameter ΔR is calculated as follows:

[0021] ΔR=|Pixel S -Pixel U |

[0022] Among them, Pixel S is the pixel density of the low-orbit satellite remote sensing sub-image, Pixel Uis the pixel density of the registered aerial image.

[0023] The present invention adjusts the spatial geometry of the drone aerial image and the low-orbit satellite remote sensing image by combining the sensor internal parameters such as the shooting device of the drone aerial image and the sensor external parameters such as the shooting device, thereby ensuring the accuracy of the pixel density calculation. At the same time, the present invention quantifies the resolution difference parameter through the pixel density to ensure that the resolution of the drone aerial image is consistent with the resolution of the low-orbit satellite remote sensing image after the resolution conversion.

[0024] Furthermore, after obtaining the resolution difference parameter, the following steps are also included:

[0025] R6: downsample the aerial image according to the resolution difference parameter to obtain a downsampled aerial image;

[0026] R7: Perform hash conversion and similarity calculation on the downsampled aerial image and the low-orbit satellite remote sensing sub-image, and determine whether the similarity is less than a preset threshold: if yes, use the current resolution difference parameter as the standard resolution difference parameter; if not, use another low-orbit satellite remote sensing sub-image as input and execute step R4.

[0027] Furthermore, the step R7 specifically includes the following sub-steps:

[0028] By graying the downsampled aerial image and the current low-orbit satellite remote sensing sub-image, and performing hash calculation on the gray-scale image through a hash algorithm, the image fingerprints of the downsampled aerial image and the current low-orbit satellite remote sensing sub-image are obtained respectively;

[0029] Next, the Hamming distance between the downsampled aerial image and the image fingerprint of the current low-orbit satellite remote sensing sub-image is calculated, and it is determined whether the Hamming distance is less than a preset threshold: if yes, the current resolution difference parameter is used as the standard resolution difference parameter; if no, another low-orbit satellite remote sensing sub-image is used as the current low-orbit satellite remote sensing sub-image, and step R4 is executed;

[0030] The Hamming distance is calculated based on the similarity of a 16-bit hexadecimal string converted from a 64-bit binary hash value.

[0031] The present invention performs Hamming distance calculation on the image fingerprint obtained through hash calculation, and uses a preset threshold to ensure that the selected resolution difference parameter can accurately reflect the resolution difference between the downsampled aerial image and the low-orbit satellite remote sensing subimage, optimize the downsampling effect, and enable the resolution difference parameter to retain the detail information in the drone aerial image to the maximum extent.

[0032] Furthermore, the steps of obtaining the color difference parameters are as follows:

[0033] C1: Perform color space conversion on the registered aerial images and low-orbit satellite remote sensing sub-images to obtain the corresponding hue, saturation and brightness;

[0034] C2: Calculate the difference parameters of hue, saturation and brightness of the registered aerial image and the low-orbit satellite remote sensing sub-image to obtain the color difference parameters;

[0035] C3: Correct the registered aerial image by using the color difference parameter to obtain the equivalent remote sensing image of the low-orbit satellite to be confirmed;

[0036] C4: Calculate the similarity score between the equivalent remote sensing image of the low-orbit satellite to be confirmed and the matching low-orbit satellite remote sensing sub-image, and determine whether it is less than a similarity threshold. If yes, use the current color difference parameter as the standard color difference parameter; if not, obtain another matching low-orbit satellite remote sensing sub-image and execute step C1.

[0037] Furthermore, the similarity score is calculated as follows:

[0038] Calculate the histograms of hue and saturation corresponding to the equivalent remote sensing image of the low-orbit satellite to be confirmed and the remote sensing sub-image of the matching low-orbit satellite respectively;

[0039] The histograms are normalized and compared using the Bhattacharyya distance to obtain the similarity score, which is calculated as follows:

[0040]

[0041] Among them, Distance B (H1, H2) is the similarity score, the smaller the value, the higher the similarity; H1(i) is the value of the normalized hue-saturation histogram of the equivalent remote sensing image of the low-orbit satellite to be confirmed in the i-th array, i.e., BIN; H2(i) is the value of the normalized hue-saturation histogram of the matching low-orbit satellite remote sensing sub-image in the i-th BIN.

[0042] An acquisition device for equivalent low-orbit satellite remote sensing images, comprising a resolution conversion unit and a color space conversion unit;

[0043] The resolution conversion unit is used to downsample the aerial image according to the resolution difference parameter to obtain a downsampled aerial image; wherein the resolution difference parameter is a difference in pixel density;

[0044] The color space conversion unit is used to perform color conversion on the downsampled aerial image according to the color difference parameters to obtain an equivalent remote sensing image of the low-orbit satellite; wherein the color difference parameters include a hue parameter ΔH, a saturation parameter ΔS, and a brightness parameter ΔV.

[0045] Furthermore, the steps of obtaining the resolution difference parameter are as follows:

[0046] The steps for obtaining the resolution difference parameter are as follows:

[0047] R1: Obtain the aerial image and its corresponding camera and other sensor parameters;

[0048] The sensor parameters of the camera device include internal reference information, external reference information and geographic location information at the time of imaging;

[0049] R2: According to the orbital parameters of the low-orbit satellite, the remote sensing images passing overhead, the sensor parameters such as the shooting device, the direction and scale of the aerial images are registered to obtain the registered aerial images;

[0050] R3: Roughly match the registered aerial image with the low-orbit satellite remote sensing image to obtain several low-orbit satellite remote sensing sub-images;

[0051] R4: Calculate the pixel density of the registered aerial image and the current low-orbit satellite remote sensing sub-image respectively;

[0052] R5: Calculate the difference between the pixel density of the aerial image and the pixel density of the current low-orbit satellite remote sensing sub-image to obtain the resolution difference parameter;

[0053] The resolution difference parameter ΔR is calculated as follows:

[0054] ΔR=|Pixel S -Pixel U |

[0055] Among them, Pixel S is the pixel density of the low-orbit satellite remote sensing sub-image, Pixel U is the pixel density of the registered aerial image;

[0056] R6: downsample the aerial image according to the resolution difference parameter to obtain a downsampled aerial image;

[0057] R7: Perform hash conversion and similarity calculation on the downsampled aerial image and the low-orbit satellite remote sensing sub-image, and determine whether the similarity is less than a preset threshold: if yes, use the current resolution difference parameter as the standard resolution difference parameter; if no, use another low-orbit satellite remote sensing sub-image as input and execute step R4;

[0058] Wherein, the step R7 specifically includes the following sub-steps:

[0059] By graying the downsampled aerial image and the current low-orbit satellite remote sensing sub-image, and performing hash calculation on the gray-scale image through a hash algorithm, the image fingerprints of the downsampled aerial image and the current low-orbit satellite remote sensing sub-image are obtained respectively;

[0060] Next, the Hamming distance between the downsampled aerial image and the image fingerprint of the current low-orbit satellite remote sensing sub-image is calculated, and it is determined whether the Hamming distance is less than a preset threshold: if yes, the current resolution difference parameter is used as the standard resolution difference parameter; if no, another low-orbit satellite remote sensing sub-image is used as the current low-orbit satellite remote sensing sub-image, and step R4 is executed;

[0061] The Hamming distance is calculated based on the 16-bit hexadecimal string converted from the 64-bit binary hash value.

[0062] The steps for obtaining the color difference parameters are as follows:

[0063] C1: Perform color space conversion on the registered aerial images and low-orbit satellite remote sensing sub-images to obtain the corresponding hue, saturation and brightness;

[0064] C2: Calculate the difference parameters of hue, saturation and brightness of the registered aerial image and the low-orbit satellite remote sensing sub-image to obtain the color difference parameters;

[0065] C3: Correct the registered aerial image by using the color difference parameter to obtain the equivalent remote sensing image of the low-orbit satellite to be confirmed;

[0066] C4: Calculate the similarity score between the equivalent remote sensing image of the low-orbit satellite to be confirmed and the matching low-orbit satellite remote sensing sub-image, and determine whether it is less than a similarity threshold. If yes, use the current color difference parameter as the standard color difference parameter; if no, obtain another matching low-orbit satellite remote sensing sub-image and execute step C1;

[0067] The similarity score is calculated as follows:

[0068] Calculate the histograms of hue and saturation corresponding to the equivalent remote sensing image of the low-orbit satellite to be confirmed and the remote sensing sub-image of the matching low-orbit satellite respectively;

[0069] The histograms are normalized and compared using the Bhattacharyya distance to obtain the similarity score, which is calculated as follows:

[0070]

[0071] Among them, Distance B(H1, H2) is the similarity score, the smaller the value, the higher the similarity; H1(i) is the value of the normalized hue-saturation histogram of the equivalent remote sensing image of the low-orbit satellite to be confirmed in the i-th array, i.e., BIN; H2(i) is the value of the normalized hue-saturation histogram of the matching low-orbit satellite remote sensing sub-image in the i-th BIN.

[0072] For better understanding and implementation, the present invention is described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 A simplified structural diagram of a device for acquiring equivalent low-orbit satellite remote sensing images;

[0074] Figure 2 The figure is a simplified flowchart of a method for acquiring an equivalent low-orbit satellite remote sensing image;

[0075] Figure 3 A schematic diagram of the process of obtaining resolution difference parameters;

[0076] Figure 4 A simplified schematic diagram of the correction process of the aerial image for registration;

[0077] Figure 5 A simple schematic diagram for calculating the similarity of image fingerprints;

[0078] Figure 6 Schematic diagram of the process of obtaining color difference parameters;

[0079] Figure 7 The drone aerial image to be converted;

[0080] Figure 8 It is a low-orbit satellite remote sensing image of visible light type;

[0081] Fig. 9 UAV aerial images for equivalent conversion;

[0082] Fig.10 This is a schematic diagram of the shooting position of the drone aerial image;

[0083] Fig.11 for Figure 7 , Figure 8 and Fig. 9 A specific comparison diagram of hue, saturation and brightness. DETAILED DESCRIPTION

[0084] In order to solve the problem of poor effect of converting drone aerial images to low-orbit satellite remote sensing images through deep learning in the prior art, the present invention performs difference analysis and color analysis on drone aerial images and low-orbit satellite remote sensing images to obtain resolution difference parameters and color difference parameters. Then, the drone aerial images are downsampled based on the resolution difference parameters, and the downsampled drone aerial images are converted into color space based on the color difference parameters to obtain an equivalent remote sensing image of the low-orbit satellite. Accordingly, the present invention characterizes the conversion relationship between the drone aerial image and the low-orbit satellite remote sensing image through resolution difference parameters and color difference parameters, thereby realizing low-cost, efficient and accurate conversion of drone aerial images to low-orbit satellite remote sensing images.

[0085] Based on the above design, the present invention proposes a method for acquiring equivalent low-orbit satellite remote sensing images, and based on this method, proposes a device for acquiring equivalent low-orbit satellite remote sensing images.

[0086] Please also see Figure 1 and Figure 2 , Figure 1 This is a simplified structural diagram of a device for acquiring equivalent low-orbit satellite remote sensing images. Figure 2 The figure is a simplified flowchart of the method for acquiring equivalent low-orbit satellite remote sensing images.

[0087] The device for acquiring equivalent low-orbit satellite remote sensing images comprises a resolution conversion unit 1 and a color space conversion unit 2.

[0088] The resolution conversion unit 1 is used to execute step S1: downsampling the drone aerial image according to the resolution difference parameter to obtain a downsampled aerial image.

[0089] Specifically, the corresponding scaling factor is obtained according to the resolution difference parameter, as follows:

[0090]

[0091] Among them, s is the scaling factor and ΔR is the resolution difference parameter. Next, the target size of the current drone aerial image is obtained. The specific calculation is as follows:

[0092] H target =H ori ×s

[0093] W target =W ori ×s

[0094] Among them, H target and W target are the height and width of the target size, H ori and W oriare the height and width of the current drone aerial image, and s is the scaling factor.

[0095] Next, bilinear interpolation is performed on all pixel coordinates in the target size according to the current drone aerial image to obtain the downsampled aerial image. The specific calculation is as follows:

[0096]

[0097] Where f(x,y) is the pixel value corresponding to the pixel coordinate (x,y) in the target size, and the pixels adjacent to the pixel coordinate (x,y) are (x1,y1), (x1,y2), (x2,y1) and (x2,y2), and the values ​​corresponding to the pixel coordinates are f(Q 11 )、f(Q 12 )、f(Q 21 ) and f(Q 22 ).

[0098] Based on this, the above calculation is performed on the pixel values ​​corresponding to all pixel coordinates in the target size to obtain a downsampled aerial image.

[0099] Since there are many types of low-orbit satellites, such as Synthetic Aperture Rader (SAR) satellites, visible light satellites, and infrared imaging satellites, their resolutions are different due to factors such as the aspect ratio and orbit height of their imaging. Therefore, the resolution difference parameters corresponding to different satellite types can be selected according to the user's application requirements.

[0100] The color space conversion unit 2 is used to execute step S2: performing color conversion on the downsampled aerial image according to the color difference parameter to obtain an equivalent remote sensing image of the low-orbit satellite.

[0101] Specifically, the pixel values ​​of the downsampled aerial image are converted to obtain hue, saturation, and brightness, as follows:

[0102]

[0103]

[0104] V=MAX

[0105] Among them, H is the hue of the downsampled aerial image, S is the saturation of the downsampled aerial image, and V is the brightness of the downsampled aerial image; RGB is the pixel value of the three primary colors of the downsampled aerial image, representing the red value, green value, and blue value of the pixel value respectively; MAX is the maximum value of the three primary colors of RGB; MIN is the minimum value of the three primary colors of RGB.

[0106] The target hue parameter ΔH, the target saturation parameter ΔS, and the target lightness parameter ΔV in the hue, saturation, and lightness are added to the color difference parameters, and converted into the RGB color space to obtain the equivalent remote sensing image of the low-orbit satellite, as follows:

[0107] H′=H+ΔH

[0108] S′=S+ΔS

[0109] V′=V+ΔV

[0110] Among them, H′ is the hue of the equivalent remote sensing image of the low-orbit satellite, S′ is the saturation of the equivalent remote sensing image of the low-orbit satellite, and V′ is the brightness of the equivalent remote sensing image of the low-orbit satellite; H is the hue of the downsampled aerial image, S is the saturation of the downsampled aerial image, and V is the brightness of the downsampled aerial image.

[0111] Since different types of low-orbit satellites are affected by lighting, atmospheric environment and electromagnetic interference during imaging, the imaging results are different. Therefore, each low-orbit satellite has its own unique color difference parameters, and the same color difference parameters are not applicable to different types of satellites.

[0112] See also Figure 3 , Figure 3 Schematic diagram of the process for obtaining the resolution difference parameter.

[0113] In order to more accurately characterize the difference between the drone aerial image and the low-orbit satellite remote sensing image, and to more accurately convert the drone aerial image to the low-orbit satellite remote sensing image, the resolution difference parameter acquisition steps are as follows:

[0114] R1: Obtain the drone aerial image and its corresponding camera and other sensor parameters.

[0115] Specifically, the sensor parameters such as the shooting device are obtained through the flight log of the drone, and the sensor parameters such as the shooting device include internal reference information, external reference information and geographical location information at the imaging moment.

[0116] In order to convert remote sensing images of low-orbit satellites into visible light, infrared light and SAR types, the drone is provided with a shooting device and a sensor. The shooting device is used to obtain visible light aerial images of low-orbit satellites, and the sensor is used to obtain infrared light, SAR and other types of aerial images of low-orbit satellites.

[0117] See also Figure 4 , Figure 4 A simplified schematic diagram of the correction process for aerial images for registration.

[0118] R2: Based on the orbital parameters of the low-orbit satellite, the remote sensing images passing overhead, the sensor parameters such as the shooting device, the direction and scale of the UAV aerial images are aligned to obtain the aligned aerial images.

[0119] Specifically, the UAV flight trajectory is calculated using the data of low-orbit satellite orbital parameters, namely the two-line element set (TLE) and the TLE update time, so as to determine the remote sensing image of the low-orbit satellite passing overhead that is consistent with the UAV aerial image.

[0120] According to the remote sensing images of low-orbit satellites passing overhead, the external parameter information such as the shooting direction of the UAV aerial images is corrected to obtain the corrected UAV aerial images.

[0121] At the same time, the scale of the calibrated UAV aerial image is aligned with the satellite remote sensing image through the internal reference information of sensor parameters such as the shooting device to obtain the registered aerial image.

[0122] The scale alignment is to adjust the ground sample distance (GSD) of the drone aerial image to be consistent with the ground sample distance of the satellite remote sensing image.

[0123] The present invention uses the orbital parameters and TLE data of low-orbit satellites to accurately calculate the satellite's trajectory, thereby accurately finding the position that matches the drone aerial image, ensuring the directional consistency of the corrected drone aerial image and the low-orbit satellite remote sensing image.

[0124] R3: Roughly match the registered aerial image with the low-orbit satellite remote sensing image to obtain several low-orbit satellite remote sensing sub-images.

[0125] Specifically, according to the geographic location information of the registered aerial image and the latitude and longitude coordinates of the low-orbit satellite remote sensing image, the pixel coordinates of the registered aerial image in the low-orbit satellite remote sensing image are calculated, and the corresponding sub-image area is extracted to obtain several low-orbit satellite remote sensing sub-images.

[0126] The sub-image area is a low-orbit satellite remote sensing sub-image including the registered aerial image.

[0127] R4: Calculate the pixel density of the registered aerial image and the current low-orbit satellite remote sensing sub-image respectively.

[0128] Specifically, the pixel density of the registered aerial image is calculated as follows:

[0129]

[0130] Among them, Pixel U is the pixel density of the registered aerial image, U ALLis the total number of pixels of the registered aerial image, Area U is the image area of ​​the registered aerial image, which is calculated as follows:

[0131]

[0132] Among them, W U is the width of the registered aerial image, H U is the height of the registered aerial image, dpi U The number of pixels per inch of the aerial image to be registered is a known parameter. -

[0133] The pixel density of the registered aerial image is calculated as follows:

[0134]

[0135] Among them, Pixel S is the pixel density of the low-orbit satellite remote sensing sub-image, S ALL is the total number of pixels of the low-orbit satellite remote sensing sub-image, Area S is the image area of ​​the low-orbit satellite remote sensing sub-image, which is calculated as follows:

[0136]

[0137] Among them, W S is the width of the low-orbit satellite remote sensing sub-image, H S is the height of the low-orbit satellite remote sensing sub-image, and dpi is the number of pixels per inch of the low-orbit satellite remote sensing sub-image, which are known parameters.

[0138] R5: Calculate the difference between the pixel density of the aerial image and the pixel density of the current low-orbit satellite remote sensing sub-image to obtain the resolution difference parameter.

[0139] Specifically, its resolution difference parameter ΔR is calculated as follows:

[0140] ΔR=|Pixel S -Pixel U |

[0141] Among them, Pixel S is the pixel density of the low-orbit satellite remote sensing sub-image, Pixel U is the pixel density of the registered aerial image.

[0142] See also Figure 5 , Figure 5 A simple schematic diagram of the similarity calculation of image fingerprints.

[0143] R6: Downsample the UAV aerial image according to the resolution difference parameter to obtain a downsampled UAV aerial image.

[0144] Specifically, the scaling factor corresponding to the resolution difference parameter is obtained, and is calculated with the height and width of the current UAV aerial image to obtain the target size.

[0145] According to the UAV aerial image, bilinear interpolation is performed on the pixel values ​​corresponding to all pixel coordinates in the target size to obtain a downsampled aerial image.

[0146] The downsampling operation is the same as that of the aforementioned step S1, and the present invention will not be repeated here.

[0147] R7: Perform hash conversion and similarity calculation on the downsampled UAV aerial image and the low-orbit satellite remote sensing sub-image, and determine whether the similarity is less than a preset threshold: if yes, use the current resolution difference parameter as the standard resolution difference parameter; if no, use another low-orbit satellite remote sensing sub-image as input and execute step R4.

[0148] Specifically, the downsampled aerial image and the current low-orbit satellite remote sensing sub-image are gray-scaled, and the gray-scale image is hashed by a hash algorithm (such as image difference hashing, Difference Hash) to obtain image fingerprints of the downsampled aerial image and the current low-orbit satellite remote sensing sub-image respectively.

[0149] Calculate the Hamming distance between the downsampled aerial image and the image fingerprint of the current low-orbit satellite remote sensing sub-image, and determine whether the Hamming distance is less than a preset threshold: if yes, use the current resolution difference parameter as the standard resolution difference parameter; if not, use another low-orbit satellite remote sensing sub-image as the current low-orbit satellite remote sensing sub-image, and execute step R4.

[0150] The preset threshold is set according to the user's accuracy requirements. The smaller the preset threshold, the closer the resolution of the drone aerial image is to the low-orbit satellite remote sensing image after conversion. The Hamming distance is based on the 64-bit binary hash value converted into a 16-bit hexadecimal string for similarity calculation.

[0151] Furthermore, the Hamming distance performs similarity calculation based on a 64-bit binary hash value to increase the precision of the Hamming distance in representing the degree of similarity.

[0152] See also Figure 6 , Figure 6 Schematic diagram of the process of obtaining color difference parameters.

[0153] However, the difference between the drone aerial image and the low-orbit satellite remote sensing image is not limited to the resolution difference, but also includes the difference in color space. Therefore, in order to obtain a more accurate low-orbit satellite equivalent remote sensing image, the present invention further obtains the color difference parameter, and the acquisition steps are as follows:

[0154] C1: Perform color space conversion on the registered aerial images and low-orbit satellite remote sensing sub-images to obtain the corresponding hue, saturation and brightness.

[0155] The specific content is consistent with the content of color conversion of the pixel values ​​of the downsampled aerial image in the aforementioned step S2, and the present invention will not be described in detail here.

[0156] C2: Calculate the difference parameters of hue, saturation and brightness between the two to obtain color difference parameters.

[0157] Specifically, it is calculated as follows:

[0158] ΔH=H Target -H Original

[0159] ΔS=S Target -S Original

[0160] ΔV=V Target -V Original

[0161] Among them, ΔH is the hue parameter, ΔS is the saturation parameter, and ΔV is the brightness parameter; H Target , S Target and V Target are the hue, saturation and brightness of the low-orbit satellite remote sensing sub-image; H Original , S Original and V Original are the hue, saturation and brightness of the registered aerial images, respectively.

[0162] C3: Correct the registered aerial image through color difference parameters to obtain the equivalent remote sensing image of the low-orbit satellite to be confirmed.

[0163] Specifically, the hue, saturation and brightness of the registered aerial image are added to the hue parameter ΔH, saturation parameter ΔS and brightness parameter ΔV in the color difference parameter, and converted into RGB color space to obtain the equivalent remote sensing image of the low-orbit satellite to be confirmed.

[0164] The content of the correction calculation is consistent with the equivalent remote sensing image of the low-orbit satellite in the aforementioned step S2, and the present invention will not be described in detail here.

[0165] C4: Calculate the similarity score between the equivalent remote sensing image of the low-orbit satellite to be confirmed and the matching low-orbit satellite remote sensing sub-image, and determine whether it is less than a similarity threshold. If yes, use the current color difference parameter as the standard color difference parameter; if not, obtain another matching low-orbit satellite remote sensing sub-image and execute step C1.

[0166] Specifically, the histograms of hue and saturation corresponding to the equivalent remote sensing image of the low-orbit satellite to be confirmed and the matching low-orbit satellite remote sensing sub-image are calculated respectively.

[0167] The histograms are normalized and compared using the Bhattacharyya distance to obtain the similarity score, which is calculated as follows:

[0168]

[0169] Among them, Distance B (H1, H2) is the similarity score, the smaller the value, the higher the similarity; H1(i) is the value of the normalized hue-saturation histogram of the equivalent remote sensing image of the low-orbit satellite to be confirmed on the i-th BIN (array); H2(i) is the value of the normalized hue-saturation histogram of the matching low-orbit satellite remote sensing subimage on the i-th BIN (array).

[0170] In this embodiment, the similarity threshold is 0.2. However, due to different application scenarios and different requirements for conversion accuracy, users can set different similarity thresholds. Therefore, the present invention does not specifically limit the similarity threshold.

[0171] Compared with the prior art, the present invention accurately adjusts the resolution of the drone aerial image through the resolution difference parameter, so that it is consistent with the low-orbit satellite remote sensing image in spatial resolution. At the same time, the color difference parameter is used to adjust the hue, saturation and brightness of the drone aerial image to ensure that the color characteristics are consistent with the low-orbit satellite remote sensing image, thereby achieving a high degree of simulation of the image in visual effect. In addition, the present invention ensures that the drone is in a situation with large environmental differences, that is, different climatic conditions or different terrain features, through the resolution difference and color difference parameters, to achieve complete image conversion and obtain a conversion effect consistent with the low-orbit satellite remote sensing image.

[0172] Furthermore, in order to ensure the validity of the resolution difference parameters and the color difference parameters, the present invention analyzes the resolution difference parameters by using image fingerprint and Hamming distance, and analyzes the color difference parameters by using color histogram normalization and Bhattacharyya distance, thereby ensuring that the drone aerial images can be equivalently converted into satellite remote sensing images through the resolution difference parameters and the color difference parameters.

[0173] Based on this, the low-orbit satellite equivalent remote sensing images converted by the present invention can be widely used in a variety of business needs, such as path navigation, disaster detection and resource exploration in geographic information systems (GIS). It ensures that in various application scenarios, drone aerial images can be seamlessly integrated with low-orbit satellite remote sensing images, thereby achieving full coverage of low-orbit satellites, reducing blind spots, and providing high-quality data support.

[0174] In order to reflect the technical effect of the present invention, please also refer to Figure 7 , Figure 8 and Fig. 9 . Figure 7 is the drone aerial image to be converted, Figure 8 It is a low-orbit satellite remote sensing image of visible light type. Fig. 9 For equivalent conversion of drone aerial images.

[0175] Figure 7 The drone aerial images show clear ground details, while Figure 8 The low-orbit satellite remote sensing image blocked by clouds is shown. The image from the heterogeneous source has significant differences in resolution and color. However, the acquisition method of the equivalent low-orbit satellite remote sensing image of the present invention is Figure 7 Obtained by conversion Fig. 9 , which is closer to low-orbit satellite remote sensing images in terms of resolution and color ( Figure 8 ), the present invention enhances the consistency and availability of UAV aerial images and low-orbit satellite remote sensing images.

[0176] at the same time, Fig.10 The location where the drone aerial image was taken is shown. If you need to apply equivalent transformed drone aerial images, you can directly embed them into Fig.10 The image is placed in the box in the figure, thereby reducing the blind area caused by cloud cover in low-orbit satellite remote sensing images.

[0177] Also, see Fig.11 , which shows the differences in hue, saturation and brightness between the drone aerial image to be converted, the equivalently converted drone aerial image and the visible light low-orbit satellite remote sensing image. The y-axis is the number of pixels, and the x-axis is the corresponding hue, saturation and brightness values. As can be seen from the figure, the drone aerial image after equivalent conversion is very close to the visible light low-orbit satellite remote sensing image in hue, which further proves the effectiveness of the present invention in color consistency.

[0178] Through comparative analysis of these images and data, the present invention demonstrates the technical effect of converting drone aerial images into low-orbit satellite remote sensing images, and not only accurately adjusts the resolution and color space of drone aerial images, thereby ensuring that the converted images have high consistency and usability with low-orbit satellite images. This technical effect makes the present invention of great value in practical applications, especially in scenarios where multiple remote sensing data need to be combined for analysis.

[0179] Based on the same inventive concept, the present application also provides an electronic device, which may be a terminal device such as a server, a desktop computing device or a mobile computing device (e.g., a laptop computing device, a handheld computing device, a tablet computer, a netbook, etc.). The device includes one or more processors and a memory, wherein the processor is used to execute a program to implement the method for acquiring an equivalent low-orbit satellite remote sensing image according to an embodiment of the present invention; and the memory is used to store a computer program executable by the processor.

[0180] Based on the same inventive concept, the present application also provides a computer-readable storage medium, corresponding to an embodiment of the aforementioned method for acquiring an equivalent low-orbit satellite remote sensing image, wherein the computer-readable storage medium stores a computer program thereon, and when the program is executed by a processor, the steps of the method for acquiring an equivalent low-orbit satellite remote sensing image recorded in any of the aforementioned embodiments are implemented.

[0181] The present application may take the form of a computer program product implemented on one or more storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing program code. Computer-usable storage media include permanent and non-permanent, removable and non-removable media, and information storage can be achieved by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include but are not limited to: phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, tape disk storage or other magnetic storage devices or any other non-transmission medium that can be used to store information that can be accessed by a computing device.

[0182] The above-mentioned embodiments only express several implementation methods of the present invention, and the description is relatively specific and detailed, but it cannot be understood as limiting the scope of the invention patent. It should be pointed out that for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present invention, and the present invention is also intended to include these modifications and modifications.

Claims

1. A method for acquiring an equivalent low-orbit satellite remote sensing image, characterized in that: The following steps are involved: S1: downsampling the aerial image according to a standard resolution difference parameter to obtain a downsampled aerial image; The resolution difference parameter is the difference in pixel density. The steps for obtaining the standard resolution difference parameter are as follows: R1: Obtain aerial images and their corresponding sensor parameters; The sensor parameters include internal reference information, external reference information and geographic location information at the time of imaging by the camera; R2: According to the orbital parameters of the low-orbit satellite, the remote sensing image passing overhead and the sensor parameters, the aerial image is oriented and scaled to obtain a registered aerial image; R3: Roughly match the registered aerial image with the low-orbit satellite remote sensing image to obtain several low-orbit satellite remote sensing sub-images; R4: Calculate the pixel density of the registered aerial image and the current low-orbit satellite remote sensing sub-image respectively; R5: Calculate the difference between the pixel density of the aerial image and the pixel density of the current low-orbit satellite remote sensing sub-image to obtain the resolution difference parameter; The resolution difference parameter ΔR is calculated as follows: ΔR=|Pixel S -Pixel U | Among them, Pixel S is the pixel density of the low-orbit satellite remote sensing sub-image, Pixel U is the pixel density of the registered aerial image; R6: downsample the aerial image according to the resolution difference parameter to obtain a downsampled aerial image; R7: Perform hash conversion and similarity calculation on the downsampled aerial image and the low-orbit satellite remote sensing sub-image, and determine whether the similarity is less than a preset threshold: if yes, use the current resolution difference parameter as the standard resolution difference parameter; if no, use another low-orbit satellite remote sensing sub-image as input and execute step R4; S2: Perform color conversion on the downsampled aerial image according to the color difference parameter to obtain an equivalent remote sensing image of the low-orbit satellite; The color difference parameters include a hue parameter ΔH, a saturation parameter ΔS, and a brightness parameter ΔV; the steps for obtaining the color difference parameters are as follows: C1: Perform color space conversion on the registered aerial images and low-orbit satellite remote sensing sub-images to obtain the corresponding hue, saturation and brightness; C2: Calculate the difference parameters of hue, saturation and brightness of the registered aerial image and the low-orbit satellite remote sensing sub-image to obtain the color difference parameters; C3: Correct the registered aerial image by using the color difference parameter to obtain the equivalent remote sensing image of the low-orbit satellite to be confirmed; C4: Calculate the similarity score between the equivalent remote sensing image of the low-orbit satellite to be confirmed and the matching low-orbit satellite remote sensing sub-image, and determine whether it is less than a similarity threshold. If yes, use the current color difference parameter as the standard color difference parameter; if not, obtain another matching low-orbit satellite remote sensing sub-image and execute step C1.

2. The method for acquiring an equivalent low-orbit satellite remote sensing image according to claim 1, characterized in that: The step R7 specifically includes the following sub-steps: By graying the downsampled aerial image and the current low-orbit satellite remote sensing sub-image, and performing hash calculation on the gray-scale image through a hash algorithm, the image fingerprints of the downsampled aerial image and the current low-orbit satellite remote sensing sub-image are obtained respectively; Next, the Hamming distance between the downsampled aerial image and the image fingerprint of the current low-orbit satellite remote sensing sub-image is calculated, and it is determined whether the Hamming distance is less than a preset threshold: if yes, the current resolution difference parameter is used as the standard resolution difference parameter; if not, another low-orbit satellite remote sensing sub-image is used as the current low-orbit satellite remote sensing sub-image, and step R4 is executed; The Hamming distance is calculated based on the similarity of a 16-bit hexadecimal string converted from a 64-bit binary hash value.

3. The method for acquiring an equivalent low-orbit satellite remote sensing image according to claim 2, characterized in that: The similarity score is calculated as follows: Calculate the histograms of hue and saturation corresponding to the equivalent remote sensing image of the low-orbit satellite to be confirmed and the remote sensing sub-image of the matching low-orbit satellite respectively; The histograms are normalized and compared using the Bhattacharyya distance to obtain the similarity score, which is calculated as follows: Among them, Distance B (H1, H2) is the similarity score, the smaller the value, the higher the similarity; H1(i) is the value of the normalized hue-saturation histogram of the equivalent remote sensing image of the low-orbit satellite to be confirmed in the i-th array, that is, BIN; H2(i) is the value of the normalized hue-saturation histogram of the matching low-orbit satellite remote sensing sub-image in the i-th BIN.

4. A device for acquiring equivalent low-orbit satellite remote sensing images, characterized in that: including a resolution conversion unit and a color space conversion unit; The resolution conversion unit is used to downsample the aerial image according to a standard resolution difference parameter to obtain a downsampled aerial image; wherein the resolution difference parameter is a difference in pixel density; the steps for obtaining the standard resolution difference parameter are as follows: R1: Obtain aerial images and their corresponding sensor parameters; The sensor parameters include internal reference information, external reference information and geographic location information at the time of imaging by the camera; R2: According to the orbital parameters of the low-orbit satellite, the remote sensing image passing overhead and the sensor parameters, the aerial image is oriented and scaled to obtain a registered aerial image; R3: Roughly match the registered aerial image with the low-orbit satellite remote sensing image to obtain several low-orbit satellite remote sensing sub-images; R4: Calculate the pixel density of the registered aerial image and the current low-orbit satellite remote sensing sub-image respectively; R5: Calculate the difference between the pixel density of the aerial image and the pixel density of the current low-orbit satellite remote sensing sub-image to obtain the resolution difference parameter; The resolution difference parameter ΔR is calculated as follows: ΔR=|Pixel S -Pixel U | Among them, Pixel S is the pixel density of the low-orbit satellite remote sensing sub-image, Pixel U is the pixel density of the registered aerial image; R6: Downsample the aerial image according to the resolution difference parameter to obtain a downsampled drone aerial image; R7: Perform hash conversion and similarity calculation on the downsampled aerial image and the low-orbit satellite remote sensing sub-image, and determine whether the similarity is less than a preset threshold: if yes, use the current resolution difference parameter as the standard resolution difference parameter; if no, use another low-orbit satellite remote sensing sub-image as input and execute step R4; The color space conversion unit is used to perform color conversion on the downsampled aerial image according to the color difference parameter to obtain an equivalent remote sensing image of the low-orbit satellite; wherein the color difference parameter includes a hue parameter ΔH, a saturation parameter ΔS, and a brightness parameter ΔV; the steps for obtaining the color difference parameter are as follows: C1: Perform color space conversion on the registered aerial images and low-orbit satellite remote sensing sub-images to obtain the corresponding hue, saturation and brightness; C2: Calculate the difference parameters of hue, saturation and brightness of the registered aerial image and the low-orbit satellite remote sensing sub-image to obtain the color difference parameters; C3: Correct the registered aerial image by using the color difference parameter to obtain the equivalent remote sensing image of the low-orbit satellite to be confirmed; C4: Calculate the similarity score between the equivalent remote sensing image of the low-orbit satellite to be confirmed and the matching low-orbit satellite remote sensing sub-image, and determine whether it is less than a similarity threshold. If yes, use the current color difference parameter as the standard color difference parameter; if not, obtain another matching low-orbit satellite remote sensing sub-image and execute step C1.

5. The device for acquiring equivalent low-orbit satellite remote sensing images according to claim 4, characterized in that: in, The step R7 specifically includes the following sub-steps: By graying the downsampled aerial image and the current low-orbit satellite remote sensing sub-image, and performing hash calculation on the gray-scale image through a hash algorithm, the image fingerprints of the downsampled aerial image and the current low-orbit satellite remote sensing sub-image are obtained respectively; Next, the Hamming distance between the downsampled aerial image and the image fingerprint of the current low-orbit satellite remote sensing sub-image is calculated, and it is determined whether the Hamming distance is less than a preset threshold: if yes, the current resolution difference parameter is used as the standard resolution difference parameter; if not, another low-orbit satellite remote sensing sub-image is used as the current low-orbit satellite remote sensing sub-image, and step R4 is executed; The Hamming distance is calculated based on the 16-bit hexadecimal string converted from the 64-bit binary hash value. The similarity score is calculated as follows: Calculate the histograms of hue and saturation corresponding to the equivalent remote sensing image of the low-orbit satellite to be confirmed and the remote sensing sub-image of the matching low-orbit satellite respectively; The histograms are normalized and compared using the Bhattacharyya distance to obtain the similarity score, which is calculated as follows: Among them, Distance B (H1, H2) is the similarity score, the smaller the value, the higher the similarity; H1(i) is the value of the normalized hue-saturation histogram of the equivalent remote sensing image of the low-orbit satellite to be confirmed in the i-th array, that is, BIN; H2(i) is the value of the normalized hue-saturation histogram of the matching low-orbit satellite remote sensing sub-image in the i-th BIN.

6. An electronic device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method for acquiring an equivalent low-orbit satellite remote sensing image as described in any one of claims 1 to 3 when executing the computer program.

7. A computer-readable storage medium storing computer-executable instructions, characterized in that: The computer executable instructions are used in the method for acquiring equivalent low-orbit satellite remote sensing images as described in any one of claims 1 to 3.

Images