A no-reference remote sensing image quality evaluation method

Abstract

The application is a no-reference remote sensing image quality evaluation method. The application relates to the technical field of satellite remote sensing image data quality evaluation, extracts image data to be evaluated; carries out block multi-thread processing on the image data, sets a blurring threshold, determines whether there is a blurring problem; detects whether there is a string problem in the image data, calculates a string evaluation index edge energy; sets a string problem index threshold, judges whether the index value calculated by each block exceeds the threshold, determines whether there is a string problem, and completes the inspection and evaluation. The application can automatically, accurately and objectively evaluate the quality of remote sensing image data.

Description

Technical Field

[0001] This invention relates to the field of satellite remote sensing image data quality assessment technology, and is a reference-free remote sensing image quality assessment method. Background Technology

[0002] Spaceborne remote sensing technology is characterized by its rich information content and large coverage area, and its applications are becoming increasingly widespread. It plays a crucial role in agriculture, forestry, military reconnaissance, geodesy, and geographic information acquisition. The quality of remote sensing imagery reflects the performance of spacecraft and the capabilities of image processing algorithms. With the rapid increase in remote sensing image data, the importance of remote sensing image quality evaluation is becoming increasingly prominent. Relying solely on subjective human evaluation can no longer meet the demands of efficient and rapid remote sensing applications. Therefore, a reference-free remote sensing image quality evaluation method is proposed.

[0003] The quality assessment of remote sensing imagery is mainly divided into two types: subjective assessment methods and objective assessment methods. Subjective assessment methods involve evaluation by personnel with professional knowledge. Subjective assessment is influenced by the evaluator's professional background and psychological state, thus possessing a certain degree of subjectivity. Objective assessment methods extract texture and radiometric feature parameters from the image and evaluate it by calculating objective indicators. Commonly used image evaluation indicators include: grayscale statistics, signal-to-noise ratio, information entropy, and second moment of angle. The calculation results of objective evaluation indicators are significantly affected by ground cover type. Existing remote sensing image quality assessment methods primarily involve independently calculating objective evaluation indicators or establishing simple comprehensive evaluation models, without addressing specific image quality issues, and therefore lack practical application.

[0004] This method analyzes remote sensing links, detecting image blurring caused by atmospheric scattering, optical systems, and satellite vibrations, as well as image crosstalk issues caused by imaging electronics. Based on the detection results, image quality is evaluated, which has practical significance for the use of remote sensing image data.

[0005] The key feature of this method is that it can automatically, accurately, and objectively evaluate the quality of remote sensing image data. Summary of the Invention

[0006] To overcome the shortcomings of existing technologies, this invention proposes a reference-free remote sensing image quality assessment method, based on current remote sensing image quality assessment techniques, to address the inability of existing quality assessment techniques to detect the actual existence of problems in image data during satellite imaging.

[0007] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0008] This invention provides a reference-free method for assessing the quality of remote sensing images. The invention offers the following technical solutions:

[0009] A reference-free remote sensing image quality assessment method, the method comprising the following steps:

[0010] Step 1: Extract the image data to be evaluated;

[0011] Step 2: Perform block-based multi-threaded processing on the image data. First, calculate the blur of the extracted blocks. The blur is used to characterize the degree of blur of the image features.

[0012] Step 3: Set a fuzziness threshold to determine if fuzziness issues exist;

[0013] Step 4: Detect whether there is a crosstalk problem in the image data, calculate the crosstalk evaluation index. The crosstalk problem detection is evaluated by two indices together to improve the reliability of the detection.

[0014] Step 5: Calculate the edge energy of the series evaluation index;

[0015] Step 6: Calculate the contrast ratio of the series evaluation index;

[0016] Step 7: Set the threshold for the serialization problem indicator, determine whether the calculated indicator value of each block exceeds the threshold, determine whether there is a serialization problem, and complete the inspection and evaluation.

[0017] Preferably, step 2 is divided into blocks as follows:

[0018] To detect image blurring, after comparing detection results with different block sizes, the block size was fixed at 200×200 pixels. For detecting image jamming caused by imaging electronics, the block size was set based on the jamming width. `width` refers to the width of the string.

[0019] Preferably, step 3 specifically includes the following steps:

[0020] Step 3.1: For the image block F to be evaluated, with a size of m×n pixels, perform vertical and horizontal filtering to obtain the corresponding blurred image B:

[0021] B Ver =h v *F

[0022] B Hor =h h *F

[0023] Among them, B Ver B Hor The blurred images obtained after vertical and horizontal filtering of image block F are shown below, h. v ,h h The vertical and horizontal models of the filter are respectively expressed as follows:

[0024]

[0025] Step 3.2: Calculate the absolute error D_F of the image block F before filtering. Ver ,D_F Hor The absolute error D_B of the filtered image patch Ver ,D_B Hor This allows us to obtain the grayscale value changes of adjacent pixels.

[0026] Subtract the absolute errors, i.e., D_F Ver With D_B Ver Difference, D_F Hor With D_B Hor By subtracting the values, we can obtain the difference between the image patches before and after filtering.

[0027] V Ver =max(0,D_F Ver (i,j)-D_B Ver (i,j))

[0028] V Hor =max(0,D_F Hor (i,j)-D_B Hor (i,j))

[0029] Where i takes values ​​in the range [1, m-1] and j takes values ​​in the range [1, n-1].

[0030] Step 3.3: To compare the overall differences between image patches, the absolute D_F values ​​of the image patches before and after filtering are calculated. Ver ,D_F Hor Error D_B Ver D_B Hor Summing yields the differences across the entire image patch:

[0031]

[0032]

[0033]

[0034]

[0035] Step 3.4: Normalize the results to [0, 1]:

[0036]

[0037]

[0038] Step 3.5: Select the calculated b_F Ver and b_F Hor The maximum value among the values ​​is used as the final evaluation value:

[0039] blur F =max(b_F Ver ,b_F Hor )

[0040] Among them, blur F ∈[0,1], when blur F A value of 0 indicates the lowest image blur and the best image sharpness; when blur is 0, the image is clearest. F A value of 1 indicates the highest image blur and the worst image clarity.

[0041] A threshold is set to determine whether the calculated blur value of each block exceeds the threshold; the threshold is obtained statistically based on a large amount of experimental image data.

[0042] Step 3.6: Record the total number of blocks m that exceeds the threshold. When the total number of blocks m > Block, it is considered that there is a blur problem in the scene, where Block is the set threshold for the total number of blocks.

[0043] Preferably, step 5 specifically comprises:

[0044] When there is a problem with image data being strung together, the index value will show a peak at the position of the strung pixels, resulting in obvious changes;

[0045] The contrast ratio, a key performance indicator for series evaluation, is calculated using the following method:

[0046]

[0047] Where |ij|=n, These are the elements of the normalized gray-level co-occurrence matrix; The gray-level co-occurrence matrix (GLCM) is the total number of times the gray value of a point and its neighboring points appear in a point pair on an image. Normalization is then performed to obtain the probability of occurrence. δ represents the distance between two points, where δ = 1; θ represents the angle between the distance and the coordinate axis. Since the stringing problem occurs in the column direction, θ = 0 here, meaning the pixel pair is in the horizontal direction.

[0048]

[0049] Preferably, step 6 specifically comprises:

[0050] The edge energy calculation method is as follows:

[0051]

[0052] e(x,y)=E1(f(x,y))+E2(f(x,y))

[0053] in:

[0054]

[0055] Preferably, step 7 specifically comprises:

[0056] Because of overly bright ground features such as clouds and snow, the evaluation index values ​​for these features may exceed the set threshold. Therefore, a strategy of continuous statistical analysis of abnormal blocks is adopted to avoid misjudgment of these features; block size... `width` is the width of the sequence. According to this block division rule, the span of abnormal pixel areas is a maximum of 3 blocks. This rule is used to avoid the problem of normal images being misjudged.

[0057] Preferably, the threshold is set separately according to the land cover type with high, medium and low brightness values; when the set threshold is exceeded, the pixel position and the number of abnormal blocks are recorded.

[0058] A referenceless remote sensing image quality assessment system, the system comprising:

[0059] A data extraction module, which extracts the image data to be evaluated;

[0060] The data processing module performs block-based multi-threaded processing on the image data. The extracted blocks are first subjected to blur calculation, and the blur is used to characterize the degree of blur of the image features.

[0061] A fuzziness threshold setting module, wherein the fuzziness threshold setting module sets a fuzziness threshold to determine whether a fuzziness problem exists;

[0062] The detection module detects whether there is a crosstalk problem in the image data and calculates a crosstalk evaluation index. The crosstalk problem detection is evaluated by two indicators together to improve the reliability of the detection.

[0063] A contrast calculation module, which calculates the contrast ratio of the series evaluation index;

[0064] An edge energy calculation module, which calculates the edge energy, a serial evaluation index;

[0065] The evaluation module sets a threshold for the serialization problem index, determines whether the calculated index value of each block exceeds the threshold, identifies whether a serialization problem exists, and completes the inspection and evaluation.

[0066] A computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement a referenceless remote sensing image quality assessment method.

[0067] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement a referenceless remote sensing image quality assessment method.

[0068] The present invention has the following beneficial effects:

[0069] This invention analyzes remote sensing links, detecting image blurring caused by atmospheric scattering, optical systems, and satellite vibrations, as well as image jamming issues caused by imaging electronics. Based on the detection results, image quality is evaluated, which has practical significance for the use of remote sensing image data.

[0070] This invention can automatically, accurately, and objectively evaluate the quality of remote sensing image data. Attached Figure Description

[0071] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0072] Figure 1 Flowchart of a method for assessing the quality of remote sensing images without reference;

[0073] Figure 2 Global graph for test cases;

[0074] Figure 3 This is a magnified illustration of the test sample.

[0075] Figure 4 Global graph for test cases;

[0076] Figure 5 This is a magnified illustration of the test sample. Detailed Implementation

[0077] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. 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.

[0078] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0079] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances. Specific Implementation Example 1:

[0081] according to Figures 1 to 5 As shown, the specific optimized technical solution adopted by the present invention to solve the above-mentioned technical problems is: The present invention relates to a method for evaluating the quality of remote sensing images without reference.

[0082] A reference-free remote sensing image quality assessment method, the method comprising the following steps:

[0083] A referenceless remote sensing image quality assessment method, characterized by the following steps:

[0084] Step 1: Extract the image data to be evaluated;

[0085] Step 2: Perform block-based multi-threaded processing on the image data. First, calculate the blur of the extracted blocks. The blur is used to characterize the degree of blur of the image features.

[0086] Step 3: Set a fuzziness threshold to determine if fuzziness issues exist;

[0087] Step 4: Detect whether there is a crosstalk problem in the image data, calculate the crosstalk evaluation index. The crosstalk problem detection is evaluated by two indices together to improve the reliability of the detection.

[0088] Step 5: Calculate the contrast ratio of the series evaluation indicators;

[0089] Step 6: Calculate the edge energy of the series evaluation index;

[0090] Step 7: Set the threshold for the serialization problem indicator, determine whether the calculated indicator value of each block exceeds the threshold, determine whether there is a serialization problem, and complete the inspection and evaluation.

[0091] The specific steps of step 2, segmentation, are as follows:

[0092] To detect image blurring, after comparing detection results with different block sizes, the block size was fixed at 200×200 pixels. For detecting image jamming caused by imaging electronics, the block size was set based on the jamming width. `width` refers to the width of the string.

[0093] Step 3 specifically includes the following steps:

[0094] Step 3.1: For the image block F to be evaluated, with a size of m×n pixels, perform vertical and horizontal filtering to obtain the corresponding blurred image B:

[0095] B Ver =h v *F

[0096] B Hor =h h *F

[0097] Among them, B Ver B Hor The blurred images obtained after vertical and horizontal filtering of image block F are shown below, h. v ,h h The vertical and horizontal models of the filter are respectively expressed as follows:

[0098]

[0099] Step 3.2: Calculate the absolute error D_F of the image block F before filtering. Ver ,D_F Hor The absolute error D_B of the filtered image patch Ver ,D_B Hor This allows us to obtain the grayscale value changes of adjacent pixels.

[0100] Subtract the absolute errors, i.e., D_F Ver With D_B Ver Difference, D_F Hor With D_B Hor By subtracting the values, we can obtain the difference between the image patches before and after filtering.

[0101] V Ver =max(0,D_F Ver (i,j)-D_B Ver (i,j))

[0102] V Hor =max(0,D_F Hor (i,j)-D_B Hor (i,j))

[0103] Where i takes values ​​in the range [1, m-1] and j takes values ​​in the range [1, n-1].

[0104] Step 3.3: To compare the overall differences between image patches, the absolute D_F values ​​of the image patches before and after filtering are calculated. Ver ,D_F Hor Error D_B Ver D_B Hor Summing yields the differences across the entire image patch:

[0105]

[0106]

[0107]

[0108]

[0109] Step 3.4: Normalize the results to [0, 1]:

[0110]

[0111]

[0112] Step 3.5: Select the calculated b_F Ver and b_F Hor The maximum value among the values ​​is used as the final evaluation value:

[0113] blur F =max(b_F Ver ,b_F Hor )

[0114] Among them, blur F∈[0,1], when blur F A value of 0 indicates the lowest image blur and the best image sharpness; when blur is 0, the image is clearest. F A value of 1 indicates the highest image blur and the worst image clarity.

[0115] A threshold is set to determine whether the calculated blur value of each block exceeds the threshold; the threshold is obtained statistically based on a large amount of experimental image data.

[0116] Step 3.6: Record the total number of blocks m that exceeds the threshold. When the total number of blocks m > Block, it is considered that there is a blur problem in the scene, where Block is the set threshold for the total number of blocks.

[0117] Step 5 specifically involves:

[0118] When there is a problem with image data being strung together, the index value will show a peak at the position of the strung pixels, resulting in obvious changes;

[0119] The contrast ratio, a key performance indicator for series evaluation, is calculated using the following method:

[0120]

[0121] Where |ij|=n, These are the elements of the normalized gray-level co-occurrence matrix; The gray-level co-occurrence matrix (GLCM) is the total number of times the gray value of a point and its neighboring points appear in a point pair on an image. Normalization is then performed to obtain the probability of occurrence. δ represents the distance between two points, where δ = 1; θ represents the angle between the distance and the coordinate axis. Since the stringing problem occurs in the column direction, θ = 0 here, meaning the pixel pair is in the horizontal direction.

[0122]

[0123] Step 6 specifically involves:

[0124] The edge energy calculation method is as follows:

[0125]

[0126] e(x,y)=E1(f(x,y))+E2(f(x,y))

[0127] in:

[0128]

[0129] Step 7 specifically involves:

[0130] Because of overly bright ground features such as clouds and snow, the evaluation index values ​​for these features may exceed the set threshold. Therefore, a strategy of continuous statistical analysis of abnormal blocks is adopted to avoid misjudgment of these features; block size... `width` is the width of the sequence. According to this block division rule, the span of abnormal pixel areas is a maximum of 3 blocks. This rule is used to avoid the problem of normal images being misjudged.

[0131] Thresholds are set separately for different land cover types with high, medium, and low brightness values; when the set threshold is exceeded, the pixel location and the number of abnormal blocks are recorded. Specific Implementation Example 2:

[0133] The only difference between Embodiment 2 and Embodiment 1 of this application is that:

[0134] A referenceless remote sensing image quality assessment method, the method comprising:

[0135] Image blur detection caused by atmospheric scattering, optical systems, and satellite vibration;

[0136] Detection of image jamming issues caused by imaging electronics;

[0137] The implementation process of this invention includes the following steps:

[0138] Step 1: Extract the image data to be evaluated and perform block-based multi-threaded processing on the image data. Since blurring and crosstalk issues do not exist in the entire scene, a block-based approach is used for local detection to improve the accuracy and efficiency of detection.

[0139] Furthermore, the rules for the block size are as follows:

[0140] For image blur detection, after comparing detection results with different block sizes, the block size was fixed at 200×200 pixels. For image jamming caused by imaging electronics, the block size was set according to the jamming width. Block size `width` refers to the width of the string.

[0141] Step 2: First, perform blur calculation on the extracted small patches. Blur is used to characterize the degree of blur in image features. MTF is a relatively mature method for evaluating sharpness, but its measurement is complex. To simplify and quickly reflect the sharpness changes of the entire image, a blur index is used to evaluate image sharpness. The idea behind this invention is: for a sharp image, especially in texture-rich areas, the difference between adjacent pixels is large. After low-pass filtering, adjacent pixels will approach the same gray level, significantly reducing the difference. Conversely, for a blurry image, adjacent pixels are already close to the same gray level, and the difference change after low-pass filtering is not as significant. For a blurred image already affected by atmospheric scattering, optical systems, and satellite vibrations, edge enhancement is ineffective for severely blurred images, resulting in minimal enhancement. For slightly blurred or sharp images, the enhancement effect is significant. Therefore, edge enhancement is first performed on the image data to be evaluated. The purpose of this step is to make sharp images even sharper, widening the gap between sharp and blurred images in terms of blur index.

[0142] Step 3: Perform vertical and horizontal filtering on the image block F to be evaluated (with a size of m×n pixels) to obtain the corresponding blurred image B:

[0143]

[0144] Among them, B Ver B Hor The blurred images obtained after vertical and horizontal filtering of image block F are shown below, h. v ,h h The vertical and horizontal models of the filter are respectively expressed as follows:

[0145]

[0146] Step 4: Calculate the absolute error D_F of the image block F before filtering. Ver ,D_F Hor The absolute error D_B of the filtered image patch Ver ,D_B Hor To obtain the grayscale value changes of adjacent pixels:

[0147]

[0148] Step 5: Subtract the absolute error from Step 4, i.e., D_F Ver With D_B Ver Difference, D_F Hor With D_B Hor By subtracting the values, we can obtain the difference between the image patches before and after filtering.

[0149] V Ver =max(0,D_FVer (i,j)-D_B Ver (i,j))

[0150] V Hor =max(0,D_F Hor (i,j)-D_B Hor (i,j))

[0151] Formula 3

[0152] Where i takes values ​​in the range [1, m-1] and j takes values ​​in the range [1, n-1].

[0153] Step 6: To compare the overall differences between image patches, calculate the absolute error D_F of the image patches before and after filtering. Ver ,D_F Hor ,difference D_B Ver D_B Hor Summing yields the differences across the entire image patch:

[0154]

[0155] Step 7: Normalize the result to [0, 1]:

[0156]

[0157] Step 8: Select b_F obtained in Step 7 Ver and b_F Hor The maximum value among the values ​​is used as the final evaluation value:

[0158] blur F =max(b_F Ver ,b_F Hor ) Formula 6

[0159] Among them, blur F ∈[0,1]. When blur F A value of 0 indicates the lowest image blur and the best image sharpness; when blur is 0, the image is clearest. F A value of 1 indicates the highest image blur and the worst image clarity.

[0160] Step 9: Set a threshold to determine whether the calculated blur value for each block exceeds the threshold. The threshold is obtained statistically based on a large amount of experimental image data.

[0161] Step 10: Record the total number of blocks m that exceeds the threshold. When the total number of blocks m > Block, the scene is considered to have a blur problem. Here, Block is the set threshold for the total number of blocks.

[0162] Step 11: Detect whether there is a crosstalk problem in the image data, calculate the crosstalk evaluation index. The crosstalk problem detection is evaluated by two indices together to improve the reliability of the detection.

[0163] If there is a crosstalk problem in the image data, the index value will show a peak at the crosstalk pixel position, resulting in obvious changes.

[0164] Step 12: Calculate the contrast ratio, an evaluation metric for the series. The method for calculating the contrast ratio is as follows:

[0165]

[0166] In the formula, |ij|=n, These are the elements of the normalized gray-level co-occurrence matrix. The gray-level co-occurrence matrix (GLCM) is the total number of times the gray value of a point and its neighboring points appear in a point pair on an image. Normalization is then performed to obtain the probability of occurrence. δ represents the distance between two points, where δ = 1; θ represents the angle between the distance and the coordinate axis. Since the string problem occurs in the column direction, θ = 0 here, meaning the pixel pair is in the horizontal direction.

[0167]

[0168] Step 13: Calculate the edge energy of the series evaluation index. The edge energy calculation method is as follows:

[0169]

[0170] e(x,y)=E1(f(x,y))+E2(f(x,y)) Formula 10

[0171] in:

[0172]

[0173] Step 14: Set indicator thresholds and determine whether the calculated indicator value for each block exceeds the threshold. Thresholds are set separately for different land cover types (high, medium, and low brightness values). If the threshold is exceeded, record the pixel location and the number of abnormal blocks.

[0174] Step 15: Due to overly bright ground features such as clouds and snow, the evaluation index values ​​of these features may exceed the set threshold. This invention employs a strategy of continuous statistical analysis of abnormal blocks to avoid misjudgment of such features. Block size. `width` refers to the width of the sequence. According to this block division rule, the span of abnormal pixel regions is at most 3 blocks. This rule can effectively avoid the problem of normal images being misjudged. Specific Implementation Example 3:

[0176] The only difference between Embodiment 3 and Embodiment 2 of this application is that:

[0177] This invention provides a referenceless remote sensing image quality assessment system, the system comprising:

[0178] A data extraction module, which extracts the image data to be evaluated;

[0179] The data processing module performs block-based multi-threaded processing on the image data. The extracted blocks are first subjected to blur calculation, and the blur is used to characterize the degree of blur of the image features.

[0180] A fuzziness threshold setting module, wherein the fuzziness threshold setting module sets a fuzziness threshold to determine whether a fuzziness problem exists;

[0181] The detection module detects whether there is a crosstalk problem in the image data and calculates a crosstalk evaluation index. The crosstalk problem detection is evaluated by two indicators together to improve the reliability of the detection.

[0182] A contrast calculation module, which calculates the contrast ratio of the series evaluation index;

[0183] An edge energy calculation module, which calculates the edge energy, a serial evaluation index;

[0184] The evaluation module sets a threshold for the serialization problem index, determines whether the calculated index value of each block exceeds the threshold, identifies whether a serialization problem exists, and completes the inspection and evaluation. Specific Implementation Example 4:

[0186] The present invention provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement a referenceless remote sensing image quality assessment method. Specific Implementation Example 5:

[0188] The present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement a referenceless remote sensing image quality assessment method. Specific Implementation Example Six:

[0190] Step 1: Extract the image data to be evaluated. The image data has a resolution of 0.75 meters and is panchromatic. The image data has already undergone edge enhancement processing. For blur detection, the image block size is 200×200 pixels. For sequence detection, the image block size is 400×34575 pixels. Test sample information is shown in Table 1.

[0191] Table 1. Detailed information on test samples

[0192] Shooting time 2023-05-17 15:27:34 Side swing angle 0.8° Cloud cover 0 Place Africa

[0193] Calculate the ambiguity index for each block according to steps 2 through 8. A total of 29,240 blocks were calculated.

[0194] Step 9: Set a threshold to determine whether the calculated fuzzy value for each block exceeds the threshold. Set the threshold to 0.40.

[0195] Step 10: Record the total number of blocks m exceeding the threshold. When the total number of blocks m > Block, the scene is considered to have a blur problem. Here, Block is the set threshold for the total number of blocks. Block is set to 5000, and the calculation shows that m is 15032, therefore the scene is considered to have a blur problem. Visual inspection confirms that the scene image does indeed have a blur problem.

[0196] Step 11: Check if there is a problem with image data being strung together.

[0197] Steps 12-13: Calculate the evaluation index for each image data block.

[0198] Step 14: Set indicator thresholds and determine whether the calculated indicator value for each block exceeds the threshold. Thresholds are divided according to land cover type. Thresholds are set separately for high, medium, and low brightness land cover types. If the set threshold is exceeded, record the pixel location and the number of blocks. The contrast threshold for low brightness land cover is set to 1800, and the edge energy threshold is set to 650. The contrast threshold for medium brightness land cover is set to 3500, and the edge energy threshold is set to 2000. The contrast threshold for high brightness land cover is set to 12000, and the edge energy threshold is set to 3500.

[0199] Step 15: The panoramic image is divided into 85 blocks. Calculations show that the maximum contrast ratio is 189.962 and the maximum edge energy is 118.672. Both the maximum contrast ratio and edge energy are less than the set thresholds, therefore, the image is considered to have no crosstalk issue. Visual inspection confirms that the image indeed does not have a crosstalk problem. Specific Implementation Example 7:

[0201] Step 1: Extract the image data to be evaluated. The image data has a resolution of 0.75 meters and is panchromatic. The image data has already undergone edge enhancement processing. For blur detection, the image block size is 200×200 pixels. For sequence detection, the image block size is 400×28669 pixels. Test sample information is shown in Table 2.

[0202] Table 2. Detailed information on test samples

[0203] Shooting time 2023-05-2510:56:09 Side swing angle 4.02° Cloud cover 0 Place Chengdu

[0204] Calculate the ambiguity index for each block according to steps 2 through 8. A total of 20,449 blocks were calculated.

[0205] Step 9: Set a threshold to determine whether the calculated fuzzy value for each block exceeds the threshold. Set the threshold to 0.40.

[0206] Step 10: Record the total number of blocks *m* exceeding the threshold. When the total number of blocks *m* > *Block*, the scene is considered to have a blur problem. Here, *Block* is the set threshold for the total number of blocks. *Block* is set to 5000. The calculation shows that *m* is 275, which does not exceed the set threshold; therefore, the scene image does not have a blur problem. Visual inspection confirms that the scene image indeed does not have a blur problem.

[0207] Step 11: Check if there is a problem with image data being strung together.

[0208] Steps 12-13: Calculate the evaluation index for each image data block.

[0209] Step 14: Set indicator thresholds and determine whether the calculated indicator value for each block exceeds the threshold. Thresholds are divided according to land cover type. Thresholds are set separately for high, medium, and low brightness land cover types. If the set threshold is exceeded, record the pixel location and the number of blocks. The contrast threshold for low brightness land cover is set to 1800, and the edge energy threshold is set to 650. The contrast threshold for medium brightness land cover is set to 3500, and the edge energy threshold is set to 2000. The contrast threshold for high brightness land cover is set to 12000, and the edge energy threshold is set to 3500.

[0210] Step 15: The panoramic image is divided into 71 blocks. Calculations show that the contrast and edge energy indicators in blocks 44, 45, and 46 are abnormal, exceeding the set thresholds. The calculation results are shown in Table 3 and meet the strategy conditions for continuous statistical analysis of abnormal blocks. Therefore, it is determined that the image data has a crosstalk problem. Based on the location of the abnormal blocks mapped to the image, the vertical coordinate of the crosstalk problem is between 17400 and 18200. Visual inspection confirms that the image does indeed have a crosstalk problem.

[0211] Table 3 Calculation Results of Indicators

[0212] Block ID Contrast Edge energy 40 312.440276 82.630906 41 304.013637 77.811991 42 309.139518 81.787408 43 377.337334 119.420981 44 2873.693336 966.645105 45 5771.280258 2132.584250 46 5990.191203 2073.756946 47 462.547972 117.352134 48 587.699195 167.024104 49 459.325963 113.599772 50 490.474593 149.404916 51 416.817969 105.176117 52 349.784452 77.280011 53 556.015238 154.700878 54 476.320886 115.504001 55 554.536579 139.510664 56 399.301844 92.852794 57 341.654184 86.482252 58 397.762944 114.014166 59 418.307689 110.449835 60 445.823791 125.141100

[0213] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or N embodiments or examples. Furthermore, those skilled in the art can combine and integrate the different embodiments or examples described in this specification and the features of different embodiments or examples without contradiction. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of the present invention, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified. Any process or method described in the flowcharts or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more N executable instructions for implementing custom logical functions or processes, and the scope of preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order according to the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain. The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection having one or N wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic device, and portable optical disc read-only memory (CDROM).Furthermore, the computer-readable medium can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory. It should be understood that various parts of the invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0214] The above description is merely a preferred embodiment of a referenceless remote sensing image quality assessment method. The scope of protection for a referenceless remote sensing image quality assessment method is not limited to the above embodiments; all technical solutions falling within this conceptual framework are within the scope of protection of this invention. It should be noted that for those skilled in the art, any improvements and variations made without departing from the principles of this invention should also be considered within the scope of protection of this invention.

Claims

1. A reference-free remote sensing image quality assessment method, characterized by: The method includes the following steps: Step 1: Extract the image data to be evaluated; Step 2: Perform block-based multi-threaded processing on the image data. First, calculate the blur of the extracted blocks. The blur is used to characterize the degree of blur of the image features. Step 3: Set a fuzziness threshold to determine if fuzziness issues exist; Step 4: Detect whether there is a crosstalk problem in the image data, calculate the crosstalk evaluation index. The crosstalk problem detection is evaluated by two indices together to improve the reliability of the detection. Step 5: Calculate the contrast ratio of the series evaluation indicators; Step 5 specifically involves: When there is a problem with image data being strung together, the index value will show a peak at the position of the strung pixels, resulting in obvious changes; The contrast ratio, a key performance indicator for series evaluation, is calculated using the following method: in, , These are the elements of the normalized gray-level co-occurrence matrix; The gray-level co-occurrence matrix is ​​the total number of times the gray value of a point and its adjacent points appear in a point pair on the image. δ represents the distance between the two points, where δ=1; θ represents the angle between the distance and the coordinate axis. Since the chaining problem occurs in the column direction, θ=0 here, meaning the pixel pair is in the horizontal direction. ； Step 6: Calculate the edge energy of the series evaluation index; Step 6 specifically involves: The edge energy calculation method is as follows: in: ； Step 7: Set the threshold for the serialization problem indicator, determine whether the calculated indicator value of each block exceeds the threshold, determine whether there is a serialization problem, and complete the inspection and evaluation.

2. The method according to claim 1, characterized in that: The specific steps of step 2, segmentation, are as follows: To detect image blurring, the detection results were compared with different block sizes, and the block size was fixed at 200×200 pixels; to detect image jamming caused by imaging electronics, the block size was set according to the jamming width. Block size Size = ⌊width / 2⌋, where width is the width of the string.

3. The method according to claim 1, characterized in that: Step 3 specifically includes the following steps: Step 3.1: For the image block F to be evaluated, the size is Each pixel is used to perform vertical and horizontal filtering to obtain the corresponding blurred image B: in, Image blocks F The blurred image obtained after vertical and horizontal filtering. The vertical and horizontal models of the filter are respectively expressed as follows: Step 3.2: Calculate the absolute error of the image block F before filtering. The absolute error of the filtered image patch This allows us to obtain the grayscale value changes of adjacent pixels. Subtract the absolute errors, i.e. and Make a mistake, and By subtracting the values, we can obtain the difference between the image patches before and after filtering. Where the range of values ​​for i is: The range of values ​​for j is ; Step 3.3: To compare the overall differences between image patches, calculate the absolute error of the image patches before and after filtering. , Summing yields the differences across the entire image patch: Step 3.4: Normalize the results to : Step 3.5: Select the calculated result and The maximum value among the values ​​is used as the final evaluation value: in, ,when A value of 0 indicates the lowest image blur and the best image clarity; when... A value of 1 indicates the highest image blur and the worst image clarity. A threshold is set to determine whether the calculated blur value of each block exceeds the threshold; the threshold is obtained statistically based on a large amount of experimental image data. Step 3.6: Record the total number of blocks m that exceeds the threshold. When the total number of blocks m > Block, the image is considered to have a blur problem, where Block is the set threshold for the total number of blocks.

4. The method according to claim 3, characterized in that: Step 7 specifically involves: Because of the overly bright cloud and snow features, the evaluation index values ​​of such features may exceed the set threshold. Therefore, a strategy of continuous statistical analysis of abnormal blocks is adopted to avoid misjudgment of such features. Block size Size = ⌊width / 2⌋, where width is the width of the sequence. According to this block division rule, the span of abnormal pixel areas is at most 3 blocks. This rule is used to avoid the problem of normal images being misjudged.

5. The method according to claim 4, characterized in that: Thresholds are set separately for different land cover types with high, medium, and low brightness values; when the set threshold is exceeded, the pixel location and the number of abnormal blocks are recorded.

6. A referenceless remote sensing image quality assessment system, characterized by: The system includes: A data extraction module, which extracts the image data to be evaluated; The data processing module performs block-based multi-threaded processing on the image data. The extracted blocks are first subjected to blur calculation, and the blur is used to characterize the degree of blur of the image features. A fuzziness threshold setting module, wherein the fuzziness threshold setting module sets a fuzziness threshold to determine whether a fuzziness problem exists; The detection module detects whether there is a crosstalk problem in the image data and calculates a crosstalk evaluation index. The crosstalk problem detection is evaluated by two indexes together to improve the reliability of the detection. An edge energy calculation module, which calculates the edge energy, a serial evaluation index; The evaluation module sets a threshold for the serialization problem index, determines whether the calculated index value of each block exceeds the threshold, determines whether a serialization problem exists, and completes the inspection and evaluation. When there is a problem with image data being strung together, the index value will show a peak at the position of the strung pixels, resulting in obvious changes; The contrast ratio, a key performance indicator for series evaluation, is calculated using the following method: in, , These are the elements of the normalized gray-level co-occurrence matrix; The gray-level co-occurrence matrix is ​​the total number of times the gray value of a point and its adjacent points appear in a point pair on the image. δ represents the distance between the two points, where δ=1; θ represents the angle between the distance and the coordinate axis. Since the chaining problem occurs in the column direction, θ=0 here, meaning the pixel pair is in the horizontal direction. ； The edge energy calculation method is as follows: in: 。 7. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the method as described in any one of claims 1-5.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the method according to any one of claims 1-5.

Images