A method, device, electronic device and storage medium for evaluating the degree of haze

Abstract

The present invention relates to a method, device, electronic device and storage medium for evaluating the haze degree, including the following steps: S1. Traverse the haze image through a preset sliding window, and respectively obtain the dark channel model corresponding to the haze image through minimization operation and the bright channel model corresponding to the haze image through maximization operation based on the atmospheric scattering model; S2. Obtain the first difference information between the bright channel information of the original image and the bright channel information of the haze image based on the dark channel model and the bright channel model, wherein the first difference information is a first function including the dark channel information of the haze image and the bright channel information of the haze image; S3. Obtain the mean value of the first difference information as the first evaluation index based on the traversal process and the first function, so as to obtain the haze degree evaluation result according to the first evaluation index. Implementing the present invention simplifies the haze degree evaluation process while ensuring the haze degree evaluation result.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and more particularly, to a method, apparatus, electronic device, and storage medium for evaluating the haze degree. Background Art

[0002] The evaluation of the haze degree has broad application prospects. For example, by evaluating the haze degree of an image to analyze visibility, the driving safety of driverless vehicles can be ensured. Another example is that at locations such as deserts and mountaintops where it is difficult to install air quality monitoring equipment, by evaluating the haze degree in surveillance videos, real-time monitoring of air quality can be achieved.

[0003] In terms of image processing, the evaluation of the haze degree can be classified into the category of haze image quality evaluation. Similar to general image quality evaluation methods, haze image quality evaluation is also divided into subjective methods and objective methods. Among them, subjective image quality evaluation takes a long time and is difficult to be applied in real time in embedded devices, so it cannot be directly applied to the field of video surveillance. In objective haze image quality evaluation methods, since there is no corresponding original haze-free image for the haze image itself, the focus of research should be the problem of no-reference haze image quality evaluation.

[0004] However, the currently adopted haze degree evaluation methods mainly use traditional full-reference image quality evaluation algorithms for haze image quality evaluation. The common haze image quality evaluation algorithms are described as follows.

[0005] (1) Contrast method, which belongs to the no-reference image quality evaluation method. It mainly uses the mean value of local contrast of the image to evaluate the quality of the dehazed image, as shown in the following formula

[0006]

[0007] where C is the mean value of the contrast of all 3×3 image blocks, max(i) is the highest gray value in the current image block, min(i) is the lowest gray value in the current image block, and n is the total number of image blocks.

[0008] (2) Point sharpness method, which belongs to the no-reference image quality evaluation method. The higher its value, the better the image quality evaluation result. The greater the change in edge gray level, the higher the clarity and the lower the haze degree. Therefore, the image quality can be evaluated by statistically calculating the point sharpness. The formula is as follows:

[0009]

[0010] where dI / dx represents the gray derivative in the edge direction, and I(b) - I(a) represents the overall gray change in the edge direction.

[0011] This method only counts specific image regions, and these regions need to be manually selected, which is not conducive to automation.

[0012] (3) Entropy method, which also belongs to the no-reference image quality assessment method. The greater the entropy of an image, the better the image quality. Image entropy is based on statistical features and is used to measure the richness of image information. It is an important indicator for measuring the amount of information in an image. The formula is as follows:

[0013]

[0014] where P i is the probability that a pixel with a gray value of i appears in the image, and L is the total number of gray levels.

[0015] This method has a high sensitivity and is not likely to produce evaluation results contrary to the haze level.

[0016] (4) Gray scale difference method (SMD), which also belongs to the no-reference image quality assessment method. The greater the gray scale variance value of an image, the better the image quality. The lower the haze level of an image, the more high-frequency components it has. Therefore, the change in gray scale can be used as the basis for evaluating the quality of haze images. The formula is as follows:

[0017] SMD = ∑ y ∑ x (f(x, y) - f(x, y - 1) + |f(x, y) - f(x + 1, y)|)

[0018] where f(x, y) represents the gray value of the pixel at coordinates (x, y) on the image.

[0019] The gray scale variance evaluation method is convenient and fast to calculate. Its disadvantage is that it is not highly sensitive in areas with dense gradients.

[0020] (5) Method based on the atmospheric scattering model. The haze level is evaluated by the brightness of the image and the amount of original information received, and these two are respectively related to the atmospheric light A and the atmospheric transmittance t(x). The formula is as follows:

[0021]

[0022] where A(i, j) and t(i, j) are the atmospheric light and the atmospheric transmittance at coordinates (i, j) respectively, and W and H are the width and height of the image respectively.

[0023] In the existing solutions (1)-(4), the average gradient or point sharpness of the image, etc. are used as indicators to evaluate the haze level of the image. Such solutions fail to consider the physical model in the degradation process of the haze image and only evaluate the haze level from the haze image itself. Moreover, images in different scenarios have different average gradient or point sharpness characteristics. Therefore, it is difficult to generalize such methods to the comparison of the quality of haze images in different scenarios.

[0024] Solution (5) is a haze image quality evaluation method based on prior information and the atmospheric scattering model. The haze degree of the image is evaluated through two parameters, namely, the atmospheric light and the atmospheric transmittance. However, it is not easy to design an explicit expression using these two parameters.

[0025] In summary, there is still a large room for improvement in the haze image quality evaluation performance of the existing solutions, and it is necessary to improve the haze image quality evaluation method and the scope of image acquisition. Summary of the Invention

[0026] The technical problem to be solved by the present invention is to provide a haze degree evaluation method, device, electronic device and storage medium.

[0027] The technical solution adopted by the present invention to solve its technical problems is to construct a haze degree evaluation method, including the following steps:

[0028] S1. Traverse the haze image through a preset sliding window, and based on the atmospheric scattering model, respectively obtain the dark channel model corresponding to the haze image through a minimization operation and the bright channel model corresponding to the haze image through a maximization operation;

[0029] S2. Obtain the first difference information between the bright channel information of the original image and the bright channel information of the haze image based on the dark channel model and the bright channel model, where the first difference information is a first function including the dark channel information of the haze image and the bright channel information of the haze image;

[0030] S3. Obtain the mean value of the first difference information as the first evaluation index based on the traversal process and the first function, so as to obtain the haze degree evaluation result according to the first evaluation index.

[0031] Preferably, in the haze degree evaluation method of the present invention, it further includes:

[0032] S4. Obtain the second difference information between the bright channel information of the original image and the bright channel information of the haze image based on the dark channel model and the bright channel model, where the second difference information is a second function including the dark channel information of the haze image and the bright channel information of the haze image;

[0033] S5. Obtain the global threshold of the second function based on the second difference information, and use the global threshold as the second evaluation index to obtain the haze degree evaluation result according to the second evaluation index.

[0034] Preferably, in the haze degree evaluation method of the present invention, it further includes:

[0035] S6. Obtain a third function according to the first function and the second function;

[0036] S7. Obtain the third evaluation index of the haze image based on the traversal process and the third function, and obtain the haze degree evaluation result according to the third evaluation index.

[0037] Preferably, in the haze degree evaluation method of the present invention, in the step S2, the first function is

[0038]

[0039] where J b (x) is the bright channel information of the original image, I b (x) is the bright channel information of the haze image, I d (x) is the dark channel information of the haze image, and A is the atmospheric light;

[0040] In the step S3, the first evaluation index satisfies the following function:

[0041]

[0042] where DBCP-I is the first evaluation index, and w and h are the width and height of the haze image respectively.

[0043] Preferably, in the haze degree evaluation method of the present invention, in the step S4, the second function is:

[0044]

[0045] where C db (x) is the second difference information;

[0046] In the step S5, the second evaluation index satisfies the following function:

[0047]

[0048] where DBCP-II is the second evaluation index, and T[·] represents the global threshold of C db (x) obtained by the maximum inter-class variance method.

[0049] Preferably, in the haze degree evaluation method of the present invention, in the step S6, the third function is

[0050]

[0051] In the step S7, the third evaluation index satisfies the following function:

[0052]

[0053] where DBCP-III is the third evaluation index, and Ω l is the region in the haze image where C db (x) is less than the preset value, and Ω h is the region in the haze image where C db (x) is greater than or equal to the preset value, is the preset atmospheric transmittance corresponding to the haze image.

[0054] Preferably, in the haze degree evaluation method of the present invention, the preset value is the global threshold.

[0055] In addition, the present invention also constructs a haze degree evaluation device, including:

[0056] A model establishment unit, configured to traverse a haze image through a preset sliding window, and based on an atmospheric scattering model, respectively obtain a dark channel model corresponding to the haze image through a minimization operation and a bright channel model corresponding to the haze image through a maximization operation;

[0057] A first function establishment unit, configured to obtain first difference information between the bright channel information of the original image and the bright channel information of the haze image based on the dark channel model and the bright channel model, where the first difference information is a first function including the dark channel information of the haze image and the bright channel information of the haze image;

[0058] A first evaluation index acquisition unit, configured to obtain the mean value of the first difference information as a first evaluation index based on the traversal process and the first function, so as to obtain the haze degree evaluation result according to the first evaluation index.

[0059] Preferably, in the haze degree evaluation device of the present invention, it further includes:

[0060] A second function establishment unit, configured to obtain second difference information between the bright channel information of the original image and the bright channel information of the haze image based on the dark channel model and the bright channel model, where the second difference information is a second function including the dark channel information of the haze image and the bright channel information of the haze image;

[0061] A second evaluation index acquisition unit, configured to obtain a global threshold of the second function based on the second difference information, and use the global threshold as a second evaluation index to obtain a haze degree evaluation result according to the second evaluation index.

[0062] Preferably, the haze degree evaluation device according to the present invention further includes:

[0063] A third function establishment unit, configured to obtain a third function according to the first function and the second function;

[0064] A third evaluation index acquisition unit, configured to obtain a third evaluation index of the haze image based on a traversal process and the third function, and obtain a haze degree evaluation result according to the third evaluation index.

[0065] In addition, the present invention also constructs a computer storage medium, on which a computer program is stored, characterized in that when the computer program is executed by a processor, the haze degree evaluation method described in any one of the above is implemented.

[0066] In addition, the present invention also constructs an electronic device, characterized in that it includes a memory and a processor;

[0067] The memory is used to store a computer program;

[0068] The processor is used to execute the computer program to implement the haze degree evaluation method described in any one of the above.

[0069] Implementing a haze degree evaluation method, device, electronic device and storage medium of the present invention has the following beneficial effects: while simplifying the haze degree evaluation process, ensuring the haze degree evaluation result. Description of the Drawings

[0070] The present invention will be further described below in conjunction with the drawings and embodiments. In the drawings:

[0071] Figure 1 is a program flowchart of an embodiment of a haze degree evaluation method of the present invention;

[0072] Figure 2 is a program flowchart of another embodiment of a haze degree evaluation method of the present invention;

[0073] Figure 3 is a program flowchart of another embodiment of a haze degree evaluation method of the present invention;

[0074] Figure 4 is a logic block diagram of an embodiment of a haze degree evaluation device of the present invention;

[0075] Figure 5It is the logic block diagram of another embodiment of a haze degree evaluation device of the present invention;

[0076] Figure 6 It is the logic block diagram of another embodiment of a haze degree evaluation device of the present invention. Detailed implementation manners

[0077] For a clearer understanding of the technical features, objectives, and effects of the present invention, the detailed implementation manners of the present invention will now be described in detail with reference to the accompanying drawings.

[0078] As Figure 1 shown, in the first embodiment of a haze degree evaluation method of the present invention, the following steps are included: S1. Traverse the haze image through a preset sliding window, and based on the atmospheric scattering model, respectively obtain the dark channel model corresponding to the haze image through a minimization operation and the bright channel model corresponding to the haze image through a maximization operation; specifically, in the field of haze image quality evaluation, the atmospheric scattering process is described by the atmospheric scattering model, and the mathematical expression of this model is as follows:

[0079] I(x) = J(x)t(x) + A(1 - t(x)) (1)

[0080] where I(x) is the haze image, J(x) is the scene reflected light, t(x) is the atmospheric transmittance, and A is the atmospheric light. Traverse the haze image through a preset sliding window, and during the traversal process, the preset atmospheric transmittance is a constant and denoted as The atmospheric light of the entire scene is constant and the same in all channels and denoted as A.

[0081] Calculate the dark channel prior on both sides of Equation (1), that is, perform the minimum value operation as shown in the following formula:

[0082]

[0083] And according to the data corresponding to the dark channel prior (Dark Channel Prior, DCP), obtain the mathematical expression of its haze-free image as follows:

[0084]

[0085] where J d (x) is the dark channel of the image J(x), J c is a color channel of J, Ω(x) is the window at pixel point x, and {r, g, b} are the three color channels of red, green, and blue respectively.

[0086] The dark channel model corresponding to the haze image can be obtained by transforming Formula (2) with the above formula:

[0087]

[0088] Similarly, calculate the bright channel prior on both sides of Equation (1) simultaneously, that is, perform the maximum value operation and obtain the mathematical expression of its haze-free image according to the data of the bright channel prior (BCP) as follows:

[0089]

[0090] where J b (x) is the bright channel of the image J(x). For an 8-bit color image, J b has a maximum value of 255.

[0091] Finally, obtain the bright channel model corresponding to the haze image:

[0092]

[0093] S2. Obtain the first difference information between the bright channel information of the original image and the bright channel information of the haze image based on the dark channel model and the bright channel model, where the first difference information is a first function including the dark channel information and the bright channel information of the haze image; specifically, obtain the first function that can reflect the first difference information according to the obtained dark channel model and bright channel model.

[0094] In one embodiment, the obtained first function is

[0095]

[0096] where J b (x) is the bright channel information of the original image, I b (x) is the bright channel information of the haze image, I d (x) is the dark channel information of the haze image, and A is the atmospheric light.

[0097] S3. Obtain the mean value of the first difference information as the first evaluation index based on the traversal process and the first function, so as to obtain the haze degree evaluation result according to the first evaluation index. Specifically, since the dark channel and bright channel information of the haze image can be calculated, and the atmospheric light can also be estimated, the difference between the haze image and the original image on the bright channel can be calculated. Calculate this difference over the entire image and take the average to obtain the haze level evaluation index DBCP-I.

[0098] In one embodiment, in step S3, the first evaluation index satisfies the following function:

[0099]

[0100] Among them, DBCP-I is the first evaluation index, and w and h are the width and height of the haze image respectively.

[0101] As Figure 2 shown, in an embodiment, the haze degree evaluation method of the present invention further includes: S4. Obtain the second difference information between the bright channel information of the original image and the bright channel information of the haze image based on the dark channel model and the bright channel model, where the second difference information is a second function including the dark channel information and the bright channel information of the haze image; specifically, obtain the second function that can reflect the second difference information based on the obtained dark channel model and bright channel model.

[0102] In an embodiment, the process of obtaining the second function can be based on formula (6), and the process is as follows.

[0103]

[0104]

[0105]

[0106] Among them, C db (x) is the contrast information between the dark channel I d (x) and the bright channel I b (x) of the haze image, is the estimated atmospheric transmittance. The final formula (9) corresponds to the second function reflecting the second difference information between the bright channel information of the original image and the bright channel information of the haze image, and the corresponding evaluation index is obtained through the second function.

[0107] S5. Obtain the global threshold of the second function based on the second difference information, and use the global threshold as the second evaluation index to obtain the haze degree evaluation result according to the second evaluation index.

[0108] Since it can be seen from formula (8) that the contrast information between the dark channel and the bright channel can be used for haze level evaluation. Regarding this contrast information as an image and performing threshold segmentation on it, the haze level evaluation index DBCP-II, that is, the second evaluation index, can be obtained, as shown in the following formula:

[0109]

[0110] Among them, T[·] represents calculating the global threshold of C db (x) using the Otsu method.

[0111] As Figure 3As shown, in one embodiment, the haze degree evaluation method of the present invention further includes: S6. Obtaining a third function according to the first function and the second function; S7. Obtaining a third evaluation index of the haze image based on the traversal process and the third function, so as to obtain the haze degree evaluation result according to the third evaluation index. Specifically, in the function corresponding to the second evaluation index, the contrast information C db in the haze image is divided into two parts, namely the low-contrast region Ω l and the high-contrast region Ω h . For the pixels in the low-contrast region Ω l , when calculating using Equation (8), both the numerator and the denominator are very small, which will bring more noise and uncertainty to the result of the division. Therefore, by obtaining its second evaluation index DBCP-II for determination, the result is more accurate. And in the high-contrast region Ω h , the pixels generally represent the scene details and content. Therefore, by obtaining its first evaluation index DBCP-I for determination, the difference between the dark channel and the bright channel can be obtained more accurately. Therefore, based on the process of establishing the first evaluation index and the second evaluation index, a third function is obtained through mathematical processing according to the first function and the second function, and a third evaluation index is established through the third function, so as to evaluate the haze degree of the image through the third evaluation index.

[0112] In one embodiment, the process of obtaining the third evaluation index is as follows,

[0113]

[0114]

[0115] where DBCP-III is the third evaluation index, Ω l is the region in the haze image where C db (x) is less than the preset value, and Ω h is the region in the haze image where C db (x) is greater than or equal to the preset value, is the preset atmospheric transmittance corresponding to the haze image. That is, Equation (12) corresponds to the third function, which is obtained according to the first function and the second function, and according to the traversal result of the third function, the third evaluation index DBCP-III is finally obtained, and its corresponding expression is Equation (13).

[0116] Furthermore, for the division of the low-contrast region Ω db and the high-contrast region Ω l in the contrast information C h in the haze image, it can be based on the global threshold T[C dbCalibrate [(x)]. When the comparison information C db (x) is higher than this global threshold, then define this area as the high-contrast area Ω h , when the comparison information C db (x) is lower than this global threshold, then define this area as the low-contrast area Ω l .

[0117] In addition, as Figure 4 shown, a haze degree evaluation device of the present invention includes a model establishment unit 110, which is used to traverse a haze image through a preset sliding window, and based on the atmospheric scattering model, respectively obtain the dark channel model corresponding to the haze image through a minimization operation and obtain the bright channel model corresponding to the haze image through a maximization operation; a first function establishment unit 121, which is used to obtain the first difference information between the bright channel information of the original image and the bright channel information of the haze image based on the dark channel model and the bright channel model, where the first difference information is a first function including the dark channel information of the haze image and the bright channel information of the haze image; a first evaluation index acquisition unit 131, which is used to obtain the mean value of the first difference information as the first evaluation index based on the traversal process and the first function, so as to obtain the haze degree evaluation result according to the first evaluation index.

[0118] Optionally, as Figure 5 shown, a haze degree evaluation device of the present invention further includes: a second function establishment unit 122, which is used to obtain the second difference information between the bright channel information of the original image and the bright channel information of the haze image based on the dark channel model and the bright channel model, where the second difference information is a second function including the dark channel information of the haze image and the bright channel information of the haze image; a second evaluation index acquisition unit 132, which is used to obtain the global threshold of the second function based on the second difference information, and use the global threshold as the second evaluation index, so as to obtain the haze degree evaluation result according to the second evaluation index.

[0119] Optionally, as Figure 6 shown, a haze degree evaluation device of the present invention further includes: a third function establishment unit 123, which is used to obtain a third function according to the first function and the second function; a third evaluation index acquisition unit 133, which is used to obtain the third evaluation index of the haze image based on the traversal process and the third function, so as to obtain the haze degree evaluation result according to the third evaluation index.

[0120] Specifically, the specific cooperation operation process between the units of the haze degree evaluation device here can specifically refer to the above haze degree evaluation method, which will not be elaborated here.

[0121] In addition, an electronic device according to the present invention includes a memory and a processor; the memory is used for storing a computer program; the processor is used for executing the computer program to implement any of the above haze degree evaluation methods. Specifically, according to an embodiment of the present invention, the process described with reference to the flowchart above can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes program codes for executing the method shown in the flowchart. In such an embodiment, when the computer program is downloaded and installed by the electronic device and executed, it executes the above functions defined in the method of the embodiment of the present invention. The electronic device in the present invention can be a terminal such as a notebook, a desktop computer, a tablet computer, a smart phone, etc., or a server.

[0122] In addition, a computer storage medium according to the present invention has a computer program stored thereon, and when the computer program is executed by a processor, it implements any of the above haze degree evaluation methods. Specifically, it should be noted that the above computer-readable medium in the present invention can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, the computer-readable storage medium can be any tangible medium that contains or stores a program, and the program can be used by or combined with an instruction execution system, apparatus, or device. In the present invention, the computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program codes. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium can also be any computer-readable medium other than the computer-readable storage medium, and the computer-readable signal medium can send, propagate, or transmit a program for use by or combined with an instruction execution system, apparatus, or device. The program codes contained on the computer-readable medium can be transmitted by any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0123] The above computer-readable medium can be included in the above electronic device; or it can exist separately and not be assembled into the electronic device.

[0124] In a specific implementation process, the performance of its evaluation results is compared with that of a general no-reference image quality assessment method. Among them, the methods in the first group directly extract features related to the haze level from haze images, such as the C, R, and entropy methods, etc. The methods in the second group include pre-trained machine learning and deep learning models, and the application scenarios of these models include natural images, screen content images, and contrast distortion images, etc. The second group of models includes methods such as NIQE, dipIQ, and MEON, etc. The methods in the third group include two deep neural network models, CNN and DIQaM-NR. 80% of the data in the dataset is used for training, and 20% of the data is used for testing. In this implementation process, the window sizes for calculating the dark channel and the bright channel are both 5.

[0125] The performance comparison mainly uses two metrics, the Spearman rank correlation coefficient (SROCC) and the Pearson correlation coefficient (PCC). Among them, SROCC is used to evaluate the monotonicity of the prediction, and PCC is used to evaluate the linear correlation between the prediction score and the subjective quality score. The closer these two coefficients are to 1, the better the algorithm performance. Their calculation methods are as follows:

[0126]

[0127]

[0128] where M is the number of images, x i and y i are respectively the subjective evaluation score and the prediction score of the i-th image, and are respectively the mean of the subjective evaluation scores and the mean of the prediction scores, and d i is the difference between the ranks of the subjective evaluation score and the prediction score of the i-th image.

[0129] The RHID_AQI dataset is a dataset designed specifically for the haze level evaluation task and contains the subjective evaluation scores of haze images. Based on this dataset, the performances of the above-mentioned various methods are compared, and the results are shown in Table 1.

[0130] Table 1 Comparison of SROCC and PCC on the RHID_AQI dataset

[0131]

[0132]

[0133] According to Table 1, considering the comprehensive SROCC and PCC, the effect of the second evaluation index DBCP-II is the best, and the models with better effects are concentrated in the deep learning methods and the DBCP method described in this application, that is, the index process.

[0134] To further compare the performance of each model in different scenarios, the SROCC of the above models on different RHID_AQI scenario subsets are shown in Table 2. Hit count in the table indicates the number of times the model ranks in the top three. It can be seen from Table 2 that the performance of the index evaluation process of this application is still excellent, especially the third evaluation index evaluation process DBCP-Ⅲ algorithm ranks in the top three in the six scenario subsets, which is better than the deep learning algorithm.

[0135] Table 2 SROCC comparison on different scene subsets of RHID_AQI dataset

[0136]

[0137]

[0138] In order to prove the generalization of the index evaluation process of this application, namely the DBCP method, a comparative experiment was conducted based on the exBeDDE dataset. Since the subjective evaluation scores of the exBeDDE dataset are only valid within the group of each scene, in this comparison process, only the performance of the algorithms in a single scene was compared. The results are shown in Table 3. As can be seen from Table 3, the third evaluation index evaluation process DBCP-Ⅲ is still the best performing algorithm.

[0139] Table 3 Comparison of SROCC and PCC on the RHID_AQI dataset

[0140]

[0141]

[0142]

[0143]

[0144] Therefore, it can be seen that the haze degree assessment process of the present application is far superior to the existing general assessment process. Table 4 is a description of the English terms involved in the present application.

[0145] Table 4 Chinese-English comparison table of proper nouns

[0146]

[0147]

[0148] It can be understood that the above embodiments only represent the preferred embodiments of the present invention, and the description thereof is relatively specific and detailed. However, it should not be construed as a limitation on the scope of the patent for the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, the above technical features can be freely combined, and several modifications and improvements can also be made, which all fall within the protection scope of the present invention. Therefore, all equivalent transformations and modifications made to the scope of the claims of the present invention shall fall within the scope covered by the claims of the present invention.

Claims

1. A method for evaluating the haze level, characterized in that, It includes the following steps: S1. Traverse the haze image through a preset sliding window, and based on the atmospheric scattering model, respectively obtain the dark channel model corresponding to the haze image through a minimization operation and the bright channel model corresponding to the haze image through a maximization operation; S2. Obtain the first difference information between the bright channel information of the original image and the bright channel information of the haze image based on the dark channel model and the bright channel model, where the first difference information is a first function including the dark channel information of the haze image and the bright channel information of the haze image, and the original image is the haze-free image corresponding to the haze image; S3. Obtain the mean value of the first difference information as the first evaluation index based on the traversal process and the first function, so as to obtain the haze degree evaluation result according to the first evaluation index; wherein, the first function is Among them, is the bright channel information of the original image, is the bright channel information of the haze image, is the dark channel information of the haze image, A is the atmospheric light; The first evaluation index satisfies the following function: -Ⅰ ； Among them, DBCP-I is the first evaluation index, and are the width and height of the haze image, respectively.

2. The haze degree evaluation method according to claim 1, wherein The method further includes: S4. Obtain the second difference information between the bright channel information of the original image and the bright channel information of the haze image based on the dark channel model and the bright channel model, where the second difference information is a second function including the dark channel information of the haze image and the bright channel information of the haze image; S5. Obtain the global threshold of the second function based on the second difference information, and use the global threshold as the second evaluation index, so as to obtain the haze degree evaluation result according to the second evaluation index.

3. The haze degree evaluation method according to claim 2, characterized in that, The method further includes: S6. Obtain a third function according to the first function and the second function; S7. Obtain the third evaluation index of the haze image based on the traversal process and the third function, so as to obtain the haze degree evaluation result according to the third evaluation index.

4. The haze degree evaluation method according to claim 3, wherein In the step S4, the second function is: Among them, the is the second difference information; In the step S5, the second evaluation index satisfies the following function: DBCP-Ⅱ Among them, DBCP-Ⅱ is the second evaluation index, which represents the global threshold obtained by the Otsu method for 5. The haze degree evaluation method according to claim 4, characterized in that, In the step S6, the third function is In the step S7, the third evaluation index satisfies the following function: . wherein, DBCP-Ⅲ is the third evaluation index, is the area in the haze image less than the preset value, is the area in the haze image greater than or equal to the preset value, is the preset atmospheric transmittance corresponding to the haze image.

6. The haze degree evaluation method according to claim 5, characterized in that, The preset value is the global threshold.

7. An apparatus for evaluating the degree of haze, characterized in that, It includes: A model establishment unit, configured to traverse the haze image through a preset sliding window, and based on the atmospheric scattering model, respectively obtain the dark channel model corresponding to the haze image through a minimization operation and the bright channel model corresponding to the haze image through a maximization operation; A first function establishment unit, configured to obtain the first difference information between the bright channel information of the original image and the bright channel information of the haze image based on the dark channel model and the bright channel model, where the first difference information is a first function including the dark channel information of the haze image and the bright channel information of the haze image, and the original image is the haze-free image corresponding to the haze image; A first evaluation index acquisition unit, configured to obtain the mean value of the first difference information as the first evaluation index based on the traversal process and the first function, so as to obtain the haze degree evaluation result according to the first evaluation index; wherein, the first function is Among them, is the bright channel information of the original image, is the bright channel information of the hazy image, is the dark channel information of the hazy image, A is the atmospheric light; The first evaluation index satisfies the following function: -Ⅰ ； Among them, DBCP-I is the first evaluation index, and are the width and height of the haze image respectively.

8. A computer storage medium, on which a computer program is stored, characterized in that, When the computer program is executed by a processor, it implements the haze degree evaluation method according to any one of claims 1-6.

9. An electronic device, characterized in that, It includes a memory and a processor; The memory is used for storing a computer program; The processor is used for executing the computer program to implement the haze degree evaluation method according to any one of claims 1-6.

Images