Water area detection method and device based on multi-attribute consistency

Abstract

The present invention provides a water area detection method and device based on multi-attribute consistency, wherein the method includes: determining a multidimensional water area attribute characteristic value of at least one water area feature point in a surface image; determining a multidimensional water area attribute characteristic value of multiple points to be detected in the surface image; determining all points to be detected whose similarity with the multidimensional water area attribute characteristic value of at least one water area feature point is greater than or equal to a preset threshold value based on the multidimensional water area attribute characteristic value of at least one water area feature point and the multidimensional water area attribute characteristic value of multiple points to be detected in the surface image; obtaining a water area detection result of the surface image based on the water area of ​​at least one water area feature point and the water area of ​​all points to be detected whose similarity with the multidimensional water area attribute characteristic value of at least one water area feature point is greater than or equal to a preset threshold value. The above technical solution realizes rapid and accurate water area detection, and improves the efficiency and accuracy of water area detection.

Description

Technical Field

[0001] The present invention relates to the technical field of petroleum geophysical prospecting, and in particular to a water area detection method and device based on multi-attribute consistency. Background Art

[0002] In the process of seismic exploration and construction, field surveys and indoor surveys are required before the observation system is designed to mark the obstacles on the surface, that is, to mark the obstacle areas in the exploration area where blasting is not allowed or not suitable. Among them, water areas, such as rivers, ponds, lakes, etc., are an important content that needs to be calibrated. They are an obstacle that needs to be calibrated in large quantities when geophysical exploration is carried out in rainy areas such as southern my country. The calibration of water obstacles mainly combines the actual survey results with the results of manual calibration on indoor satellite images. Manual calibration on satellite images is monotonous, time-consuming, inefficient, and prone to errors.

[0003] Currently, no effective solution has been proposed to the above problems. Summary of the invention

[0004] The embodiment of the present invention provides a water area detection method based on multi-attribute consistency, which is used to quickly and accurately perform water area detection. The method includes:

[0005] Determine a multidimensional water area attribute characteristic value of at least one water area characteristic point in the surface image;

[0006] Determine the multi-dimensional water area attribute characteristic values ​​of multiple points to be detected in the surface image;

[0007] According to the multidimensional water area attribute characteristic value of at least one water area feature point and the multidimensional water area attribute characteristic values ​​of multiple points to be detected in the surface image, all points to be detected whose similarity with the multidimensional water area attribute characteristic value of at least one water area feature point is greater than or equal to a preset threshold are determined;

[0008] The water area detection result of the surface image is obtained based on the water area of ​​at least one water area feature point and the water areas of all points to be detected whose similarity with the multi-dimensional water area attribute characteristic value of at least one water area feature point is greater than or equal to a preset threshold.

[0009] The embodiment of the present invention further provides a water area detection device based on multi-attribute consistency, which is used to quickly and accurately detect water areas. The device includes:

[0010] A first determining unit is used to determine a multi-dimensional water area attribute characteristic value of at least one water area characteristic point in a surface image;

[0011] The second determination unit is used to determine the multi-dimensional water area attribute characteristic values ​​of multiple points to be detected in the surface image;

[0012] A detection unit, configured to determine, based on the multidimensional water area attribute characteristic value of at least one water area feature point and the multidimensional water area attribute characteristic values ​​of a plurality of points to be detected in the surface image, all points to be detected whose similarity to the multidimensional water area attribute characteristic value of at least one water area feature point is greater than or equal to a preset threshold;

[0013] A processing unit is used to obtain the water area detection result of the surface image based on the water area of ​​at least one water area feature point and the water areas of all points to be detected whose similarity with the multi-dimensional water area attribute characteristic value of at least one water area feature point is greater than or equal to a preset threshold.

[0014] An embodiment of the present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned water area detection method based on multi-attribute consistency when executing the computer program.

[0015] An embodiment of the present invention further provides a computer-readable storage medium, which stores a computer program for executing the above-mentioned water area detection method based on multi-attribute consistency.

[0016] The embodiment of the present invention provides a technical solution by: determining a multidimensional water area attribute characteristic value of at least one water area feature point selected by a user in a surface image; determining the multidimensional water area attribute characteristic values ​​of multiple points to be detected in the surface image; determining all points to be detected whose similarity with the multidimensional water area attribute characteristic value of at least one water area feature point is greater than or equal to a preset threshold value based on the multidimensional water area attribute characteristic value of at least one water area feature point and the multidimensional water area attribute characteristic values ​​of multiple points to be detected in the surface image; obtaining a water area detection result of the surface image based on the water area of ​​at least one water area feature point and the water area of ​​all points to be detected whose similarity with the multidimensional water area attribute characteristic value of at least one water area feature point is greater than or equal to the preset threshold value, thereby realizing fast and accurate water area detection, improving the efficiency and accuracy of water area detection, and thereby improving the efficiency and accuracy of water area obstacle calibration. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 is a flow chart of a water area detection method based on multi-attribute consistency in an embodiment of the present invention;

[0019] Figure 2ais a schematic diagram of the principle of water area detection based on multi-attribute consistency in an embodiment of the present invention;

[0020] Figure 2b is a schematic diagram of seed points and related concepts in an embodiment of the present invention;

[0021] Figure 3a This is a schematic diagram of standard test results in an example;

[0022] Figure 3b is a schematic diagram of the effect of water area detection based on multi-attribute consistency in an embodiment of the present invention;

[0023] Figure 4a This is another example of a standard test result diagram;

[0024] Figure 4b is a schematic diagram of the effect of water area detection based on multi-attribute consistency in another embodiment of the present invention;

[0025] Figure 5 It is a structural schematic diagram of a water area detection device based on multi-attribute consistency in an embodiment of the present invention. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] Since the inventors have considered the technical problems existing in the water area detection schemes in the prior art, a water area detection scheme based on multi-attribute consistency is proposed. This scheme is a detection technology and method for water areas proposed in the process of calibrating water area obstacles in surface images during indoor exploration in the field of petroleum geophysical exploration. This scheme calibrates several water area feature points on satellite images or aerial images, that is, through the seed points of given water areas, it can automatically check out and calibrate the water areas where these seed points are located or similar water areas. Specifically, after the user adds water area seed points to the surface image (such as satellite images, aerial images), the multi-dimensional attribute features of each point to be detected in the surface image, such as color, gradient, etc., are automatically calculated, and the similarity between the attributes of these points to be detected and the attributes of the seed points is calculated to find the water area containing the attribute features of the seed points.

[0028] Figure 1 is a flow chart of a water area detection method based on multi-attribute consistency in an embodiment of the present invention, such as Figure 1 As shown, the method comprises the following steps:

[0029] Step 101: determining a multi-dimensional water area attribute characteristic value of at least one water area characteristic point in a surface image;

[0030] Step 102: determining multi-dimensional water area attribute feature values ​​of a plurality of points to be detected in the surface image;

[0031] Step 103: determining all the points to be detected whose similarity with the multidimensional water area attribute characteristic value of at least one water area feature point is greater than or equal to a preset threshold value according to the multidimensional water area attribute characteristic value of at least one water area feature point and the multidimensional water area attribute characteristic values ​​of a plurality of points to be detected in the surface image;

[0032] Step 104: Obtain the water area detection result of the surface image based on the water area of ​​at least one water area feature point and the water areas of all points to be detected whose similarity with the multi-dimensional water area attribute characteristic value of at least one water area feature point is greater than or equal to a preset threshold.

[0033] The following is combined with Figure 2a and Figure 4b , each step involved in the embodiment of the present invention is described in detail as follows.

[0034] 1. First, the preparation steps before the above step 101 are introduced.

[0035] Before the above step 101, a step of acquiring a surface image may be included. After acquiring the surface image, the user opens the surface image and calibrates the seed points. Therefore, the water feature points, namely the seed points, may be selected by the user in the surface image.

[0036] When implementing it, Figure 2a As shown, load and open the surface image, and the user can use the mouse to calibrate the seed points of the water area on the image to obtain one or more seed points, as shown in the following formula (1). Set the threshold discrimination parameters of the similarity degree, etc.

[0037]

[0038] Among them, P s is the set of seed points, s is the number of seed points, p s (x,y) is the seed point p s The coordinates (x,y) of

[0039] When implementing it, Figure 2bAs shown in the figure, the red circle is a selected seed point, and the 3×3 (or 5×5 or other size) grid points around the seed point are the small neighborhood where the seed point is located (the preset neighborhood where the seed point is located). Neighborhood means area and range, and the small neighborhood is an area around a range, a small adjacent area. For the concept of the point to be detected and its small neighborhood (preset neighborhood) mentioned in the following steps, please refer to the above explanation of the concept of seed point and its small neighborhood.

[0040] 2. Next, the above step 101 is introduced.

[0041] In the specific implementation, after selecting the seed points, it is necessary to calculate the multidimensional attribute characteristic values ​​of these seed points, and then use these multidimensional attribute characteristic values ​​as a standard to find similar data or areas. Specifically, for the detailed method of calculating the multidimensional attribute characteristic values ​​of these seed points, please refer to the detailed method of calculating the multidimensional attribute characteristic values ​​of the points to be detected in step 102 below.

[0042] 3. Next, the above step 102 is introduced.

[0043] In one embodiment, the multi-dimensional water area attribute feature may include any combination of the following features: color attribute feature, gradient attribute feature, statistical attribute feature based on local standard deviation, and local information entropy attribute feature.

[0044] In specific implementation, multi-dimensional water area attribute feature values ​​obtained by any combination of color attribute features, gradient attribute features, statistical attribute features based on local standard deviation and local information entropy attribute features are used to perform water area detection, which can further improve the efficiency and accuracy of water area detection.

[0045] In one embodiment, any combination of features of the multi-dimensional water area attributes may also include: angular second moment attribute features.

[0046] In specific implementation, adding the angular second-order moment attribute feature to any of the above-mentioned combined features can further improve the efficiency and accuracy of water area detection.

[0047] In one embodiment, determining the multi-dimensional water area attribute characteristic values ​​of a plurality of to-be-detected points in a surface image may include determining the multi-dimensional water area attribute characteristic value of each to-be-detected point in the surface image according to the following method:

[0048] For each point to be detected, the calculation steps of each characteristic component value are performed:

[0049] Convert the surface image to the HSI color space, take the two HS components, and obtain the color attribute characteristic component value of each point to be detected;

[0050] The surface image is converted into a grayscale image, and the gradient attribute characteristic component value of each point to be detected is obtained according to the grayscale image;

[0051] Perform data statistical analysis on the preset neighborhood of each point to be detected, and obtain the statistical attribute characteristic component value based on the local standard deviation of each point to be detected according to the analysis results;

[0052] The entropy of each point to be detected is calculated for the preset neighborhood, and the characteristic component value of the local information entropy attribute of each point to be detected is obtained according to the calculation result;

[0053] The step of fusing the characteristic component values ​​of each point to be detected (the fusion step can be referred to the relevant introduction of formula (11)): the color attribute characteristic component value, gradient attribute characteristic component value, statistical attribute characteristic component value based on local standard deviation and local information entropy attribute characteristic component value of each point to be detected are fused to obtain the multi-dimensional water area attribute characteristic value of each point to be detected.

[0054] The following describes how to obtain the values ​​of each characteristic component and the process of obtaining the characteristic values ​​of the water area attributes based on the values ​​of each characteristic component.

[0055] In specific implementation, the multi-dimensional water area attribute feature value of each point in the image is calculated, and based on this, the attribute similarity judgment is performed with the seed point to determine whether it is an area similar to the seed point.

[0056] 1. First, we will introduce how to determine the color attribute feature component values.

[0057] In one embodiment, the surface image is converted to the HSI color space, and the two HS components are taken to obtain the color attribute characteristic component value of each point to be detected, including obtaining the color attribute characteristic component value of each point to be detected according to the following formula:

[0058]

[0059]

[0060] In the specific implementation, the image is converted to the HSI (hue, color saturation and brightness) color space, and the two HS components are taken as color attribute features. The HSI color space is based on the human visual system and uses hue, color saturation and brightness to describe color. The color characteristics of water areas are easier to distinguish under the HSI color system. Here, the two components of hue and saturation are selected as feature components.

[0061] In specific implementation, the image is converted to the HSI (hue, color saturation and brightness) color space, and the two components of HS (H: hue; S: color saturation) are used as color attribute features. The advantage is that in the commonly used RGB color system, RGB represents the amount of the three components of red, green and blue respectively, and the color and these three components are mixed together, making it difficult to distinguish colors. In the HIS color system, hue, saturation and brightness are independent of each other, and color can be distinguished from hue, and color depth can be distinguished from saturation. Therefore, the accuracy and efficiency of water area detection are further improved.

[0062] In the above formulas (2) and (3), H represents hue, S represents saturation, and R, G, and B represent the red, green, and blue color components of the original image in the RGB color system, respectively.

[0063] Normalize H and S respectively to obtain two color attribute features P describing the color of the water area. H and P S , where P H Represents the color feature component, P S Represents the saturation feature component.

[0064] 2. Next, we will introduce how to determine the characteristic component values ​​of gradient attributes.

[0065] In one embodiment, the surface image is converted into a grayscale image, and the gradient attribute characteristic component value of each point to be detected is obtained according to the grayscale image, including determining the gradient attribute characteristic component value according to the following formula:

[0066]

[0067] In the specific implementation, the surface image is converted into a grayscale image, the gradient feature K of each point is calculated, and the result is normalized to obtain the gradient attribute feature P of the current point (which can be the point to be detected or the seed point) K .

[0068] In the above formula (4), K x and K y They are the gradient information of a point (which can be a seed point or any point to be detected) in the x and y directions respectively.

[0069] 3. Next, we introduce how to determine the characteristic component values ​​of statistical attributes based on local standard deviation.

[0070] In one embodiment, a data statistical analysis is performed on a preset neighborhood where each to-be-detected point is located, and according to the analysis result, a statistical attribute characteristic component value based on a local standard deviation of each to-be-detected point is obtained, including: obtaining a statistical attribute characteristic component value based on a local standard deviation according to the following formula:

[0071]

[0072]

[0073] In the specific implementation, by performing data statistical analysis on the small neighborhood where the target point (the target point can be the point to be detected, and of course when calculating the attribute characteristics of the seed point, the target point can also be the seed point) is located, the attribute characteristics of the point based on statistical analysis are obtained. Since the overall representation of the water area is relatively consistent, the present invention uses the local standard deviation as a discriminant feature P σ .

[0074] In specific implementation, the above formula (5) and formula (6) are used for statistical analysis: first calculate the sum of the grayscale values ​​of all points in the small area (small neighborhood, preset neighborhood) around the target point (seed point or point to be detected), and then divide it by n (the number of points) to obtain the mean of the small area where the target point is located. Then, use formula (6) to calculate the variance of the small area where the target point is located, that is, first calculate the square of the difference between each point in the neighborhood and the mean calculated by formula (5), then sum it, and then divide it by n and take the square root.

[0075] In the above formulas (5) and (6), n represents the size of the small neighborhood (a preset small surrounding area), f i represents the gray value of the i-th point in the small neighborhood, μ is the mean of this small neighborhood (a preset small surrounding area, which can be set in advance), and σ is the variance of this small neighborhood.

[0076] 4. Next, we will introduce how to determine the local information entropy attribute feature component values.

[0077] In one embodiment, entropy calculation is performed on a preset neighborhood where each to-be-detected point is located, and obtaining a local information entropy attribute characteristic component value of each to-be-detected point according to the calculation result may include obtaining a local information entropy attribute characteristic component value according to the following formula:

[0078]

[0079] In the specific implementation, the small neighborhood where the point (target point: seed point or point to be detected) is calculated. Before the entropy calculation, the histogram statistics of the entire image need to be performed to obtain the probability of each value appearing in each image. Then the value of each point in the small neighborhood around the target point is calculated (using p(f i )log(p(f i)), and then calculate the results for each, and then calculate the average value), the information entropy attribute characteristics of the point can be obtained. Since the water area is flat, the characteristics are simple and similar, showing similar information characteristics. The present invention calculates the information entropy e of the small neighborhood where the point (target point: seed point or point to be detected) is located as its attribute feature, as shown in formula (7), and performs normalized mapping processing on the result to obtain the local information entropy attribute feature P e .

[0080] In the above formula (7), n represents the size of the small neighborhood, f i represents the gray value of the i-th point in the small neighborhood, p(f i ) represents the probability corresponding to this gray value.

[0081] In specific implementation, the above-mentioned implementation method of determining the values ​​of each characteristic component further improves the efficiency and accuracy of water area detection.

[0082] 5. Next, we will introduce how to determine the characteristic component values ​​of the angular second moment attributes.

[0083] In one embodiment, the step of calculating each characteristic component value for each point to be detected may further include:

[0084] The image moment feature analysis is performed on the preset neighborhood of each point to be detected, and the angular second-order moment attribute feature component value of each point to be detected is obtained according to the analysis result.

[0085] In one embodiment, image moment feature analysis is performed on a preset neighborhood where each point to be detected is located, and according to the analysis result, the angular second-order moment attribute feature component value of each point to be detected is obtained, including obtaining the angular second-order moment attribute feature component value of each point to be detected according to the following formula:

[0086]

[0087] In specific implementation, since the angular second-order moment is a measure of the uniformity of image distribution and texture coarseness, and the water area appears relatively flat in the image, this feature can be used as a basis for discrimination. By calculating the small neighborhood where the point is located, it is necessary to perform histogram statistics on the entire image before calculating to obtain the probability of each value in each image (grayscale texture is used here to increase the speed of the algorithm). Then calculate the value of each point in the small neighborhood around the target point (using p in formula (8) 2 (f i )), then calculate the result, and then calculate the average value to obtain the moment attribute feature M of the point, as shown in formula (8), and perform normalized mapping on the result to obtain the angular second-order moment attribute feature P of the water area M .

[0088] In the above formula (8), n represents the size of the small neighborhood, f i represents the gray value of the i-th point in the small neighborhood, p(f i ) represents the probability corresponding to this gray value.

[0089] 6. Next, the steps of fusing the characteristic component values ​​to obtain the multi-dimensional water area attribute characteristic values ​​are introduced.

[0090] In one embodiment, the step of fusing the characteristic component values ​​of each point to be detected may include: fusing the color attribute characteristic component value, gradient attribute characteristic component value, statistical attribute characteristic component value based on local standard deviation, local information entropy attribute characteristic component value and angular second-order moment attribute characteristic component value of each point to be detected to obtain a multi-dimensional water area attribute characteristic value of each point to be detected.

[0091] In one embodiment, the color attribute feature component value, the gradient attribute feature component value, the statistical attribute feature component value based on the local standard deviation, the local information entropy attribute feature component value and the angular second-order moment attribute feature component value of each point to be detected are fused to obtain the multidimensional water area attribute feature value of each point to be detected, including obtaining the multidimensional water area attribute feature value of each point to be detected according to the following formula:

[0092] p=[P H P s P K P σ P e P M ]; (9)

[0093] r=[r H r s r K r σ r e r M ]; (10)

[0094] P=p×r=[P H × H P s × s P K × K P σ × σ P e × e P M × M ];(11)

[0095] Wherein, in the above formulas (10) and (11), r is the characteristic component weight coefficient.

[0096] In specific implementation, the above scheme of determining the angular second-order moment attribute characteristic component value and further considering the angular second-order moment attribute characteristic component value to obtain the multi-dimensional water area attribute characteristic value further improves the efficiency and accuracy of water area detection.

[0097] 4. Next, the above steps 103 and 104 are introduced.

[0098] In one embodiment, based on the multidimensional water area attribute characteristic value of at least one water area feature point and the multidimensional water area attribute characteristic values ​​of multiple points to be detected in the surface image, determining all points to be detected whose similarity to the multidimensional water area attribute characteristic value of at least one water area feature point is greater than or equal to a preset threshold may include:

[0099] Add a water feature point as an original seed point to the water set;

[0100] Taking the original seed point as the starting point, the following water area detection steps are performed: calculating the similarity between the multidimensional water area attribute characteristic value of the original seed point and the multidimensional water area attribute characteristic value of a current point to be detected in the area corresponding to the original seed point; when it is determined that the similarity is greater than or equal to a preset threshold, adding the water area of ​​the current point to be detected as the water area corresponding to the original seed point to the water area set, and updating the multidimensional water area attribute characteristic value corresponding to the water area set;

[0101] Take the current point to be detected as a new seed point, repeat the steps of water area detection, and look for a point to be detected whose similarity with the multi-dimensional water area attribute characteristic value of the new seed point is greater than or equal to the preset threshold, or whose similarity with the multi-dimensional water area attribute characteristic value corresponding to the updated water area set is greater than or equal to the preset threshold, and so on until all the points to be detected using the original seed point are obtained.

[0102] The above step 103 is described in detail below.

[0103] 1. First, the similarity determination method involved in the embodiment of the present invention is introduced.

[0104] In one embodiment, the similarity between the point to be detected and the seed point or the water feature can be calculated by the following formula:

[0105]

[0106] Where P is the feature vector data to be detected (the multi-dimensional water area attribute feature value of the point to be detected can be obtained by the above formula (11)), P sis the standard feature vector data (it can be the multidimensional water area attribute feature value of the seed point, or it can be obtained by formula (11). The standard feature vector data can also be the multidimensional water area attribute feature value corresponding to the water area set, that is, the multidimensional water area attribute feature value of the entire water area, which can be obtained by the following formula (14). See the following example for details), EP is the mean value of P, EP s P s The mean of P, D(P) is the variance of P, D(P s ) is P s The variance of , E() indicates the mean calculation. s(P,P s ) is the similarity (similarity) between two feature vectors.

[0107] 2. Secondly, the water area search (detection) method based on similarity discrimination method is introduced.

[0108] In specific implementation, a series of original seed points are obtained through user interaction, and the water areas are searched (detected) respectively with these seed points as starting points. Of course, one original seed point is also acceptable.

[0109] ① First, divide the image into the union of three regions: water area r w 、Non-water area n and undetected area r u .

[0110] I={r w ∪r u ∪r n}; (13)

[0111] In the initial state, r u is the entire image area, and the water area r w and r n Non-water areas are empty sets.

[0112] Take an original seed point p s (x,y) adds it to the queue to be checked and uses it as the starting point for searching. And adds it to the water area set r w among.

[0113] The above step ① explains the above: adding a water feature point as an original seed point to the water area set.

[0114] ②Then take the seed point p s (x, y) is the starting point, and the points in its neighborhood that have not been searched are used as seed points to be grown (points to be detected), and the characteristics of the points to be grown and the seed points or the entire water area (that is, the multidimensional water area attribute characteristic value corresponding to the water area set, that is, the characteristic feature vector calculated for the entire point that has been determined to be a water area using formula (11)) p wThe similarity coefficient (similarity) between them is calculated as shown in the above formula (12), where the water feature p w It can be obtained by the following formula (14):

[0115]

[0116] where p w Represents the characteristics of the entire water area, N RW (Number-Region-water) is the number of data points in the water area (all points that have been determined to be water areas), P i Represents the characteristic vector of the i-th point in the water area (the multi-dimensional water area attribute characteristic value, which can be obtained by the above formula (11)).

[0117] If the similarity is extremely low (less than the preset threshold), directly move the point to be detected into the non-water set r n If the similarity is extremely high (greater than or equal to the preset threshold), the point to be detected is moved into the water area set r w , then update p w Value (As the new feature points are added to the water area, the characteristics of the entire water area will change, so it is necessary to calculate and update p w . Because the newly added point needs to be compared with the overall characteristics of the water area), the point is added to the queue of seed points to be detected, that is, the point becomes a new seed point.

[0118] The above step ② explains: taking the original seed point as the starting point, the following water area detection steps are performed: calculating the similarity between the multi-dimensional water area attribute characteristic value of the original seed point and the multi-dimensional water area attribute characteristic value of a current point to be detected in the area corresponding to the original seed point; when it is determined that the similarity is greater than or equal to the preset threshold, the water area of ​​the current point to be detected is added to the water area set as the water area corresponding to the original seed point, and the multi-dimensional water area attribute characteristic value corresponding to the water area set is updated.

[0119] ③If p s If all the points in the (x, y) neighborhood have been detected, the seed point is deleted from the queue to be detected, and the next point to be detected is found from the set to be detected and step ② is repeated.

[0120] If all points in the queue to be inspected (the queue to be inspected stores the original seeds and the temporarily added seed points to be identified) are detected, then points are taken from the undetected area (all undetected areas other than the seed points) as the seed points to be grown for detection, and step ② is repeated. When all points are detected, the water area detected using the seed point can be obtained.

[0121] This step ③ describes: taking the current point to be detected as a new seed point, repeating the water area detection steps, searching for a point to be detected whose similarity with the multi-dimensional water area attribute characteristic value of the new seed point is greater than or equal to the preset threshold, or whose similarity with the multi-dimensional water area attribute characteristic value corresponding to the updated water area set is greater than or equal to the preset threshold, until all the points to be detected using the original seed point are obtained.

[0122] The above steps ①-③ are the process of detecting water areas using a seed point, that is, the process of detecting all water areas in a large area corresponding to a seed point.

[0123] ④When all points are detected, take the next original seed point from the original seed point set and start the calculation from step ①.

[0124] This step ④ introduces that when there are multiple seed points, the process of using each seed point to detect the water area is as shown in the above steps ①-③.

[0125] ⑤ When all the original seed points are detected, all the detected water areas are merged to obtain the final detection result.

[0126]

[0127] in: represents the water area corresponding to the seed point s (including all the seed points found, including: the water area of ​​the original seed point (water area feature point), and the water area of ​​all points to be detected whose multi-dimensional water area attribute characteristic value similarity with at least one water area feature point is greater than or equal to the preset threshold), I w Indicates all water areas found.

[0128] This step ⑤ is the above-mentioned step 104.

[0129] 5. Next, the step of extracting the boundary of the water area after the water area is detected in the above steps 103 and 104 is introduced.

[0130] In one embodiment, the above-mentioned water area detection method based on multi-attribute consistency may further include: after obtaining the water area detection result of the surface image, performing an operation process of extracting the water area boundary.

[0131] In specific implementation, after finding the water area, it is necessary to extract the water area boundary and use the structure operator Q to extract the water area image I w Dilate and then subtract I w The water boundary can be obtained.

[0132]

[0133] in:

[0134] Finally, the technical solution provided by the embodiment of the present invention is introduced as a whole.

[0135] 1) Data preparation:

[0136] The image to be inspected (surface image) is opened, and the user marks the seed point (water feature point) in the image with the mouse, and gives the threshold parameter (such as the preset threshold) for judging the attribute similarity.

[0137] 2) Water area detection based on multi-attribute consistency.

[0138] ① Obtain the multi-dimensional water attribute characteristics of each point in the image: calculate the color characteristics, gradient characteristics, statistical characteristics, entropy characteristics, and moment characteristics of each point in the image.

[0139] ② Use seed points to search for water areas according to attribute similarity: Start from the original seed point and search for each point in its neighborhood. Find points that have not been searched and have similar attributes and add them to the water area. Set them as seed points to be searched for recursive search. After the search is completed, the water area consistent with the current original seed point is obtained. Then repeat this step to find the water area corresponding to all the original seed points, and merge the results to get the final search result.

[0140] ③ Extract water area boundaries: After finding the water area, use morphological methods to extract the found water area boundaries.

[0141] The inventor has conducted a lot of experiments and demonstrated the effect of water area detection based on multi-attribute consistency in the embodiment of the present invention. Figure 3b Effects with standards Figure 3a Comparison, and schematic diagram of the effect of water area detection based on multi-attribute consistency in another embodiment of the present invention Figure 4b Standard renderings Figure 4a By comparison, it is found that the water area detection based on multi-attribute consistency in the embodiment of the present invention has a good effect, high efficiency and accuracy.

[0142] In summary, the present invention uses a water area search method for the multi-dimensional attribute similarity of water areas to search for water areas similar to the seeds based on the water area seeds calibrated by the user, and automatically extracts the boundaries of the water areas. Through this method, the user can automatically search for water areas and extract corresponding boundaries by only calibrating a few mouse points. This greatly improves the search speed and calibration accuracy of water areas.

[0143] Based on the same inventive concept, an embodiment of the present invention also provides a water area detection device based on multi-attribute consistency, as described in the following embodiments. Since the principle of solving the problem by the water area detection device based on multi-attribute consistency is similar to that of the water area detection method based on multi-attribute consistency, the implementation of the water area detection device based on multi-attribute consistency can refer to the implementation of the water area detection method based on multi-attribute consistency, and the repeated parts will not be repeated. As used below, the term "unit" or "module" can be a combination of software and / or hardware that implements predetermined functions. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceived.

[0144] Figure 5 is a schematic diagram of the structure of a water area detection device based on multi-attribute consistency in an embodiment of the present invention. Figure 5 As shown, the device comprises:

[0145] The first determination unit 01 is used to determine the multi-dimensional water area attribute characteristic value of at least one water area characteristic point in the surface image;

[0146] The second determination unit 02 is used to determine the multi-dimensional water area attribute characteristic values ​​of multiple points to be detected in the surface image;

[0147] The detection unit 03 is used to determine all the points to be detected whose similarity with the multidimensional water area attribute characteristic value of at least one water area feature point is greater than or equal to a preset threshold value based on the multidimensional water area attribute characteristic value of at least one water area feature point and the multidimensional water area attribute characteristic values ​​of multiple points to be detected in the surface image;

[0148] Processing unit 04 is used to obtain the water area detection result of the surface image based on the water area of ​​at least one water area feature point and the water areas of all points to be detected whose similarity with the multi-dimensional water area attribute characteristic value of at least one water area feature point is greater than or equal to a preset threshold.

[0149] In one embodiment, the multi-dimensional water area attribute feature includes any combination of the following features: color attribute feature, gradient attribute feature, statistical attribute feature based on local standard deviation, and local information entropy attribute feature.

[0150] In one embodiment, the multi-dimensional water area attribute characteristics further include: angular second-order moment attribute characteristics.

[0151] In one embodiment, the second determination unit is specifically used to determine the multi-dimensional water area attribute characteristic value of each to-be-detected point in the surface image: the second determination unit includes:

[0152] The characteristic component calculation unit is used to perform the calculation steps of each characteristic component value for each point to be detected:

[0153] Convert the surface image to the HSI color space, take the two HS components, and obtain the color attribute characteristic component value of each point to be detected;

[0154] The surface image is converted into a grayscale image, and the gradient attribute characteristic component value of each point to be detected is obtained according to the grayscale image;

[0155] Perform data statistical analysis on the preset neighborhood of each point to be detected, and obtain the statistical attribute characteristic component value based on the local standard deviation of each point to be detected according to the analysis results;

[0156] The entropy of each point to be detected is calculated for the preset neighborhood, and the characteristic component value of the local information entropy attribute of each point to be detected is obtained according to the calculation result;

[0157] A fusion processing unit is used to fuse the characteristic component values ​​of each point to be detected: the color attribute characteristic component value, the gradient attribute characteristic component value, the statistical attribute characteristic component value based on the local standard deviation and the local information entropy attribute characteristic component value of each point to be detected are fused to obtain the multi-dimensional water area attribute characteristic value of each point to be detected.

[0158] In one embodiment, the characteristic component calculation unit is further used for:

[0159] Perform image moment feature analysis on the preset neighborhood of each point to be detected, and obtain the angular second-order moment attribute feature component value of each point to be detected according to the analysis result;

[0160] The fusion processing unit is specifically used to fuse the color attribute feature component value, gradient attribute feature component value, statistical attribute feature component value based on local standard deviation, local information entropy attribute feature component value and angular second-order moment attribute feature component value of each point to be detected to obtain a multi-dimensional water area attribute feature value of each point to be detected.

[0161] In one embodiment, the second determination unit is specifically used to: determine all points to be detected whose similarity with a multi-dimensional water area attribute characteristic value of a water area characteristic point is greater than or equal to a preset threshold according to the following method:

[0162] Add a water feature point as an original seed point to the water set;

[0163] The following water area detection steps are performed with the original seed point as the starting point: based on the multidimensional water area attribute characteristic value of the original seed point and the multidimensional water area attribute characteristic value of a current point to be detected in the area corresponding to the original seed point, the similarity between the multidimensional water area attribute characteristic value of the original seed point and the multidimensional water area attribute characteristic value of the current point to be detected is calculated; when it is determined that the similarity is greater than or equal to a preset threshold, the current point to be detected is added to the water area set as the water area corresponding to the original seed point, and the multidimensional water area attribute characteristic value in the water area set is updated;

[0164] Take the current point to be detected as a new seed point, repeat the water area detection steps, and search for a point to be detected whose similarity with the multi-dimensional water area attribute characteristic value of the new seed point is greater than or equal to the preset threshold, or whose similarity with the multi-dimensional water area attribute characteristic value in the updated water area set is greater than or equal to the preset threshold, until all the points to be detected using the original seed point are obtained.

[0165] An embodiment of the present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned water area detection method based on multi-attribute consistency when executing the computer program.

[0166] An embodiment of the present invention further provides a computer-readable storage medium, which stores a computer program for executing the above-mentioned water area detection method based on multi-attribute consistency.

[0167] The beneficial technical effect of the technical solution provided by the embodiment of the present invention is: it realizes rapid and accurate water area detection, improves the efficiency and accuracy of water area detection, and further improves the efficiency and accuracy of water area obstacle calibration.

[0168] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0169] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0170] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0171] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0172] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the embodiments of the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A water area detection method based on multi-attribute consistency, characterized in that: include: Determine a multidimensional water area attribute characteristic value of at least one water area characteristic point in a surface image; Determine the multi-dimensional water area attribute characteristic values ​​of multiple points to be detected in the surface image; According to the multidimensional water area attribute characteristic value of at least one water area feature point and the multidimensional water area attribute characteristic values ​​of multiple points to be detected in the surface image, all points to be detected whose similarity with the multidimensional water area attribute characteristic value of at least one water area feature point is greater than or equal to a preset threshold are determined; Obtaining a water area detection result of the surface image according to the water area of ​​at least one water area feature point and the water areas of all points to be detected whose similarity to the multi-dimensional water area attribute characteristic value of the at least one water area feature point is greater than or equal to a preset threshold; Among them, determining the multidimensional water area attribute characteristic values ​​of multiple points to be detected in the surface image includes determining the multidimensional water area attribute characteristic values ​​of each point to be detected in the surface image according to the following method: performing the calculation steps of each characteristic component value for each point to be detected: converting the surface image to the HSI color space, taking two HS components, and obtaining the color attribute characteristic component value of each point to be detected; converting the surface image to a grayscale image, and obtaining the gradient attribute characteristic component value of each point to be detected according to the grayscale image; performing data statistical analysis on a preset neighborhood where each point to be detected is located, and obtaining the statistical attribute characteristic component value based on the local standard deviation of each point to be detected according to the analysis result; performing entropy calculation on the preset neighborhood where each point to be detected is located, and obtaining the local information entropy attribute characteristic component value of each point to be detected according to the calculation result; and performing the step of fusing the various characteristic component values ​​of each point to be detected: fusing the color attribute characteristic component value, the gradient attribute characteristic component value, the statistical attribute characteristic component value based on the local standard deviation, and the local information entropy attribute characteristic component value of each point to be detected to obtain the multidimensional water area attribute characteristic value of each point to be detected.

2. The water area detection method based on multi-attribute consistency according to claim 1, characterized in that: The multi-dimensional water area attribute features include any combination of the following features: color attribute features, gradient attribute features, statistical attribute features based on local standard deviations, and local information entropy attribute features.

3. The water area detection method based on multi-attribute consistency according to claim 2, characterized in that: The multi-dimensional water area attribute characteristics also include: angular second-order moment attribute characteristics.

4. The water area detection method based on multi-attribute consistency according to claim 1, characterized in that: The steps of calculating each characteristic component value for each point to be detected also include: Perform image moment feature analysis on the preset neighborhood of each point to be detected, and obtain the angular second-order moment attribute feature component value of each point to be detected according to the analysis result; The step of fusing the characteristic component values ​​of each point to be detected includes: fusing the color attribute characteristic component value, gradient attribute characteristic component value, statistical attribute characteristic component value based on local standard deviation, local information entropy attribute characteristic component value and angular second-order moment attribute characteristic component value of each point to be detected to obtain the multi-dimensional water area attribute characteristic value of each point to be detected.

5. The water area detection method based on multi-attribute consistency according to claim 1, characterized in that: According to the multidimensional water area attribute characteristic value of at least one water area feature point and the multidimensional water area attribute characteristic values ​​of multiple points to be detected in the surface image, all points to be detected whose similarity with the multidimensional water area attribute characteristic value of at least one water area feature point is greater than or equal to a preset threshold are determined, including: determining all points to be detected whose similarity with the multidimensional water area attribute characteristic value of a water area feature point is greater than or equal to the preset threshold according to the following method: Add a water feature point as an original seed point to the water set; Taking the original seed point as the starting point, the following water area detection steps are performed: calculating the similarity between the multidimensional water area attribute characteristic value of the original seed point and the multidimensional water area attribute characteristic value of a current point to be detected in the area corresponding to the original seed point; when it is determined that the similarity is greater than or equal to a preset threshold, adding the water area of ​​the current point to be detected as the water area corresponding to the original seed point to the water area set, and updating the multidimensional water area attribute characteristic value corresponding to the water area set; Take the current point to be detected as a new seed point, repeat the water area detection steps, and search for a point to be detected whose similarity with the multi-dimensional water area attribute characteristic value of the new seed point is greater than or equal to the preset threshold, or whose similarity with the multi-dimensional water area attribute characteristic value corresponding to the updated water area set is greater than or equal to the preset threshold, until all the points to be detected using the original seed point are obtained.

6. A water area detection device based on multi-attribute consistency, characterized in that: include: A first determining unit is used to determine a multi-dimensional water area attribute characteristic value of at least one water area characteristic point in a surface image; The second determination unit is used to determine the multi-dimensional water area attribute characteristic values ​​of multiple points to be detected in the surface image; A detection unit, configured to determine, based on the multidimensional water area attribute characteristic value of at least one water area feature point and the multidimensional water area attribute characteristic values ​​of a plurality of points to be detected in the surface image, all points to be detected whose similarity to the multidimensional water area attribute characteristic value of at least one water area feature point is greater than or equal to a preset threshold; A processing unit, configured to obtain a water area detection result of the surface image according to the water area of ​​at least one water area feature point and the water areas of all points to be detected whose similarity to the multi-dimensional water area attribute characteristic value of the at least one water area feature point is greater than or equal to a preset threshold; The second determination unit is specifically used to determine the multi-dimensional water area attribute characteristic value of each point to be detected in the surface image: the second determination unit includes: a characteristic component calculation unit, which is used to perform the calculation steps of each characteristic component value for each point to be detected: converting the surface image to the HSI color space, taking two HS components, and obtaining the color attribute characteristic component value of each point to be detected; converting the surface image to a grayscale image, and obtaining the gradient attribute characteristic component value of each point to be detected according to the grayscale image; performing data statistical analysis on the preset neighborhood where each point to be detected is located, and obtaining the statistical attribute characteristic component value based on the local standard deviation of each point to be detected according to the analysis result; performing entropy calculation on the preset neighborhood where each point to be detected is located, and obtaining the local information entropy attribute characteristic component value of each point to be detected according to the calculation result; a fusion processing unit, which is used to perform the steps of fusing the various characteristic component values ​​of each point to be detected: fusing the color attribute characteristic component value, the gradient attribute characteristic component value, the statistical attribute characteristic component value based on the local standard deviation and the local information entropy attribute characteristic component value of each point to be detected to obtain the multi-dimensional water area attribute characteristic value of each point to be detected.

7. The water area detection device based on multi-attribute consistency according to claim 6, characterized in that: The multi-dimensional water area attribute features include any combination of the following features: color attribute features, gradient attribute features, statistical attribute features based on local standard deviations, and local information entropy attribute features.

8. The water area detection device based on multi-attribute consistency according to claim 7, characterized in that: The multi-dimensional water area attribute characteristics also include: angular second-order moment attribute characteristics.

9. The water area detection device based on multi-attribute consistency according to claim 6, characterized in that: The characteristic component calculation unit is also used for: Perform image moment feature analysis on the preset neighborhood of each point to be detected, and obtain the angular second-order moment attribute feature component value of each point to be detected according to the analysis result; The fusion processing unit is specifically used to fuse the color attribute feature component value, gradient attribute feature component value, statistical attribute feature component value based on local standard deviation, local information entropy attribute feature component value and angular second-order moment attribute feature component value of each point to be detected to obtain a multi-dimensional water area attribute feature value of each point to be detected.

10. The water area detection device based on multi-attribute consistency according to claim 6, characterized in that: The second determination unit is specifically used to determine all points to be detected whose similarity with the multi-dimensional water area attribute characteristic value of a water area characteristic point is greater than or equal to a preset threshold according to the following method: Add a water feature point as an original seed point to the water set; The following water area detection steps are performed with the original seed point as the starting point: based on the multidimensional water area attribute characteristic value of the original seed point and the multidimensional water area attribute characteristic value of a current point to be detected in the area corresponding to the original seed point, the similarity between the multidimensional water area attribute characteristic value of the original seed point and the multidimensional water area attribute characteristic value of the current point to be detected is calculated; When it is determined that the similarity is greater than or equal to a preset threshold, the water area corresponding to the current point to be detected is added to the water area set as the original seed point, and the multi-dimensional water area attribute characteristic value in the water area set is updated; Take the current point to be detected as a new seed point, repeat the water area detection steps, and search for a point to be detected whose similarity with the multi-dimensional water area attribute characteristic value of the new seed point is greater than or equal to the preset threshold, or whose similarity with the multi-dimensional water area attribute characteristic value in the updated water area set is greater than or equal to the preset threshold, until all the points to be detected using the original seed point are obtained.

11. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 5 is implemented.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program for executing the method according to any one of claims 1 to 5.

Images