A method for optimizing segmentation of geological survey remote sensing images

By obtaining the angle of the connection domain and edge direction in the geological survey remote sensing image, determining the ridge line area, and optimizing the distance calculation in the SLIC algorithm, the problem of undersegment in the existing technology is solved and the segmentation accuracy of the geological survey remote sensing image is improved .

CN118941582BActive Publication Date: 2024-12-06DALIAN QIANXI NETWORK TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411412522.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-10-11
Publication Date
2024-12-06
Estimated Expiration
2044-10-11

AI Technical Summary

Technical Problem

The existing remote sensing image segmentation method of geological survey ignores the problem of blurred boundaries, resulting in undersegment, affecting the accuracy of subsequent analysis and feature extraction.

Method used

采用一种地质勘察遥感影像优化分割方法,通过获取灰度图像的连通域、边缘检测和单元格方向角度计算,确定山脊线区域,并结合SLIC算法优化聚类中心与像素点之间的距离,避免欠分割。

Benefits of technology

The accuracy of the remote sensing image segmentation results of geological surveys is improved, and the blurring of ridge areas caused by shooting angles and light is avoided, ensuring accurate identification and segmentation of land objects.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118941582B_ABST
    Figure CN118941582B_ABST
Patent Text Reader

Abstract

The present application relates to the field of image processing technology, and specifically to a method for optimizing the segmentation of geological survey remote sensing images, the method comprising: obtaining geological survey remote sensing images and converting them into grayscale images; obtaining each connected domain and each edge curve in the grayscale image, dividing the minimum circumscribed rectangle of each connected domain into each cell, and determining the direction angle of each cell based on the distribution of the gradient direction of all pixels in each cell; determining the direction angle of each edge curve in each connected domain according to the distribution of the direction angle of all cells containing each edge curve in each connected domain; obtaining each ridgeline area in each connected domain; determining the ridgeline of each ridgeline area based on the distribution characteristics of the grayscale values ​​of the pixels in each ridgeline area; and optimizing the distance between the cluster center and the pixel point in the superpixel segmentation process. The present application aims to improve the accuracy of the segmentation results of geological survey remote sensing images.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of image processing, and in particular to a method for optimizing segmentation of geological survey remote sensing images. Background Art

[0002] Geological survey is an important means to understand and evaluate the geological environment, resource distribution and geological disasters. Remote sensing technology plays an increasingly important role in geological survey with its wide coverage, large amount of information and low cost. Remote sensing images contain rich information about land objects. How to effectively analyze and process these images and extract valuable geological information is the key to geological survey. Among them, image segmentation is an important part of the preliminary processing, which directly affects the accuracy of subsequent geological feature extraction and identification.

[0003] However, there are still some areas that need to be optimized in the current geological survey remote sensing image segmentation methods. Remote sensing images are often affected by factors such as complex backgrounds, noise and shadows, which blur the boundaries between objects. Existing segmentation methods ignore the problem of blurred boundaries, which may cause objects on both sides of the boundary to be segmented into the same superpixel block, resulting in under-segmentation, that is, some originally discontinuous object areas are segmented together, affecting subsequent analysis and regional feature extraction results. For this situation of blurred boundaries between objects, how to accurately identify them to avoid under-segmentation problems is an important consideration for improving the accuracy of geological survey remote sensing image segmentation. Summary of the invention

[0004] In view of the above, it is necessary to provide a geological survey remote sensing image optimization segmentation method, which improves the accuracy of the geological survey remote sensing image segmentation results compared with the traditional geological survey remote sensing image segmentation method.

[0005] The present invention discloses a method for optimizing segmentation of geological survey remote sensing images using the following technical solutions:

[0006] An embodiment of the present application provides a method for optimizing segmentation of geological survey remote sensing images, the method comprising the following steps:

[0007] Acquire geological survey remote sensing images and convert them into grayscale images;

[0008] A preset number of pixels are randomly selected from the grayscale image as seed points, each connected domain of the grayscale image is obtained by a region growing algorithm, each edge curve in the grayscale image is obtained by an edge detection algorithm, the minimum circumscribed rectangle of each connected domain is equally divided into cells, and the direction angle of each cell is determined based on the distribution of the gradient direction of all pixels in each cell;

[0009] Determine the direction angle of each edge curve in each connected domain according to the distribution of the direction angles of all cells containing each edge curve in each connected domain;

[0010] Based on the direction angles and distribution ranges of all edge curves in each connected domain, each ridgeline area in each connected domain is obtained;

[0011] Determine the ridgeline of each ridgeline region based on the distribution characteristics of the grayscale values ​​of the pixels in each ridgeline region;

[0012] In the process of superpixel segmentation of geological survey remote sensing images, based on the positional relationship between any cluster center, any pixel point and the ridge line, the determination factor of the distance between any cluster center and any pixel point is determined, and combined with the maximum value of the distance between any cluster center and all pixel points, the distance between any cluster center and any pixel point in the superpixel segmentation process is optimized.

[0013] In one embodiment, the process of determining the direction angle of the cell is as follows:

[0014] The ratio of the number of pixels in various gradient directions in each cell to the total number of pixels is calculated, the product of various gradient directions and the corresponding ratio is calculated, and the sum of the products of all gradient directions in each cell is used as the direction angle of each cell.

[0015] In one embodiment, the process of determining the direction angle of the edge curve is:

[0016] For each edge curve in each connected domain, obtain the number of pixel points on the edge curve contained in each cell in the connected domain, calculate the ratio of the number to the number of pixel points on the edge curve, and record it as the quantity ratio;

[0017] Based on the ratio of the direction angles of all cells containing edge curves in the connected domain to the quantity, the direction angles of the edge curves in the connected domain are determined.

[0018] In one embodiment, the expression of the direction angle of the edge curve is:

[0019] ; In the formula, Represents edge curve Direction angle; Indicates that edge curves are included The number of cells; Represents edge curve The number of pixels on the Indicates that edge curves are included No. The edge curves contained in the cells The number of pixels on ; Indicates that edge curves are included No. The direction angle of each cell.

[0020] In one embodiment, the process of determining the ridgeline area is as follows:

[0021] Combining the positions and direction angles of all edge curves in each connected domain, an edge direction distribution map is constructed;

[0022] Based on the edge direction distribution map, the mean of the number of solid points in all cells containing solid points and the mean of the number of hollow points in all cells containing hollow points in any connected domain are recorded as the solid point mean and the hollow point mean, respectively;

[0023] Connecting all cells in any connected domain whose number of solid points is greater than or equal to the mean value of solid points into a solid point region; connecting all cells in any connected domain whose number of hollow points is greater than or equal to the mean value of hollow points into a hollow point region;

[0024] The ridgeline area in any connected domain is obtained by combining the solid point mean, the hollow point mean, the solid point area and the hollow point area.

[0025] In one embodiment, the process of constructing the edge direction distribution map is as follows:

[0026] The two end points of each edge curve are taken as two representative points of each edge curve;

[0027] The horizontal and vertical coordinates in the edge direction distribution diagram are the horizontal and vertical coordinate values ​​of the representative point respectively;

[0028] When the direction angle of any edge curve is If the direction angle of any edge curve is between If the two representative points are between , then their two representative points are marked as hollow points.

[0029] In one of the embodiments, each connected area formed by connecting all cells whose number of solid points is less than the average of the solid points and whose number of hollow points is less than the average of the hollow points is obtained, and if the adjacent left side of the connected area is a solid point area and the adjacent right side is a hollow point area, the connected area is taken as a ridge line area.

[0030] In one embodiment, the process of determining the ridge line is: obtaining the pixel with the largest grayscale value in each row of pixels in any ridge line area, and using the fitting curve of all the obtained pixels as the ridge line of any ridge line area.

[0031] In one embodiment, the determination factor is determined by:

[0032] The pixels on the ridge line are recorded as ridge points;

[0033] When there is a ridge point on the line between any cluster center and any pixel point, the value of the determination factor is greater than or equal to 0; otherwise, the value of the determination factor is less than 0.

[0034] In one embodiment, the distance between any cluster center and any pixel point in the process of optimizing superpixel segmentation further includes:

[0035] ; In the formula, Represents the optimized cluster center With pixels The distance between Represents the cluster center The maximum value of the Euclidean distance between all pixels in the geological survey remote sensing image; Represents the cluster center With pixels The Euclidean distance between Represents the cluster center With pixels The determining factor for measuring the distance between them.

[0036] This application has at least the following beneficial effects:

[0037] This application calculates the direction angle of each cell based on the gradient direction of all pixels in each cell of the connected domain, and obtains the direction angle of each edge curve by weighted average method, thereby improving the accuracy of calculating the direction angle of the edge curve, thereby improving the accuracy of subsequent determination of the ridge line;

[0038] Furthermore, the ridgeline area is determined by combining the direction distribution of all edge curves in the connected domain, which fully considers the direction of all edge curves in the connected domain and improves the accuracy of obtaining the ridgeline area; the ridgeline of the ridgeline area is obtained according to the grayscale value of the pixel point in the ridgeline area, which facilitates the subsequent determination of the distance between the cluster center and the pixel point according to the position of the ridgeline;

[0039] Furthermore, the distance between the cluster center and the pixel in the SLIC algorithm is optimized according to whether the pixel point and the cluster center cross the ridge, avoiding the blurring of the ridge area due to the influence of shooting angle, lighting, etc., thereby dividing the areas on both sides of the ridge into the same superpixel block, thereby improving the accuracy of the segmentation results of geological survey remote sensing images. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0041] Figure 1 A flowchart of the steps of a geological survey remote sensing image optimization segmentation method provided in this application;

[0042] Figure 2 This is a schematic diagram of the ridgeline area;

[0043] Figure 3 Get a flow diagram for the ridgeline area;

[0044] Figure 4 Schematic diagram of the pixel with the largest gray value. DETAILED DESCRIPTION

[0045] In the description of the embodiments of the present application, words such as "exemplary", "or", "for example" and the like are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary", "or", "for example" and the like is intended to present related concepts in a concrete manner.

[0046] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art in the present application. The terms used in the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. It should be understood that, unless otherwise specified, " / " means or.

[0047] It should also be noted that the terms "first" and "second" in the present application are used to distinguish similar objects rather than to describe a specific order or sequence.

[0048] The following is a detailed description of a method for optimizing segmentation of remote sensing images for geological survey provided by the present application in conjunction with the accompanying drawings.

[0049] An embodiment of the present application provides a method for optimizing segmentation of geological survey remote sensing images. Specifically, the following method for optimizing segmentation of geological survey remote sensing images is provided. Figure 1 , the method comprises the following steps:

[0050] This application takes the image segmentation of the area near the ridge in the geological survey remote sensing image as an example. Due to the influence of the shooting angle, the color distinction of the pixels on both sides of a ridge may not be obvious. When performing image segmentation, the pixels on both sides of the ridge are divided into the same super-pixel block, which in turn affects the subsequent image analysis. By analyzing the features on both sides of the ridge, the ridge position is accurately determined. When performing subsequent image segmentation, the distance is optimized to avoid the problem of under-segmentation caused by dividing the pixels on both sides of the ridge into the same super-pixel block.

[0051] Step 1: Obtain geological survey remote sensing images and convert them into grayscale images.

[0052] Acquire geological survey remote sensing images of the area near the ridge. There may be noise in the remote sensing images, which will affect the subsequent edge detection and image segmentation effects. Therefore, it is necessary to pre-process the remote sensing images to reduce the interference of noise. The geological survey remote sensing images are denoised using a filtering method and converted into grayscale images using a color space conversion technology. The color space conversion technology is a well-known technology and will not be described in detail in this application.

[0053] In this embodiment, the median filtering method is used to denoise the geological survey remote sensing image. As other implementation methods, on the basis of being able to denoise the geological survey remote sensing image, the implementer can use other existing technologies to denoise the geological survey remote sensing image, such as Gaussian filtering method, mean filtering, etc., and this application does not make any special restrictions.

[0054] Step 2: randomly select a preset number of pixels from the grayscale image as seed points, obtain the connected domains of the grayscale image through the region growing algorithm, use the edge detection algorithm to obtain the edge curves in the grayscale image, divide the minimum circumscribed rectangle of each connected domain into cells, and determine the direction angle of each cell based on the distribution of the gradient direction of all pixels in each cell.

[0055] For ridge lines with obvious boundaries with surrounding scenery in geological survey remote sensing images, the SLIC algorithm can divide the pixels on both sides of the ridge lines into different superpixel blocks. However, for ridge lines with unclear boundaries with surrounding scenery, the SLIC algorithm may divide the pixels on both sides of the ridge lines into the same superpixel block, resulting in inaccurate segmentation results. Therefore, it is necessary to detect ridge lines with unclear boundaries with surrounding scenery in geological survey remote sensing images.

[0056] Since it is necessary to detect the ridge lines whose boundaries with the surrounding scenery are not obvious in the geological survey remote sensing images, the connected domains in the grayscale image are extracted and each connected domain is analyzed to find the ridge lines in the connected domain.

[0057] A preset number of pixels are randomly selected from the grayscale image as seed points of the region growing algorithm. For any seed point, the difference in grayscale value between the seed point and the adjacent pixel point is calculated as the similarity value between the seed point and the pixel point. When the normalized value of the similarity value is greater than the preset similarity threshold, the growth is stopped, and each connected domain in the grayscale image is obtained by using the region growing algorithm. The region growing algorithm is a well-known technology and will not be described in detail in this application.

[0058] In this embodiment, the values ​​of the preset number and the preset similarity threshold are 150 and 0.1 respectively. The values ​​of the preset number and the preset similarity threshold are preset manually and can be set by the implementer. This application does not impose any special restrictions.

[0059] The edge detection algorithm is used to perform edge detection on the grayscale image to obtain the edge curves in the grayscale image.

[0060] In this embodiment, the Canny edge detection algorithm is used to perform edge detection on the grayscale image. As other implementation methods, on the basis of being able to perform edge detection on the grayscale image, the implementer may use other existing technologies to perform edge detection on the grayscale image, such as the Sobel edge detection algorithm, the Prewitt edge detection algorithm, etc., and this application does not impose any special restrictions.

[0061] Any connected domain For example, the connected domain The minimum bounding rectangle is divided into cells, insufficient The cells can be filled with the remaining space. . Connected domain Any edge curve in For example, get the edge curve All cells of , including edge curves The meaning is to include edge curves in a cell To include edge curves No. Cells For example, calculate the cell The calculation of the gradient direction is a well-known technique and will not be described in detail in this application.

[0062] In this embodiment, The value of is 3. The value of is preset manually and can be set by the implementer. This application does not impose any special restrictions.

[0063] Cell-based The distribution of the gradient direction of all pixels in the cell is determined The direction angle is expressed as:

[0064] ; In the formula, Represents a cell Direction angle; Represents a cell The number of all kinds of gradient directions in ; Represents a cell The gradient direction is The number of pixels; Represents a cell The number of pixels in ; Represents a cell Middle Gradient direction.

[0065] It should be noted that taking the weighted average of the gradient directions of all pixels in a cell as the direction angle of the cell can integrate the local gradient information into the angle information of the cell, thereby more comprehensively describing the texture information of the cell.

[0066] Step 3: Determine the direction angle of each edge curve in each connected domain according to the distribution of the direction angles of all cells containing each edge curve in each connected domain.

[0067] Furthermore, based on the inclusion edge curve The distribution of the direction angles of all cells determines the edge curve The direction angle is expressed as:

[0068] ; In the formula, Represents edge curve Direction angle; Indicates that edge curves are included The number of cells; Represents edge curve The number of pixels on the Indicates that edge curves are included No. The edge curves contained in the cells The number of pixels on ; Indicates that edge curves are included No. The direction angle of each cell.

[0069] It should be noted that: cells contain edge curves The more pixels there are, the The direction angle of each cell can better represent the edge curve , so the direction angle As the confidence weight, evaluate the The direction angle of each cell represents the edge curve The reliability of the direction angle is then calculated based on all the edge curves in the connected domain. The direction angle of the cell is used to calculate the edge curve The direction angle.

[0070] Step 4: Based on the direction angles and distribution ranges of all edge curves in each connected domain, obtain each ridgeline area in each connected domain.

[0071] According to the characteristics of the ridge line, the edge curves on both sides of the ridge line are upward according to the mountain trend, and the monotonicity of the edge curves on both sides of the ridge line is inconsistent. On one side of the ridge line, the edge curve is monotonically increasing, so the direction angle of the edge curve is On the other side of the ridge line, the edge curve is monotonically decreasing, so the direction angle of the edge curve is between between.

[0072] Since the direction angles of the edge curves on both sides of the ridge line are different, and the edge curves on both sides do not cross the ridge line, the characteristics of the edge curves can be represented according to the position and direction angles of the edge curves on both sides, and then the ridge line area can be marked. The specific process is:

[0073] Still with edge curve For example, the edge curve The two endpoints of the edge curve If the edge curve The direction angle is If the edge curve The direction angle is If the two representative points are between , mark them as hollow points to obtain the edge direction distribution map. The horizontal and vertical coordinates of the edge direction distribution map are the horizontal and vertical coordinate values ​​of the representative points respectively. The positive direction of the axis is horizontal to the right. The positive direction of the axis is recorded as the grayscale image coordinate system, wherein the horizontal and vertical coordinates of the representative point are the horizontal and vertical coordinates in the grayscale image coordinate system.

[0074] In the edge direction distribution diagram, the connected domain can be clearly seen The distribution characteristics of the edge curves in the connected domain If there is a ridge line in the image, due to the monotonic changes of the edge curves on both sides of the ridge line, it can be seen that there is an obvious boundary area between the distribution areas of solid points and hollow points, namely the ridge line area. The schematic diagram of the ridge line area is shown in Figure 2 As shown, Figure 2 The middle left figure is a schematic diagram of the edge curve, where the solid curve is the ridge line and the dotted lines are the edge curves on both sides of the ridge line; Figure 2 The middle right figure is a schematic diagram of representative points, where the solid points are the direction angles in The representative points of the edge curve between The black square part is the ridge line area.

[0075] Connected Domain For example, calculate the connected domain The average number of solid points in all cells containing solid points is recorded as the mean of solid points; calculate the connected domain The mean number of hollow points in all cells containing hollow points is recorded as the hollow point mean; the connected domain The area formed by connecting all cells whose number of solid points is greater than or equal to the mean value of the solid points is taken as the solid point area; The area formed by connecting all cells whose number of hollow points is greater than or equal to the hollow point mean is taken as the hollow point area; the connected areas formed by connecting all cells whose number of solid points is less than the solid point mean and whose number of hollow points is less than the hollow point mean are obtained. If the adjacent left side of the connected area is a solid point area and the adjacent right side is a hollow point area, the connected area is taken as the ridgeline area. The schematic diagram of the ridgeline area acquisition process is shown in the figure. Figure 3 shown.

[0076] Step 5: Determine the ridgeline of each ridgeline region based on the distribution characteristics of the grayscale values ​​of the pixels in each ridgeline region.

[0077] In most ridge areas, due to light, wind and the ridge is raised, it is relatively difficult to store water at the ridge line, which is not conducive to plant growth. Because the higher the ridge line altitude, the lower the temperature, the more unfavorable it is for plant growth. Therefore, the vegetation density at the ridge line is sparser than on both sides, and the plant density at the high altitude of the ridge line is sparser than that at the low altitude of the ridge line. Therefore, the grayscale value at the ridge line is greater than that on both sides, and the grayscale value at the high altitude of the ridge line is greater than that at the low altitude of the ridge line.

[0078] By locating the ridge line, during the subsequent image segmentation, the distance between the cluster center and the pixel point can be optimized according to the positional relationship between the pixel point, the cluster center and the ridge line in the SLIC algorithm, thereby optimizing the segmentation result and avoiding the under-segmentation problem caused by segmenting the areas on both sides of the ridge line into the same superpixel block.

[0079] Any ridgeline area For example, from the ridgeline area Start from the top and go down, taking the ridgeline area in sequence The pixel with the largest gray value in each row of pixels until the ridge line area The top of the ridgeline area refers to the intersection of the horizontal line of the pixel point with the largest ordinate in the ridgeline area and the ridgeline area in the grayscale image coordinate system, and the bottom of the ridgeline area refers to the intersection of the horizontal line of the pixel point with the smallest ordinate in the ridgeline area and the ridgeline area. Perform curve fitting on all the obtained pixels and use the fitting curve as the ridgeline area. The ridgeline The schematic diagram of the pixel with the largest gray value is as follows Figure 4 As shown in the figure, the solid curve is the ridge line, the dotted box surrounding the ridge line is the ridge line area, and the triangle is the pixel with the largest grayscale value in each row of pixels in the ridge line area.

[0080] In this embodiment, the least squares method is used for curve fitting. As other implementation methods, on the basis of being able to realize curve fitting for pixel points, the implementer may use other existing technologies to perform curve fitting for pixel points, such as polynomial fitting, spline curves, etc., and this application does not impose any special restrictions.

[0081] Step 6, during the superpixel segmentation of the geological survey remote sensing image, based on the positional relationship between any cluster center, any pixel point and the ridge line, determine the determination factor for measuring the distance between any cluster center and any pixel point, and combine the maximum value of the distance between any cluster center and all pixel points to optimize the distance between any cluster center and any pixel point during the superpixel segmentation process.

[0082] The geological survey remote sensing image is segmented using the SLIC (Simple Linear Iterative Clustering) algorithm to obtain segmented superpixel blocks. The SLIC algorithm is a well-known technology and will not be described in detail in this application.

[0083] In the process of image segmentation using the SLIC algorithm, in order to avoid segmenting the pixels on both sides of a ridge line into the same superpixel block, when calculating the distance between the pixel point and the cluster center, it is necessary to consider whether the pixel point and the cluster center cross the ridge line. , any cluster center For example, based on the cluster center , Pixels The positional relationship with the ridge line determines the cluster center With pixels The determination factor of the distance between them is as follows:

[0084] The pixels on the ridge line are recorded as ridge points; when the cluster center With pixels When there is a ridge point on the line between With pixels The value of the determination factor for measuring the distance between the cluster centers is greater than or equal to 0; With pixels When there is no ridge point on the line between With pixels The value of the determination factor for measuring the distance between them is less than 0.

[0085] Furthermore, combining the determination factors and the cluster center The maximum value of the Euclidean distance between all pixels in the geological survey remote sensing image, optimizing the cluster center in the image segmentation process With pixels The distance between them is expressed as:

[0086] ; In the formula, Represents the optimized cluster center With pixels The distance between Represents the cluster center The maximum value of the Euclidean distance between all pixels in the geological survey remote sensing image; Represents the cluster center With pixels The Euclidean distance between Represents the cluster center With pixels The determining factor for measuring the distance between them.

[0087] It should be noted that if there is a ridge point on the line between the cluster center and the pixel point, it means that the cluster center and the pixel point cross the ridge line. In order to avoid dividing the cluster center and the pixel point into the same superpixel block, the distance between the cluster center and the pixel point is set to ; When there is no ridge point on the line between the cluster center and the pixel point, it means that there is no ridge line between the cluster center and the pixel point. Therefore, the distance between the cluster center and the pixel point is still the Euclidean distance between the two.

[0088] At this point, the distance in the SLIC algorithm has been optimized. According to the optimized distance, the pixels on the same side of the ridge line can be divided into the same superpixel block, while avoiding dividing the pixels on both sides of the ridge line into the same superpixel block, thereby optimizing the segmentation results of geological survey remote sensing images.

[0089] In summary, the present application calculates the direction angle of each cell based on the gradient direction of all pixels in each cell of the connected domain, and obtains the direction angle of each edge curve by weighted average method, thereby improving the accuracy of calculating the direction angle of the edge curve, thereby improving the accuracy of subsequent determination of the ridge line;

[0090] Furthermore, the ridgeline area is determined by combining the direction distribution of all edge curves in the connected domain, which fully considers the direction of all edge curves in the connected domain and improves the accuracy of obtaining the ridgeline area; the ridgeline of the ridgeline area is obtained according to the grayscale value of the pixel point in the ridgeline area, which facilitates the subsequent determination of the distance between the cluster center and the pixel point according to the position of the ridgeline;

[0091] Furthermore, the distance between the cluster center and the pixel in the SLIC algorithm is optimized according to whether the pixel point and the cluster center cross the ridge, avoiding the blurring of the ridge area due to the influence of shooting angle, lighting, etc., thereby dividing the areas on both sides of the ridge into the same superpixel block, thereby improving the accuracy of the segmentation results of geological survey remote sensing images.

[0092] The flowchart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to the embodiment of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of the code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. In some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. In the description corresponding to the flowchart and the block diagram in the accompanying drawings, the operations or steps corresponding to different boxes can also occur in a different order from the order disclosed in the description, and sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. Each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented by a dedicated hardware-based system that performs the specified functions or actions, or may be implemented by a combination of dedicated hardware and computer instructions.

[0093] It is obvious to those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and that the present application can be implemented in other specific forms without departing from the basic features of the present application. Therefore, no matter from which point of view, the above embodiments of the present application should be regarded as exemplary and non-restrictive.

Claims

1. A method for optimizing segmentation of geological survey remote sensing images. It is characterized in that The method comprises the following steps: Acquire geological survey remote sensing images and convert them into grayscale images; A preset number of pixels are randomly selected from the grayscale image as seed points, each connected domain of the grayscale image is obtained by a region growing algorithm, each edge curve in the grayscale image is obtained by an edge detection algorithm, the minimum circumscribed rectangle of each connected domain is equally divided into cells, and the direction angle of each cell is determined based on the distribution of the gradient direction of all pixels in each cell; Determine the direction angle of each edge curve in each connected domain according to the distribution of the direction angles of all cells containing each edge curve in each connected domain; Based on the direction angles and distribution ranges of all edge curves in each connected domain, each ridgeline area in each connected domain is obtained; Determine the ridgeline of each ridgeline region based on the distribution characteristics of the grayscale values ​​of the pixels in each ridgeline region; In the process of superpixel segmentation of geological survey remote sensing images, based on the positional relationship between any cluster center, any pixel point and the ridge line, the determination factor of the distance between any cluster center and any pixel point is determined, and combined with the maximum value of the distance between any cluster center and all pixel points, the distance between any cluster center and any pixel point in the superpixel segmentation process is optimized.

2. A method for optimizing segmentation of geological survey remote sensing images according to claim 1, It is characterized in that The process of determining the direction angle of the cell is as follows: The ratio of the number of pixels in various gradient directions in each cell to the total number of pixels is calculated, the product of various gradient directions and the corresponding ratio is calculated, and the sum of the products of all gradient directions in each cell is used as the direction angle of each cell.

3. A method for optimizing segmentation of geological survey remote sensing images as claimed in claim 1, It is characterized in that The process of determining the direction angle of the edge curve is as follows: For each edge curve in each connected domain, obtain the number of pixel points on the edge curve contained in each cell in the connected domain, calculate the ratio of the number to the number of pixel points on the edge curve, and record it as the quantity ratio; Based on the ratio of the direction angles of all cells containing edge curves in the connected domain to the quantity, the direction angles of the edge curves in the connected domain are determined.

4. A method for optimizing segmentation of geological survey remote sensing images as claimed in claim 3, It is characterized in that The expression of the direction angle of the edge curve is: ; In the formula, Represents edge curve Direction angle; Indicates that edge curves are included The number of cells; Represents edge curve The number of pixels on the Indicates that edge curves are included No. The edge curves contained in the cells The number of pixels on ; Indicates that edge curves are included No. The direction angle of each cell.

5. A method for optimizing segmentation of geological survey remote sensing images as claimed in claim 1, It is characterized in that The process of determining the ridgeline area is as follows: Combining the positions and direction angles of all edge curves in each connected domain, an edge direction distribution map is constructed; Based on the edge direction distribution map, the mean of the number of solid points in all cells containing solid points and the mean of the number of hollow points in all cells containing hollow points in any connected domain are recorded as the solid point mean and the hollow point mean, respectively; Connecting all cells in any connected domain that contain a number of solid points greater than or equal to the mean value of the solid points into a solid point region; Connecting all cells in any connected domain that contain hollow points whose number is greater than or equal to the hollow point mean into a hollow point region; The ridgeline area in any connected domain is obtained by combining the solid point mean, the hollow point mean, the solid point area and the hollow point area.

6. A method for optimizing segmentation of geological survey remote sensing images as claimed in claim 5, It is characterized in that The construction process of the edge direction distribution map is as follows: The two end points of each edge curve are taken as two representative points of each edge curve; The horizontal and vertical coordinates in the edge direction distribution diagram are the horizontal and vertical coordinate values ​​of the representative point respectively; When the direction angle of any edge curve is If the direction angle of any edge curve is between If the two representative points are between , then their two representative points are marked as hollow points.

7. A method for optimizing segmentation of geological survey remote sensing images as claimed in claim 5, It is characterized in that Get the connected areas formed by connecting all cells whose number of solid points is less than the average of the solid points and whose number of hollow points is less than the average of the hollow points. If the adjacent left side of the connected area is a solid point area and the adjacent right side is a hollow point area, the connected area is taken as the ridge line area.

8. A method for optimizing segmentation of geological survey remote sensing images as claimed in claim 1, It is characterized in that The process of determining the ridge line is as follows: obtaining the pixel with the largest grayscale value in each row of pixels in any ridge line area, and taking the fitting curve of all the obtained pixels as the ridge line of any ridge line area.

9. A method for optimizing segmentation of geological survey remote sensing images as claimed in claim 1, It is characterized in that The determination method of the determination factor is: The pixels on the ridge line are recorded as ridge points; When there is a ridge point on the line connecting any cluster center and any pixel point, the value of the determination factor is greater than or equal to 0; Otherwise, the value of the determination factor is less than 0.

10. A method for optimizing segmentation of geological survey remote sensing images according to claim 1, It is characterized in that The distance between any cluster center and any pixel point in the process of optimizing superpixel segmentation further includes: ; In the formula, Represents the optimized cluster center With pixels The distance between Represents the cluster center The maximum value of the Euclidean distance between all pixels in the geological survey remote sensing image; Represents the cluster center With pixels The Euclidean distance between Represents the cluster center With pixels The determining factor for measuring the distance between them.