Methods for monitoring and assessing soil erosion based on remote sensing technology

Abstract

The present application relates to the field of image processing technology, and specifically to a method for monitoring and evaluating soil erosion based on remote sensing technology. The method includes: obtaining an image to be analyzed, dividing the area, obtaining the soil erosion area according to the regional differences; obtaining the surface variation coefficient according to the difference in the number of corner points in the same area of ​​different images, the mean square error of the pixel value, and the variance of the pixel value; obtaining the structural variation coefficient according to the difference in the number of pixels in the same area of ​​different images; and obtaining the degradation factor in combination with the surface variation coefficient; obtaining the environmental impact coefficient according to the distance between the areas and the degradation factor of the areas; and obtaining the detail loss factor in combination with the display difference coefficient; calculating the retention threshold by combining the detail loss factor with the edge strength of the pixel point, completing the image segmentation by filtering the pixel points through the retention threshold, and evaluating the segmented image. The present application improves the degree of detail retention of the area during image segmentation in the feature extraction stage in the traditional evaluation method.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a method for monitoring and evaluating soil erosion based on remote sensing technology. Background Art

[0002] Soil erosion refers to the phenomenon that soil and water are lost from the original land surface due to natural factors or human activities. Modern remote sensing technology uses satellite or aerial sensors to obtain surface images and data to accurately assess the degree, scope and trend of soil erosion. The monitoring and assessment methods based on remote sensing technology combine a variety of advanced technologies to improve the efficiency and accuracy of monitoring, and realize dynamic tracking and long-term data analysis. By continuously optimizing model algorithms and monitoring strategies, these methods provide a scientific basis for soil erosion prevention and control, and promote regional ecological environmental protection and sustainable development.

[0003] Existing soil erosion monitoring methods based on remote sensing technology mainly rely on satellite remote sensing images to evaluate soil erosion through feature extraction. However, due to the uneven spatial distribution of soil erosion, remote sensing images are difficult to accurately capture local subtle changes, especially in small-scale and complex terrain areas. The complexity of the surface environment and occlusion factors lead to the loss of local detail features, thus affecting the reliability of the monitoring results. Therefore, a more efficient and reliable soil erosion monitoring and assessment method is urgently needed. Summary of the invention

[0004] In order to solve the technical problem of loss of detail features, this application provides a soil erosion monitoring and assessment method based on remote sensing technology. The technical solutions adopted are as follows:

[0005] This application proposes a soil erosion monitoring and assessment method based on remote sensing technology, which includes the following steps:

[0006] Collect real-time remote sensing images, and obtain images to be analyzed after preprocessing the real-time remote sensing images;

[0007] The image to be analyzed is divided into several super-pixel regions, and a display difference coefficient is obtained according to the average of the differences between the mean values ​​of the pixel values ​​of the super-pixel region and all its adjacent super-pixel regions, and the soil erosion area is screened out from the super-pixel region by the display difference coefficient;

[0008] The image at the previous adjacent moment of the image to be analyzed is recorded as the adjacent image; the minimum bounding rectangle of each soil erosion area of ​​the image to be analyzed and the same minimum bounding rectangle in the adjacent image are obtained; the number of characteristic corner points in the minimum bounding rectangle is detected, and the surface variation coefficient of the soil erosion area of ​​the image to be analyzed is obtained according to the difference in the number of characteristic corner points in the same minimum bounding rectangle of the image to be analyzed and the adjacent image, the mean square error of the pixel value and the variance of all pixel points of the minimum bounding rectangle of the image to be analyzed; the variance of all pixel values ​​in the minimum bounding rectangle of the soil erosion area is calculated as the image of the soil erosion area Pixel disorder coefficient; the surface variation coefficient is negatively correlated with the difference in the number of characteristic corner points of the same minimum circumscribed rectangle of the image to be analyzed and its adjacent images, positively correlated with the mean square error of the pixel values ​​of the same minimum circumscribed rectangle of the image to be analyzed and its adjacent images, and positively correlated with the pixel disorder coefficient of the soil erosion area of ​​the image to be analyzed; the structural variation coefficient of the soil erosion area is obtained according to the difference in the number of pixels in the same minimum circumscribed rectangle of the image to be analyzed and its adjacent images; the degradation factor of the soil erosion area is obtained according to the structural variation coefficient and surface variation coefficient of the soil erosion area;

[0009] Obtain the distance between the soil erosion areas in the image to be analyzed; obtain the environmental impact coefficient based on the distance between the soil erosion area and its surrounding area, the degradation factor of the soil erosion area and the degradation factor of its surrounding area; obtain the detail loss factor of the soil and water history area based on the display difference coefficient and environmental impact coefficient of the soil erosion area;

[0010] The retention threshold is calculated by combining the detail missing factor with the edge strength of the pixel points, and the pixel points are screened by the retention threshold to complete edge detection; the image to be analyzed after edge detection is segmented, and the detail missing factor of each area after segmentation is calculated. The neural network is trained according to the detail missing factor to evaluate the degree of soil erosion.

[0011] In the above scheme, this application aims at the existing soil erosion monitoring method based on remote sensing technology. Due to the uneven spatial distribution of soil erosion, remote sensing images are difficult to accurately capture local subtle changes, especially in small-scale and complex terrain areas. The complexity of the surface environment and occlusion factors lead to the loss of local detail features. Based on this, by dividing the remote sensing image into regions and identifying different surface types, the loss of detail features in local areas is analyzed. According to these analysis results, an adaptive image segmentation strength is set to improve the degree of detail retention of local areas during image segmentation in the feature extraction stage of traditional evaluation methods, so as to ensure the accurate extraction of terrain detail features in complex terrain areas.

[0012] In one embodiment, the method for obtaining the display difference coefficient according to the difference between the mean values ​​of the pixel values ​​of the super pixel region and all its adjacent super pixel regions is:

[0013] The adjacent super-pixel regions of each super-pixel region are recorded as neighboring regions;

[0014] The mean of the difference between the mean pixel values ​​of the superpixel region and its neighboring regions is taken as the display difference coefficient.

[0015] In one embodiment, the method for obtaining the structural variation coefficient of the soil erosion area according to the difference in the number of pixels in the same minimum circumscribed rectangle of the image to be analyzed and its adjacent images is:

[0016] , Represents the number of edge pixels in the minimum bounding rectangle of the j-th soil erosion area of ​​the image a to be analyzed, It represents the number of edge pixels in the minimum bounding rectangle of the j-th soil erosion area of ​​the image a to be analyzed in the adjacent image b. Represents the structural change coefficient of the jth soil erosion area.

[0017] In one embodiment, the method for obtaining the degradation factor of the soil erosion area according to the structural variation coefficient and the surface variation coefficient of the soil erosion area is:

[0018] The degradation factor is positively correlated with the structural variation coefficient and the surface variation coefficient respectively.

[0019] In one embodiment, the method for obtaining the distance between the soil erosion areas in the image to be analyzed is:

[0020] The spatial coordinate system is used to obtain the centroid coordinates of all soil erosion areas in the image to be analyzed, and the Euclidean distance between the centroids of two soil erosion areas is used as the distance between the two soil erosion areas.

[0021] In one embodiment, the method for obtaining the environmental impact coefficient according to the distance between the soil erosion area and its surrounding area, the degradation factor of the soil erosion area and the degradation factor of its surrounding area is:

[0022] Select a preset number of soil erosion areas closest to each soil erosion area as the surrounding areas;

[0023] The environmental impact coefficient is positively correlated with the degradation factor of the soil erosion area and the degradation factor of the surrounding area, and is negatively correlated with the distance between the soil erosion area and its surrounding areas.

[0024] In one embodiment, the method for obtaining the detail missing factor of the soil and water historical area according to the display difference coefficient and the environmental impact coefficient of the soil and water loss area is:

[0025] The detail missing factor is positively correlated with the display difference coefficient and environmental impact coefficient of soil erosion areas.

[0026] In one embodiment, the method for calculating the retention threshold by combining the detail loss factor with the edge strength of the pixel point is:

[0027] , Indicates Detail missing factor for soil erosion areas, Indicates The edge strength of the oth target pixel in the soil erosion area, Indicates The retention threshold of the oth target pixel in the soil and water loss area.

[0028] In one embodiment, the method of completing edge detection by filtering pixel points through retaining thresholds is:

[0029] In the process of non-maximum suppression of the Canny operator, pixel points whose edge strength is greater than the retention threshold are selected in the direction of the uniform continuous gradient; thus, all edge points of the image to be analyzed are obtained after adjusting the non-maximum suppression step.

[0030] The beneficial effects of this application are:

[0031] This application aims at the existing soil erosion monitoring methods based on remote sensing technology. Due to the uneven spatial distribution of soil erosion, remote sensing images are difficult to accurately capture local subtle changes, especially in small-scale and complex terrain areas. The complexity of the surface environment and occlusion factors lead to the loss of local detail features. Based on this, by dividing the remote sensing images into regions and identifying different surface types, the loss of detail features in local areas can be analyzed. According to these analysis results, an adaptive image segmentation strength is set to improve the degree of detail retention in local areas during image segmentation in the feature extraction stage of traditional evaluation methods, so as to ensure the accurate extraction of terrain detail features in complex terrain areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] 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.

[0033] Figure 1 A flow chart of a method for monitoring and assessing soil and water loss based on remote sensing technology provided by one embodiment of the present application;

[0034] Figure 2 Detect the edges of the region in the image. DETAILED DESCRIPTION

[0035] In order to further explain the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of a soil and water loss monitoring and assessment method based on remote sensing technology proposed in the present application in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0036] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0037] Example of soil and water loss monitoring and assessment method based on remote sensing technology:

[0038] The specific scheme of a soil and water loss monitoring and assessment method based on remote sensing technology provided by the present application is described in detail below with reference to the accompanying drawings.

[0039] See also Figure 1 , which shows a flow chart of a soil and water loss monitoring and assessment method based on remote sensing technology provided by an embodiment of the present application, the method comprising the following steps:

[0040] Step S001, obtaining an image to be analyzed.

[0041] GIS (Geographic Information System) is a system that integrates spatial data collection, storage, management, analysis, processing and display. Its core function is to process geospatial information and combine this information with attribute data to support spatial analysis and decision making.

[0042] The GIS system is used to obtain real-time remote sensing images of the target monitoring area, and the real-time remote sensing images are transmitted to the data analysis platform through the wireless network. At the same time, the platform automatically stores the received image data once a week, and pre-processes the collected image data. The pre-processing content includes image denoising, enhancement and other related known technologies. The pre-processed real-time remote sensing images are recorded as images to be analyzed.

[0043] At this point, the image to be analyzed has been acquired.

[0044] Step S002, screening soil and water loss areas according to the pixel value differences of the superpixel areas.

[0045] Due to the complexity of the surface environment, local areas of remote sensing images may lose detailed features, which may lead to a decrease in the reliability of the monitoring and assessment results of soil erosion. Based on this, by dividing the remote sensing images into regions and identifying different surface types, the loss of detailed features in local areas can be analyzed. According to these analysis results, the corresponding image processing technology is implemented to ensure the accurate extraction of terrain detail features in complex terrain areas.

[0046] Soil erosion areas lack vegetation cover, and exposed soil or rocks reflect more light, so they appear lighter in remote sensing images. Areas with good vegetation cover appear darker due to vegetation occlusion and light absorption.

[0047] The image to be analyzed is divided into several superpixel regions with similar pixel features using a superpixel segmentation algorithm. For each superpixel region, the mean pixel value of all pixels in the superpixel region is calculated, and the adjacent superpixel regions of each superpixel region are recorded as neighboring regions; the display difference coefficient is obtained based on the difference between the mean pixel values ​​of each superpixel region and all its neighboring regions.

[0048] Preferably, in this embodiment, the expression showing the coefficient of difference is:

[0049] , represents the mean pixel value of all pixels in the i-th superpixel area, represents the mean pixel value of all pixels in the mth neighboring area of ​​the i-th superpixel area, represents all neighboring regions of the i-th superpixel region, Represents the display difference coefficient of the i-th superpixel area.

[0050] In remote sensing images, the pixel performance of soil erosion areas is usually significantly different from that of normal areas. Therefore, these areas often have obvious pixel differences in the overall image. Therefore, the pixel values ​​of each superpixel area and its neighboring areas in the image to be analyzed are analyzed to determine whether the superpixel area is a soil erosion area. If it is not a soil erosion area, the difference between the areas is small, and the difference coefficient is small; if it is a soil erosion area, the difference between the areas is large, and the difference coefficient is large.

[0051] Therefore, in this embodiment, a soil loss coefficient is preset, and if the displayed difference coefficient is greater than the soil loss coefficient, the superpixel area is regarded as a soil loss area, otherwise it is regarded as a normal area. In this embodiment, the soil loss coefficient is set to 0.5.

[0052] At this point, all superpixel areas are divided into soil erosion areas and normal areas.

[0053] Step S003, obtaining the surface variation coefficient according to the difference in the number of feature corner points between the image to be analyzed and the adjacent image in the same minimum bounding rectangle, the mean square error of the pixel values, and the variance of all the pixels in the minimum bounding rectangle of the image to be analyzed; obtaining the structural variation coefficient according to the difference in the number of pixels in the same minimum bounding rectangle between the image to be analyzed and the adjacent image; and obtaining the degradation factor by combining the two.

[0054] In the process of soil erosion, soil and sand will be washed away and flow to low-lying areas of the terrain, forming local quicksand or sediment retention areas. As the soil erosion phenomenon intensifies, the surface soil and sand are continuously taken away, exposing the underground rocks or harder minerals. After long-term weathering and erosion, these exposed rocks may form bare surface rocks.

[0055] In the early stage of soil erosion, soil may be washed into loose sand or mud and form flowing liquid materials. These materials are usually unevenly attached to the surface, but will not significantly change the surface structural characteristics of the local area, but will only affect the uneven distribution of image pixel values ​​in the local area.

[0056] For the image to be analyzed, the image at the previous adjacent moment is recorded as the adjacent image; the soil erosion area and the normal area are framed separately to obtain the minimum bounding rectangle of the area. When the images are collected in time sequence, the area is tracked by the target tracking algorithm to obtain the position of each minimum bounding rectangle in the next image.

[0057] For each soil erosion area, all characteristic corner points within the minimum bounding rectangle of the soil erosion area are obtained by a corner detection algorithm. The corner detection algorithm used in this embodiment is the Shi-Tomasi corner detection algorithm. The variance of all pixel values ​​within the minimum bounding rectangle of the soil erosion area is calculated as the pixel disorder coefficient of the soil erosion area, and the mean square error of all pixel values ​​of the minimum bounding rectangle of the image to be analyzed and the minimum bounding rectangle corresponding to its adjacent image is calculated.

[0058] The surface variation coefficient of the soil and water erosion area of ​​the image to be analyzed is obtained according to the difference in the number of characteristic corner points of the minimum circumscribed rectangle of the soil and water erosion area of ​​the image to be analyzed and the minimum circumscribed rectangle of its corresponding adjacent image, the mean square error and the pixel disorder coefficient of the soil and water erosion area of ​​the image to be analyzed.

[0059] The surface variation coefficient is negatively correlated with the difference in the number of characteristic corner points between the minimum bounding rectangle of the soil erosion area of ​​the image to be analyzed and the minimum bounding rectangle of its corresponding adjacent image, positively correlated with the mean square error between the minimum bounding rectangle of the soil erosion area of ​​the image to be analyzed and the minimum bounding rectangle of its corresponding adjacent image, and positively correlated with the pixel disorder coefficient of the soil erosion area of ​​the image to be analyzed.

[0060] It should be noted that positive correlation means that when one variable increases, the other variable also increases, and the two variables change in the same direction. When one variable changes from large to small or from small to large, the other variable also changes from large to small or from small to large. The specific relationship is determined by actual application and this application does not impose any special restrictions.

[0061] It should be noted that negative correlation means that when one variable increases, the other variable decreases accordingly, and the two variables change in opposite directions. When one variable changes from large to small or from small to large, the other variable also changes from small to large or from large to small. The specific relationship is determined by actual application and this application does not impose any special restrictions.

[0062] Preferably, in this embodiment, the expression of the surface variation coefficient is:

[0063] , The number of characteristic corner points of the minimum bounding rectangle of the j-th soil erosion area of ​​the image a to be analyzed, The number of characteristic corner points of the minimum bounding rectangle of the j-th soil erosion area in the adjacent image b, Indicates the first The pixel value set of the minimum bounding rectangle of the soil erosion area, represents the first The pixel value set of the minimum bounding rectangle of the soil erosion area, represents the mean square error function, represents the pixel disorder coefficient of the j-th soil erosion area of ​​the image a to be analyzed, Represents the surface variation coefficient of the jth soil erosion area.

[0064] Preferably, in another embodiment of the present application, the expression of the surface variation coefficient is:

[0065] , The number of characteristic corner points of the minimum bounding rectangle of the j-th soil erosion area of ​​the image a to be analyzed, The number of characteristic corner points of the minimum bounding rectangle of the j-th soil erosion area in the adjacent image b, Indicates the first The pixel value set of the minimum bounding rectangle of the soil erosion area, represents the first The pixel value set of the minimum bounding rectangle of the soil erosion area, represents the mean square error function, represents the pixel disorder coefficient of the j-th soil erosion area of ​​the image a to be analyzed, Represents the surface variation coefficient of the jth soil erosion area.

[0066] In the early stage of soil erosion, the region is disturbed by loose sand or mud, causing the pixel performance of the local area image to change continuously with the time series. However, due to the fluidity of sand or mud itself, the surface characteristics of the local area are usually not significantly changed in the initial stage. Therefore, the larger the surface variation coefficient, the more obvious the quicksand phenomenon is and the more serious the soil erosion is.

[0067] In the early stages of soil erosion, quicksand exists in liquid form and can evenly fill local areas, resulting in an inconspicuous increase in edge pixels. However, as soil erosion intensifies, soil and sand are gradually carried away, and harder rocks or minerals underground may be exposed. After long-term weathering and erosion, these exposed rocks form obvious surface features. In extreme cases, soil erosion may also cause surface subsidence or depression, further exacerbating soil loss, thereby increasing edge pixels in remote sensing images.

[0068] Therefore, the edge detection algorithm can be used to obtain all edge curves inside the target soil erosion area and count the number of edge pixels in the soil erosion area. In this embodiment, the edge detection algorithm used is the Canny edge detection algorithm, where the edge detection image is as follows: Figure 2 shown.

[0069] The structural variation coefficient of the soil and water erosion area is obtained according to the difference between the number of edge pixels in the minimum bounding rectangle of the soil and water erosion area and the number of edge pixels in the minimum bounding rectangle of its corresponding adjacent image.

[0070] Preferably, in this embodiment, the expression of the structural variation coefficient is:

[0071] , Represents the number of edge pixels in the minimum bounding rectangle of the j-th soil erosion area of ​​the image a to be analyzed, Represents the number of edge pixels in the minimum bounding rectangle of the j-th soil erosion area in the neighboring image b, Represents the structural change coefficient of the jth soil erosion area.

[0072] For the soil erosion area, when the flowing sand particles continue to decrease, resulting in the exposure of the lower surface rock, there will be significant structural changes in the image of the area. This change is manifested as a gradual increase in the number of pixels at the edge of the area, reflecting a significant change in the surface characteristics. Among them, if the difference in the number of edge pixels of the minimum circumscribed rectangle of the soil erosion area and the minimum circumscribed rectangle of the soil erosion area corresponding to its adjacent image is greater, it means that the soil erosion furniture has led to the exposure of rocks, that is, the greater the structural change coefficient, the more serious the soil erosion.

[0073] The degradation factor of the soil and water erosion area is obtained based on the structural variation coefficient and surface variation coefficient of the soil and water erosion area.

[0074] The degradation factor is positively correlated with the structural variation coefficient and the surface variation coefficient respectively.

[0075] Preferably, in this embodiment, the degradation factor is expressed as:

[0076] , represents the surface variation coefficient of the j-th soil erosion area, represents the structural change coefficient of the jth soil erosion area, represents the degradation factor of the jth soil erosion area.

[0077] Preferably, in another embodiment of the present application, the expression of the degradation factor is:

[0078] , represents the surface variation coefficient of the j-th soil erosion area, represents the structural change coefficient of the jth soil erosion area, represents the degradation factor of the jth soil erosion area, is an exponential function with a natural constant as its base.

[0079] For the soil erosion area, the larger the surface flow coefficient, the more obvious the quicksand phenomenon is and the higher the degree of soil erosion deterioration is. In addition, when there are large structural changes in the soil erosion area, the local soil erosion situation may deteriorate sharply.

[0080] At this point, the degradation factor of each soil erosion area was obtained.

[0081] Step S004, obtaining the environmental impact coefficient according to the distance between the soil erosion areas and the degradation factor; obtaining the detail missing factor of the soil and water history area according to the display difference coefficient and the environmental impact coefficient of the soil and water erosion area.

[0082] Because there is a consistency in the environmental ecology between different soil erosion areas, this will lead to a certain degree of mutual interference. Therefore, when evaluating the impact of each soil erosion area, the density of surrounding soil erosion areas should be comprehensively considered.

[0083] The spatial coordinate system is used to obtain the centroid coordinates of all soil erosion areas in the image to be analyzed. For each soil erosion area, the distance between the soil erosion area and all other soil erosion areas is calculated. The distance is calculated as the Euclidean distance between the centroids of the two areas.

[0084] The N nearest soil erosion areas around each soil erosion area are selected as surrounding areas, and the environmental impact coefficient is obtained according to the distance between the soil erosion area and its surrounding areas, the degradation factor of the soil erosion area and the degradation factor of its surrounding areas.

[0085] The environmental impact coefficient is positively correlated with the degradation factor of the soil erosion area and the degradation factor of the surrounding area, and is negatively correlated with the distance between the soil erosion area and its surrounding areas.

[0086] Preferably, in this embodiment, the expression of the environmental impact coefficient is:

[0087] , Represents the The number of surrounding areas corresponding to each soil erosion area, Represents the The degradation factor of the surrounding area, Represents the Soil erosion areas and The Euclidean distance between the centroid coordinates of the soil erosion areas, Represents the The degradation factor of the soil erosion area is Represents the Environmental impact coefficient of soil erosion area, Represents an exponential function with a natural constant as its base.

[0088] Preferably, in another embodiment of the present application, the expression of the environmental impact coefficient is:

[0089] , Represents the The number of surrounding areas corresponding to each soil erosion area, Represents the The degradation factor of the surrounding area, Represents the Soil erosion areas and The Euclidean distance between the centroid coordinates of the soil erosion areas, Represents the The degradation factor of the soil erosion area is Represents the Environmental impact coefficient of soil erosion area, Represents an exponential function with a natural constant as its base.

[0090] When there are dense and severe soil erosion areas around the soil erosion area, the soil erosion in these areas may carry sand and mud to the target area through wind or water flow, thereby further worsening the soil erosion problem in the target area. That is, when the distance to the soil erosion area around the soil erosion area is closer and the degradation factor is greater, the soil erosion area will be more affected and the soil erosion will be more serious. This diffusion effect not only increases the environmental burden of the region, but also increases its impact coefficient on the surrounding environment.

[0091] The detail missing factor of the soil and water history area is obtained based on the display difference coefficient and environmental impact coefficient of the soil and water loss area.

[0092] The detail missing factor is positively correlated with the display difference coefficient and environmental impact coefficient of soil erosion areas.

[0093] Preferably, in this embodiment, the expression of the detail loss factor is:

[0094] , Represents the The environmental impact coefficient of the soil erosion area is Indicates The displayed coefficient of difference of soil erosion areas is represents the linear normalization function, Indicates The detail missing factor of the soil erosion area.

[0095] Preferably, in another embodiment of the present application, the expression of the detail loss factor is:

[0096] , Represents the The environmental impact coefficient of the soil erosion area is Indicates The displayed coefficient of difference of soil erosion areas is represents the linear normalization function, Indicates The detail missing factor of the soil erosion area.

[0097] When the soil erosion area shows serious soil erosion and has a significant impact on the regional environment, it indicates that the soil erosion phenomenon in the area is relatively serious. When using traditional methods to obtain remote sensing images of the area, it will lead to a significant loss of regional detail features.

[0098] At this point, the detail missing factors of each soil erosion area were obtained.

[0099] Step S005, calculating the retention threshold according to the detail loss factor, and filtering the pixel points according to the retention threshold to complete edge detection; segmenting the image to be analyzed after edge detection, and evaluating the degree of soil and water loss according to the detail loss factor of each area after segmentation.

[0100] For the images to be analyzed, the local areas of the images have serious lack of details, which will lead to inaccurate monitoring and assessment of soil erosion by remote sensing images.

[0101] In the Canny edge detection image segmentation method, the image is first smoothed by a Gaussian filter to reduce noise. Then, the edge strength and direction of each point are calculated using the Sobel operator. Subsequently, local maximum suppression is performed to obtain the pixel with the maximum edge strength in the same continuous gradient direction, and this pixel is marked, and the marked pixel is recorded as the target pixel.

[0102] However, due to the complex surface environment and vegetation occlusion in soil erosion areas, a large number of overlapping edges may appear in the image. Traditional non-maximum suppression methods may ignore some important details of the clustered areas, resulting in loss of local information.

[0103] Therefore, in order to ensure more details, the edge information is not obtained by non-maximum suppression, but the retention threshold is set, and the target pixel is the largest pixel in the gradient direction. The larger the detail loss factor, the more serious the soil erosion, which means that more edge points are needed to reflect local information. Therefore, the retention threshold is calculated by the target pixel in the gradient direction.

[0104] The formula for the retention threshold is:

[0105] , Indicates Detail missing factor for soil erosion areas, Indicates The edge strength of the oth target pixel in the soil erosion area, Indicates The retention threshold of the oth target pixel in the soil and water loss area.

[0106] In the process of non-maximum suppression of the Canny operator, the step of selecting the maximum value in the uniform continuous gradient direction is modified to selecting pixels whose edge strength is greater than the retention threshold in the uniform continuous gradient direction. Thus, all edge points of the image to be analyzed are obtained after adjusting the non-maximum suppression step.

[0107] The image to be analyzed is segmented according to all edge points detected in the image to be analyzed, thereby obtaining multiple regions, and the detail loss factor of each region is calculated by the above-mentioned method for calculating the detail loss factor in the soil erosion region.

[0108] Obtain the detail missing factors of several historical areas in the historical archives, mark each historical area according to the severity of soil erosion, mark the area with severe soil erosion as 1, and mark the area without soil erosion as 0, and train it through the CNN neural network. The input is the detail missing factor of each area of ​​the image to be analyzed, and the output is a number between 0 and 1. The closer it is to 1, the more serious the soil erosion. The loss function used is the mean square error loss function.

[0109] The degree of soil erosion was thus assessed for each area of ​​the image to be analyzed.

[0110] It should be noted that the above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

[0111] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A soil and water loss monitoring and assessment method based on remote sensing technology, characterized in that: The method comprises the following steps: Collect real-time remote sensing images, and obtain images to be analyzed after preprocessing the real-time remote sensing images; The image to be analyzed is divided into several super-pixel regions, and a display difference coefficient is obtained according to the average of the differences between the mean values ​​of the pixel values ​​of the super-pixel region and all its adjacent super-pixel regions, and the soil erosion area is screened out from the super-pixel region by the display difference coefficient; The image at the previous adjacent moment of the image to be analyzed is recorded as the adjacent image; the minimum bounding rectangle of each soil erosion area of ​​the image to be analyzed and the same minimum bounding rectangle in the adjacent image are obtained; the number of characteristic corner points in the minimum bounding rectangle is detected, and the surface variation coefficient of the soil erosion area of ​​the image to be analyzed is obtained according to the difference in the number of characteristic corner points in the same minimum bounding rectangle of the image to be analyzed and the adjacent image, the mean square error of the pixel value and the variance of all pixel points of the minimum bounding rectangle of the image to be analyzed; the variance of all pixel values ​​in the minimum bounding rectangle of the soil erosion area is calculated as the image of the soil erosion area The pixel disorder coefficient and the surface variation coefficient are negatively correlated with the difference in the number of characteristic corner points of the same minimum circumscribed rectangle of the image to be analyzed and its adjacent images, positively correlated with the mean square error of the pixel values ​​of the same minimum circumscribed rectangle of the image to be analyzed and its adjacent images, and positively correlated with the pixel disorder coefficient of the soil erosion area of ​​the image to be analyzed; the structural variation coefficient of the soil erosion area is obtained according to the difference in the number of pixels in the same minimum circumscribed rectangle of the image to be analyzed and its adjacent images; the degradation factor of the soil erosion area is obtained according to the structural variation coefficient and the surface variation coefficient of the soil erosion area; Obtain the distance between the soil erosion areas in the image to be analyzed; obtain the environmental impact coefficient based on the distance between the soil erosion area and its surrounding area, the degradation factor of the soil erosion area and the degradation factor of its surrounding area; obtain the detail loss factor of the soil and water history area based on the display difference coefficient and environmental impact coefficient of the soil erosion area; The retention threshold is calculated by combining the detail missing factor with the edge strength of the pixel points, and the pixel points are screened by the retention threshold to complete edge detection; the image to be analyzed after edge detection is segmented, and the detail missing factor of each area after segmentation is calculated. The neural network is trained according to the detail missing factor to evaluate the degree of soil erosion.

2. The method for monitoring and assessing soil and water loss based on remote sensing technology as claimed in claim 1, characterized in that: The method for obtaining the display difference coefficient according to the difference between the mean values ​​of the pixel values ​​of the super pixel area and all its adjacent super pixel areas is: The adjacent super-pixel regions of each super-pixel region are recorded as neighboring regions; The mean of the difference between the mean pixel values ​​of the superpixel region and its neighboring regions is taken as the display difference coefficient.

3. The method for monitoring and assessing soil and water loss based on remote sensing technology as claimed in claim 1, characterized in that: The method for obtaining the structural variation coefficient of the soil erosion area according to the difference in the number of pixels in the same minimum circumscribed rectangle of the image to be analyzed and its adjacent images is: , Represents the number of edge pixels in the minimum bounding rectangle of the j-th soil erosion area of ​​the image a to be analyzed, It represents the number of edge pixels in the minimum bounding rectangle of the j-th soil erosion area of ​​the image a to be analyzed in the adjacent image b. Represents the structural change coefficient of the jth soil erosion area.

4. The method for monitoring and assessing soil and water loss based on remote sensing technology as claimed in claim 1, characterized in that: The method for obtaining the degradation factor of the soil and water loss area according to the structural variation coefficient and the surface variation coefficient of the soil and water loss area is: The degradation factor is positively correlated with the structural variation coefficient and the surface variation coefficient respectively.

5. The method for monitoring and evaluating soil and water loss based on remote sensing technology as claimed in claim 1, characterized in that: The method for obtaining the distance between the soil erosion areas in the image to be analyzed is: The spatial coordinate system is used to obtain the centroid coordinates of all soil erosion areas in the image to be analyzed, and the Euclidean distance between the centroids of two soil erosion areas is used as the distance between the two soil erosion areas.

6. The method for monitoring and evaluating soil and water loss based on remote sensing technology as claimed in claim 1, characterized in that: The method for obtaining the environmental impact coefficient according to the distance between the soil erosion area and its surrounding area, the degradation factor of the soil erosion area and the degradation factor of its surrounding area is: Select a preset number of soil erosion areas closest to each soil erosion area as the surrounding areas; The environmental impact coefficient is positively correlated with the degradation factor of the soil erosion area and the degradation factor of the surrounding area, and is negatively correlated with the distance between the soil erosion area and its surrounding areas.

7. The method for monitoring and evaluating soil and water loss based on remote sensing technology as claimed in claim 1, characterized in that: The method for obtaining the detail missing factor of the soil and water historical area according to the display difference coefficient and environmental impact coefficient of the soil and water loss area is: The detail missing factor is positively correlated with the display difference coefficient and environmental impact coefficient of soil erosion areas.

8. The method for monitoring and assessing soil and water loss based on remote sensing technology as claimed in claim 1, characterized in that: The method for calculating the retention threshold by combining the detail loss factor with the edge strength of the pixel point is: , Indicates Detail missing factor for soil erosion areas, Indicates The edge strength of the oth target pixel in the soil erosion area, Indicates The retention threshold of the oth target pixel in the soil and water loss area.

9. The method for monitoring and assessing soil and water loss based on remote sensing technology as claimed in claim 1, characterized in that: The method for completing edge detection by filtering pixel points through retaining thresholds is: In the process of non-maximum suppression of the Canny operator, pixel points whose edge strength is greater than the retention threshold are selected in the direction of the uniform continuous gradient; thus, all edge points of the image to be analyzed are obtained after adjusting the non-maximum suppression step.

Images