CN110363857B - Method and device for constructing and observing soil erosion surface morphology DEM in rainfall process - Google Patents
Method and device for constructing and observing soil erosion surface morphology DEM in rainfall process Download PDFInfo
- Publication number
- CN110363857B CN110363857B CN201910515164.0A CN201910515164A CN110363857B CN 110363857 B CN110363857 B CN 110363857B CN 201910515164 A CN201910515164 A CN 201910515164A CN 110363857 B CN110363857 B CN 110363857B
- Authority
- CN
- China
- Prior art keywords
- soil erosion
- rain
- dem
- underlying surface
- obtaining
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/20—Finite element generation, e.g. wire-frame surface description, tesselation
Abstract
The invention discloses a method and a device for constructing and observing a soil erosion earth surface form DEM in a rainfall process, wherein a gray value is processed pixel by using a K-means algorithm, and a pixel value after rain and fog removal is reconstructed from a plurality of pixel values, so that the accuracy and the efficiency of removing rain and fog of an image are improved; the shape of the underlying surface can be observed in the continuous rainfall process, and the process of soil erosion occurrence and development can be expressed in a digital and graphical mode. More importantly, the simultaneous description of the underlying surface form from two dimensions of time and space is realized, and a powerful means is provided for the scientific research of the soil erosion process and mechanism.
Description
Technical Field
The invention relates to a soil erosion morphology evolution observation method, in particular to a soil erosion earth surface morphology DEM construction and observation method and device in a rainfall process.
Background
A Digital Elevation Model (DEM) realizes digital simulation of ground topography (namely digital expression of topography surface morphology) through limited topography elevation data, and is an entity ground model for expressing ground elevation in a group of ordered numerical array forms.
On the basis of obtaining the soil erosion ground surface form DEM, the soil erosion ground surface form DEM at multiple moments can be processed to realize dynamic observation of soil erosion, and the soil erosion is a process that soil or other ground component substances are degraded, destroyed, separated, transported and deposited under the action of external force. At each point in the process, the morphology of the soil erosion surface changes constantly. Meanwhile, particles including raindrops, aggregates, silt and the like involved in the soil erosion process reach the size of a sub-centimeter or millimeter. Therefore, the method for accurately acquiring the morphological changes of the erosion surface under the microscale has important significance for researching soil erosion mechanism, soil erosion evolution process and the like, and parameters reflecting the soil erosion process mainly comprise runoff, sediment content, the topographic changes of the erosion surface and the like.
Disclosure of Invention
The invention aims to provide a method and a device for constructing and observing a soil erosion ground surface form DEM (digital elevation model) in a rainfall process, which are used for solving the problems that rain and fog in an acquired image are difficult to process by using a three-dimensional modeling method in the rainfall process and three-dimensional modeling on a soil erosion underlying surface can not be realized under a continuous rainfall condition in the prior art.
In order to realize the task, the invention adopts the following technical scheme:
1. a method for constructing a soil erosion surface morphology DEM in a rainfall process is used for obtaining the DEM of a soil erosion underlying surface in the rainfall process and is executed according to the following steps:
the method comprises the following steps that 1, N cameras are arranged right above a soil erosion underlying surface to be modeled, N is an integer larger than 1, a plurality of images of the soil erosion underlying surface to be modeled are collected simultaneously through each camera, N initial image groups are obtained, and the size of each initial image in the initial image groups is the same;
wherein, carry out the fog processing of removing rain to nth initial image group, obtain nth and remove the rain fog image, specifically include:
step 2.1, obtaining the gray value of the ith pixel point of each initial image in the nth initial image group to obtain a gray value sequence, wherein I belongs to I, I is the total number of the pixel points in each initial image in the nth initial image group, and I is a positive integer;
step 2.2, taking the minimum value of the gray values in the gray value sequence as an initial seed point C 0 Taking the intermediate value of the gray values in the gray value sequence as the initial seed point C 1 Taking the maximum value of the gray values in the gray value sequence as an initial seed point C 2 ；
Step 2.3, performing iterative clustering on all gray values in the gray value sequence obtained in the step 2.1 by utilizing a K-means clustering algorithm to obtain three new clustering centers X 0 New cluster center X 1 And a new cluster center X 2 The new cluster center X 1 Greater than new cluster center X 0 And is smaller than the new cluster center X 2 ；
Step 2.4, new clustering center X 1 The gray value of the ith pixel point after rain and fog removal is taken as the gray value of the ith pixel point after rain and fog removal;
step 2.5, repeating the steps 2.1 to 2.5 until I gray values without rain and fog are obtained;
step 2.6, collecting the I rain and fog removed gray values to obtain an nth rain and fog removed image;
step 3, reconstructing the N rain and fog removing images by using a three-dimensional reconstruction method to obtain three-dimensional point cloud data of a soil erosion underlying surface to be modeled;
and 4, interpolating the three-dimensional point cloud data of the soil erosion underlying surface to be modeled to obtain the DEM of the soil erosion underlying surface.
Further, when iterative clustering is performed on all gray values in the gray value sequence obtained in step 2.1 by using the K-means clustering algorithm in step 2.3, the iteration stop condition is that the difference between clustering centers obtained after two adjacent iterations is less than or equal to 10 -6 Or the iteration times are more than 2000, and three new clustering centers are obtained.
A soil erosion surface morphology observation method in a rainfall process is used for obtaining soil erosion amount between adjacent observation moments in the rainfall process and is executed according to the following steps:
a, collecting a DEM of a soil erosion underlying surface to be observed before rainfall;
b, respectively obtaining the DEM of the soil erosion underlying surface at a plurality of observation moments in the rainfall process by adopting the construction method of the soil erosion surface morphology DEM in the rainfall process;
step C, obtaining the volume difference between the DEM of the soil erosion underlying surface at each observation time and the DEM of the soil erosion underlying surface to be observed before rainfall, which is obtained in the step A, and obtaining the accumulated soil erosion amount at each observation time;
and D, obtaining a difference value between the accumulated soil erosion amounts at the adjacent moments, and obtaining the soil erosion amount between the adjacent observation moments.
A soil erosion earth surface form DEM construction device in a rainfall process comprises an image acquisition module:
the image acquisition module is used for simultaneously acquiring a plurality of images of the soil erosion underlying surface to be modeled to obtain N initial image groups, wherein the size of each initial image in the initial image groups is the same;
the system also comprises an image rain and fog removing module, a three-dimensional data obtaining module and a modeling module;
the image defogging module is used for respectively performing defogging processing on the N initial image groups to obtain N images for defogging;
the image rain and fog removing module comprises a gray value sequence obtaining submodule, an initial seed point selecting submodule, a clustering submodule, a gray value updating submodule and an image reconstruction submodule;
the gray value obtaining submodule is used for obtaining the gray value of the ith pixel point of each image in the nth initial image group to obtain a gray value sequence, I belongs to I, I is the total number of the pixel points in each initial image in the nth initial image group, and I is a positive integer;
the initial seed point selection submodule is used for taking the minimum value of the gray values in the gray value sequence as an initial seed point C 0 Taking the intermediate value of the gray values in the gray value sequence as the initial seed point C 1 Taking the maximum value of the gray values in the gray value sequence as an initial seed point C 2 ；
The clustering submodule is used for carrying out iterative clustering on all the gray values in the obtained gray value sequence by utilizing a K-means clustering algorithm to obtain three new clustering centers, namely a new clustering center X 0 New cluster center X 1 And a new cluster center X 2 The new cluster center X 1 Greater than new cluster center X 0 And is smaller than the new cluster center X 2 ；
The grey value updating submodule is used for updating a new clustering center X 1 The gray value of the ith pixel point after rain and fog removal is taken as the gray value of the ith pixel point after rain and fog removal;
the image reconstruction submodule is used for collecting the I rain and fog removed gray values to obtain an nth rain and fog removed image;
the three-dimensional data acquisition module is used for reconstructing the N rain and fog removing images by using a three-dimensional reconstruction method to acquire three-dimensional point cloud data of a soil erosion underlying surface to be modeled;
and the modeling module is used for interpolating the three-dimensional point cloud data of the soil erosion underlying surface to be modeled to obtain the DEM of the soil erosion underlying surface.
Further, when all the gray values in the obtained gray value sequence are subjected to iterative clustering by utilizing a K-means clustering algorithm in the clustering submodule, iteration is stoppedThe condition is that the difference value between the cluster centers obtained after two adjacent iterations is less than or equal to 10 -6 Or the iteration times are more than 2000, and three new clustering centers are obtained.
The utility model provides a rainfall in-process soil erosion earth's surface form observation device, includes reference model acquisition module, observation model acquisition module, accumulative total erosion amount acquisition module and soil erosion amount acquisition module:
the reference model acquisition module is used for acquiring a DEM of the soil erosion underlying surface to be observed before rainfall;
the observation model obtaining module is also used for respectively obtaining the DEM of the soil erosion underlying surface at a plurality of observation moments in the rainfall process by adopting the DEM constructing device of the soil erosion earth surface form in the rainfall process;
the accumulated erosion amount obtaining module is used for obtaining the difference between the DEM of the soil erosion underlying surface at each observation moment and the obtained volume of the DEM of the soil erosion underlying surface to be observed before rainfall to obtain the accumulated soil erosion amount at each observation moment;
the soil erosion amount obtaining module is used for obtaining a difference value between the accumulated soil erosion amounts at adjacent moments and obtaining the soil erosion amount between the adjacent observation moments.
Compared with the prior art, the invention has the following technical effects:
1. according to the method and the device for constructing the DEM (digital elevation model) of the soil erosion surface form in the rainfall process, the gray values in the image are processed pixel by using the K-means algorithm, the image is restored after one gray value without rain and fog is reconstructed from a plurality of gray values, the rain and fog removing image is obtained, and the accuracy and the efficiency of removing the rain and fog from the image are improved;
2. the soil erosion surface form observation method and device in the rainfall process can observe the underlying surface form in the continuous rainfall process, and digitally and graphically express the processes of soil erosion occurrence and development. More importantly, the simultaneous description of the shape of the underlying surface from two dimensions of time and space is realized, and a powerful means is provided for the scientific research of the soil erosion process and mechanism;
3. the method and the device for constructing and observing the soil erosion surface morphology DEM in the rainfall process can obtain the digital images of the underlying surface at a plurality of time points in one rainfall process, thereby realizing the real-time observation of the soil erosion process.
Drawings
FIG. 1 is an original left trough underlying surface image provided in one embodiment of the present invention;
FIG. 2 is an original right trough underlying surface image provided in one embodiment of the present invention;
FIG. 3 is an image of the underlying surfaces of left and right earthen tanks subject to erosion (photographed after rainfall) provided in one embodiment of the present invention;
FIG. 4 is an image of a set scale in an embodiment of the invention;
FIG. 5 is a schematic view of the measurement results of the scale 1 according to an embodiment of the present invention;
FIG. 6 is a schematic representation of the measurement of scale 2 in one embodiment of the present invention;
FIG. 7 is a diagram illustrating the relationship between photogrammetric values and actual values in one embodiment of the present invention;
FIG. 8 is a schematic view of a portion of a DEM obtained in one embodiment of the present invention;
FIG. 9 is a difference plot of the DEM of the soil erosion underlying surface at a portion of the observation time and the DEM of the soil erosion underlying surface to be observed before rainfall, obtained in an embodiment of the present invention.
Detailed Description
K-means clustering algorithm: the K-means clustering algorithm is a clustering analysis algorithm for iterative solution, and comprises the steps of randomly selecting K objects as initial clustering centers, then calculating the distance between each object and each seed clustering center, assigning each object to the nearest clustering center, representing a cluster by the clustering centers and the objects assigned to the clustering centers, and assigning a sample, wherein the clustering centers of the clusters are recalculated according to the existing objects in the clusters. This process will be repeated until some termination condition is met. The termination condition may be that no (or minimum number) objects are reassigned to different clusters, no (or minimum number) cluster centers are changed again, and the sum of squared errors is locally minimal.
The following are specific examples given by the inventors for further illustrating the aspects of the invention.
Example one
The embodiment discloses a method for constructing a soil erosion ground surface form DEM in a rainfall process, which is used for obtaining the DEM of a soil erosion underlying surface in a rainfall environment.
The method comprises the following steps:
step 1, arranging N cameras right above the soil erosion underlying surface to be modeled, wherein N is an integer greater than 1, and simultaneously collecting a plurality of images of the soil erosion underlying surface to be modeled by using each camera to obtain N initial image groups, wherein the size of each initial image in the initial image groups is the same;
in this embodiment, several groups of digital image collectors are networked based on a wireless network technology, and each group of digital image collectors includes a digital camera and an industrial control level computer. The number of the digital image collectors depends on the size and the shape of the area of the underlying surface and the erection height of the digital image collectors from the underlying surface;
calibrating each camera before acquiring a plurality of images of the soil erosion underlying surface to be modeled by using the cameras, wherein the calibration comprises the determination of internal parameters, distortion parameters and external parameters of the cameras;
in the embodiment, 2 × 6-12 video cameras are arranged right above a soil erosion underlying surface to be modeled, the area of the soil erosion underlying surface is 3m × 10m, the distance between each video camera and an observation surface cannot be greater than 20 m, and the overlapping degree of pictures in a three-dimensional reconstruction area cannot be less than 3;
each camera collects a plurality of images in the same period, preferably, each camera collects the same number of images which are not less than 60 images in the same period, and the time length of one period is not more than 10 seconds;
in the present embodiment, 12 cameras each capture not less than 60 images in a period of 10 s.
Because the final purpose of the method is to establish a three-dimensional model of the soil erosion underlying surface to be modeled, and the three-dimensional model needs images acquired by a plurality of cameras at the same time to be reconstructed, if each camera acquires only one image, raindrops in the images cannot be well removed, and details of the images may be lost while the raindrops are removed.
In the invention, a method of solving the background by a plurality of continuous images is adopted to remove raindrops in the image collected by each camera, namely, in the invention, one camera is used for quickly collecting a plurality of images in a short time period, and the images are processed and integrated into a rain and fog removing image.
wherein, carry out the fog processing of removing rain to nth initial image group, obtain nth and remove the rain fog image, specifically include:
step 2.1, obtaining the gray value of the ith pixel point of each image in the nth initial image group to obtain a gray value sequence, wherein I belongs to I, I is the total number of the pixel points in each initial image in the nth initial image group, and I is a positive integer;
step 2.2, taking the minimum value of the gray values in the gray value sequence as an initial seed point C 0 Taking the intermediate value of the gray values in the gray value sequence as the initial seed point C 1 Taking the maximum value of the gray values in the gray value sequence as an initial seed point C 2 ；
Step 2.3, performing iterative clustering on all gray values in the gray value sequence obtained in the step 2.1 by utilizing a K-means clustering algorithm to obtain three new clustering centers X 0 New cluster center X 1 And a new clustering center X 2 New cluster center X 1 Greater than new cluster center X 0 And is smaller than the new cluster center X 2 ；
Step 2.4, new clustering center X 1 The gray value of the ith pixel point after rain and fog removal is taken as the gray value of the ith pixel point after rain and fog removal;
step 2.5, repeating the steps 2.1 to 2.5 until I gray values without rain and fog are obtained;
step 2.6, collecting the I gray values without rain and fog to obtain the nth image without rain and fog;
in the invention, because a plurality of photos are collected in a period, raindrops do not always exist in the period, but the raindrops are in an intermittent downward dripping state, the invention adopts a pixel-by-pixel processing method to compare pixel points at the same position in a plurality of images, and rebuilds a pixel without rain fog as a real pixel at the position on the image collected by the camera.
Specifically, the invention uses the gray value as a reference, uses the pixels with lower gray value and the pixels with higher gray value as raindrop interference or system noise, and should be eliminated, and uses the pixels with middle gray value as the final reference pixels.
In this embodiment, one camera acquires 60 images, and 60 pixel gray values are shared at the same position of the 60 images, and the K-means clustering algorithm is used to cluster the 60 gray values, which is different from the common clustering method in that the initial seed points are required to be set in the present invention (the initial seed points are randomly determined in the common clustering method), because the final purpose is to divide the gray values into three categories, 3 initial seed points are set, and the minimum value of all the gray values is respectively used as the initial seed point C 0 Taking the middle value of all gray values as the initial seed point C 1 Taking the maximum value of all gray values as the initial seed point C 2 Then, calculating the distance between each gray value and the three initial seed points by using a K-means clustering method, and then iterating to finally obtain three new clustering centers, namely X 0 、X 1 And X 2 The three cluster center seed points represent three gray values with lower gray value, middle gray value and higher gray value, the middle gray value is used as the gray value at the position, the gray value at each pixel point position is updated by adopting the method, and finally a rain and fog removing image is obtained.
In this embodiment, when the I gray values without rain fog are collected in step 2.6, it is actually a process of backfilling the gray values, and each gray value is backfilled to the position of the original pixel point according to the position of the pixel point corresponding to the gray value, so as to obtain the image without rain fog.
Optionally, when iterative clustering is performed on all gray values in the gray value sequence obtained in step 2.1 by using a K-means clustering algorithm in step 2.3, the iteration stop condition is that a difference between clustering centers obtained after two adjacent iterations is less than or equal to 10 -6 Or the iteration times are more than 2000, and three new clustering centers are obtained.
Step 3, reconstructing the N rain and fog removing images by using a three-dimensional reconstruction method to obtain three-dimensional point cloud data of a soil erosion underlying surface to be modeled;
and 4, interpolating the three-dimensional point cloud data of the soil erosion underlying surface to be modeled to obtain the DEM of the soil erosion underlying surface.
In this embodiment, the obtained DEM of the soil erosion underlying surface is as shown in fig. 8, and the DEM of the soil erosion underlying surface is 30min, 60min, 90min, 110min, 130min and 150min respectively, High represents soil erosion, and Low represents soil accumulation.
Example two
A soil erosion surface form observation method in a rainfall process is used for obtaining soil erosion amount between adjacent observation moments in the rainfall process and is executed according to the following steps:
step A, collecting a DEM of a soil erosion underlying surface to be observed before rainfall, and taking the model as a reference for erosion amount calculation;
and finishing the construction of a reference plane in the step, obtaining three-dimensional point cloud data and a digital elevation model which are not corroded, and taking the three-dimensional point cloud data and the digital elevation model as the reference plane of which the underlying surface is changed in shape after corrosion occurs.
Step B, respectively obtaining the DEM of the soil erosion underlying surface at a plurality of observation moments in the rainfall process by adopting the method for constructing the soil erosion surface morphology DEM in the rainfall process according to any one of claims 1-2;
in this embodiment, after the digital elevation model of the soil erosion underlying surface at a plurality of observation times in the rainfall process is obtained, on the basis, the digital elevation model is introduced into the GIS software, parameters such as the surface area of the slope surface and the total area of the rill can be obtained by using a profile analysis tool, a hydrological analysis tool and the like in a three-dimensional analysis module of the GIS software, and morphological parameters such as the morphology of the erosion rill and the density of the trench network of the soil erosion underlying surface to be observed at each time can be obtained by using the parameters.
Step C, obtaining the volume difference between the DEM of the soil erosion underlying surface at each observation time and the DEM of the soil erosion underlying surface to be observed before rainfall, which is obtained in the step A, and obtaining the accumulated soil erosion amount at each observation time;
and D, obtaining a difference value between the accumulated soil erosion amounts at the adjacent moments, and obtaining the soil erosion amount between the adjacent observation moments.
In this embodiment, the obtained differences between the DEM of the soil erosion underlying surface at the partial observation time and the DEM of the soil erosion underlying surface to be observed before rainfall are, as shown in fig. 9, differences between the DEM of the soil erosion underlying surface to be observed before rainfall and the DEM of the soil erosion underlying surface to be observed before rainfall, which are 30min, 60min, 90min, 110min, 130min, and 150min, respectively, High represents soil erosion, and Low represents soil accumulation.
EXAMPLE III
In this embodiment, the effectiveness of the model construction and erosion observation method provided by the present invention is tested in 2018, 10/4 in loess plateau soil erosion and artificial rainfall simulation in key laboratories of dry land agricultural countries.
The experiment is carried out in the morning and afternoon of 10 and 4 months in 2018 by arranging two steel grooves which are parallel left and right. In the experiment, the intensity of artificial rainfall simulation is respectively 90mm per hour and 120 mm per hour, and the constructed underlying surfaces are shown in figures 1, 2 and 3, wherein figures 1 and 2 are respectively pictures of left and right soil tank underlying surfaces which are not eroded (shot before rainfall), figure 3 is a picture of left and right soil tank underlying surfaces which are eroded (shot after rainfall), the bottom of a steel tank is an inclined plane, soil to be measured is placed in the steel tank, the soil type is yellow cotton soil, the soil filling bulk density is 1.15g/cm3, the sizes of the soil tanks of the left and right underlying surface devices are both 1 meter wide and 10 meter long cubic containers, and the bottom surfaces of the underlying surfaces and the horizontal plane respectively form an inclined angle of 15 degrees and 20 degrees.
Shooting by time intervals and calculating a digital image, and observing soil erosion of an underlying surface by computer vision: collecting digital images of the experiment underlying surface in a full-coverage manner from the beginning of rainfall to the end of rainfall, wherein the collecting time is respectively collected every 5min after the beginning of rainfall, and the collecting density of the digital images is 150-170 frames/m 2 ；
Collecting a full sample of water and sand and converting loss: collecting a full water-sand sample in the experimental process according to the same time interval acquired by the digital image of the soil erosion underlying surface after rainfall begins, sampling and measuring the sediment content in the sample by adopting a portable sediment measuring instrument, measuring the volume of the water-sand sample, calculating the weight of soil contained in the sample, and converting corresponding volume quantities according to the measured volume weights of 1.372g/cm3 and 1.374g/cm3 of the left soil tank and the right soil tank respectively.
In this embodiment, 12 sets of image sensors are networked; the 12 camera groups are fixed on a reinforcing steel bar plate frame which is 14 meters high away from the soil erosion underlying surface, the cameras are uniformly arranged, and the cameras and the soil erosion underlying surface are arranged in a vertical direction. Test field area for calibration: the total number of cameras to be calibrated is 12 in the case of 40 square meters, all the cameras are non-measuring cameras, and the internal parameters, the external parameters and the distortion parameters of the cameras need to be calibrated so as to obtain the high-precision point location information of the ground. The parameter calibration sequence is that the internal parameter and the distortion parameter are calibrated first, and then the external parameter is calibrated. The adopted camera calibration method is Zhangzhenyou chessboard format calibration method and the layout of control points.
Firstly, measuring the size of a scale: the measurement of the scale 1 with the size of 408.5mm and the scale 2 with the size of 310mm are measured 60 times respectively, and the results are shown in tables 1 and 2, and the size of the scale is measured under a unified reference coordinate system by using the two target sizes as length constraints, as shown in fig. 4.
Counting the size results of all the measured scales, and calculating the relative error with the actual size.
TABLE 1 measurement of the method of the invention on a ruler 1 with a dimension of 408.5mm
Table 2 measurement of the method of the invention on a scale 2 with a dimension of 310mm
The statistical analysis is carried out on the 60 measurement results of the two scales to obtain: the average length of scale 1 is 407.58mm, and the standard deviation of the measurement is 1.422 mm; the average length of the scale 2 is 309.27mm, the standard deviation of the measurement is 1.726mm, and the observation precision reaches millimeter level; further, both the scale 1 and scale 2 measurements obey a normal distribution by the K-S test, as shown in fig. 5 and 6.
(2) Accuracy detection
Under the conditions that the gradient is 15 degrees and the rain strength is 90mm/h, the system is adopted to measure the geometric dimensions of three grooves with different sizes and shapes distributed on the slope surface of the soil tank, the relative error with the actual value is calculated and counted, and the accuracy of the system is evaluated, and the result is shown in table 3.
Table 3 shows the observation results of the method of the invention on the standard material object under the conditions that the gradient is 15 degrees and the rain strength is 90mm/h
(unit: mm)
The relation between the photographic measured value and the actual value under the conditions that the gradient is 15 degrees and the rain strength is 90mm/h is 1.0014x-1.5738, and the relation between the photographic measured value and the actual value is 0.9999, which is shown in fig. 7, and R2 is 0.9999; the measured value and the measured value of the method are obviously and positively correlated, which shows that the method of the invention is accurate in observing the size of the underlying surface.
The detection method of the invention is compared with the detection observation method of the laser scanner for verification: under the same conditions observed by the method, the laser scanner is adopted to observe the slope surface of the soil groove, three-dimensional digital point cloud of the slope surface is obtained, the size of the groove of the underlying surface is calculated, the result is shown in a table 4, and the obtained partial digital elevation model is shown in a figure 8.
TABLE 4 Standard object Observation results when the gradient is 15 ° and the rain strength is 90mm/h
(unit: mm)
Parallel observation with the laser scanner showed that: the method overcomes the defect that the laser scanner cannot observe the evolution of the underlying surface in the rainfall process; and secondly, the observation precision of the method is higher than that of a laser scanner for the measurement result of the same size.
The result of parallel observation with the traditional runoff sediment collection method shows that the overall average accuracy of the left soil tank is-7.69%; the global average accuracy of the right soil box was 0.95% (tables 5-6).
The detection method of the invention is compared with the detection observation method of the three-dimensional laser scanner:
in the embodiment, after the rainfall is observed in the experiment, the observation result of the three-dimensional laser scanner is compared with the observation result of the soil tank on the right side of the method during the last rainfall, and the soil erosion amount obtained by observation is calculated and calculated as shown in the table 7.
TABLE 5 comparison of the method of the present invention with the runoff sand method (left soil box)
TABLE 6 comparison of radial silt method and photogrammetry (right trough)
TABLE 7 comparison of the method of the invention with laser scanning observations
Parallel observation with the laser scanner showed that: the total soil erosion amount observed by the method provided by the invention is 452180cm 3 The total soil erosion amount observed by laser scanning is 435322.51cm 3 Therefore, the method and the laser scanning method have the observation errors of 3.87 percent and 6.28 percent of the total soil erosion amount relative to the runoff sediment method respectively, and the method can more accurately quantify the soil erosion amount of the erosion slope; and secondly, the method adopts multi-view observation to solve the problem of missing measurement caused by the fact that the laser line at the bottom of the channel cannot be projected in place during the observation of the laser scanner, and realizes almost full-coverage digital image acquisition.
Example four
The utility model provides a soil erosion surface morphology DEM construction equipment in rainfall process, includes the image acquisition module:
the image acquisition module is used for simultaneously acquiring a plurality of images of the soil erosion underlying surface to be modeled to obtain N initial image groups, wherein the size of each initial image in the initial image groups is the same;
in this embodiment, the image acquisition module may be an industrial camera, a camera, or the like capable of acquiring an image.
The image rain and fog removing module is used for respectively carrying out rain and fog removing processing on the N initial image groups to obtain N rain and fog removing images;
in this embodiment, the image defogging module may be an industrial personal computer installed near the image acquisition module, or may be a software module disposed on a remote platform for processing an image.
The image rain and fog removing module comprises a gray value sequence obtaining submodule, an initial seed point selecting submodule, a clustering submodule, a gray value updating submodule and an image reconstruction submodule;
the gray value obtaining submodule is used for obtaining the gray value of the ith pixel point of each image in the nth initial image group to obtain a gray value sequence, wherein I belongs to I, the I is the total number of the pixel points in each initial image in the nth initial image group, and the I is a positive integer;
the initial seed point selection submodule is used for taking the minimum value of the gray values in the gray value sequence as an initial seed point C 0 Taking the intermediate value of the gray values in the gray value sequence as an initial seed point C 1 Taking the maximum value of the gray values in the gray value sequence as an initial seed point C 2 ；
The clustering submodule is used for carrying out iterative clustering on all the gray values in the obtained gray value sequence by utilizing a K-means clustering algorithm to obtain three new clustering centers, namely a new clustering center X 0 New cluster center X 1 And a new cluster center X 2 New cluster center X 1 Greater than new cluster center X 0 And is smaller than the new cluster center X 2 ；
The gray value updating submodule is used for updating a new clustering center X 1 The gray value of the ith pixel point after rain and fog removal is taken as the gray value of the ith pixel point after rain and fog removal;
the image reconstruction submodule is used for collecting the I rain and fog removed gray values to obtain an nth rain and fog removed image;
optionally, when the clustering submodule performs iterative clustering on all gray values in the obtained gray value sequence by using a K-means clustering algorithm, the iteration stop condition is that a difference between clustering centers obtained after two adjacent iterations is less than or equal to 10 -6 Or stack ofThe number of generations is more than 2000, and three new clustering centers are obtained.
The three-dimensional data acquisition module is used for reconstructing the N rain and fog removing images by using a three-dimensional reconstruction method to acquire three-dimensional point cloud data of a soil erosion underlying surface to be modeled;
and the modeling module is used for interpolating the three-dimensional point cloud data of the soil erosion underlying surface to be modeled to obtain the DEM of the soil erosion underlying surface.
In this embodiment, the three-dimensional data obtaining module and the modeling module may be both disposed on a host in a control room near the site, or may be disposed on a remote cloud platform.
EXAMPLE five
The utility model provides a rainfall in-process soil erosion earth's surface form observation device, includes reference model acquisition module, observation model acquisition module, accumulative total erosion amount acquisition module and soil erosion amount acquisition module:
the reference model obtaining module is used for collecting the DEM of the soil erosion underlying surface to be observed before rainfall and taking the model as a reference for erosion amount calculation;
the observation model obtaining module is further used for respectively obtaining the DEM of the soil erosion underlying surface at a plurality of observation moments in the rainfall process by adopting the DEM constructing device of the soil erosion surface morphology in the rainfall process in the embodiment IV;
the accumulated erosion amount obtaining module is used for obtaining the difference between the DEM of the soil erosion underlying surface at each observation moment and the obtained volume of the DEM of the soil erosion underlying surface to be observed before rainfall to obtain the accumulated soil erosion amount at each observation moment;
the soil erosion amount obtaining module is used for obtaining the difference value between the accumulated soil erosion amounts at the adjacent moments and obtaining the soil erosion amount between the adjacent observation moments.
In this embodiment, all the modules in the soil erosion surface morphology observation device in the rainfall process may be arranged on a host in a control room near the site, or may be arranged on a remote cloud platform.
Claims (6)
1. A method for constructing a soil erosion surface morphology DEM in a rainfall process is used for obtaining the DEM of a soil erosion underlying surface in the rainfall process, and is characterized by being executed according to the following steps:
the method comprises the following steps that 1, N cameras are arranged right above a soil erosion underlying surface to be modeled, wherein N is an integer larger than 1, each camera is used for simultaneously collecting a plurality of images of the soil erosion underlying surface to be modeled to obtain N initial image groups, and the size of each initial image in each initial image group is the same;
step 2, carrying out rain and fog removing treatment on the N initial image groups respectively to obtain N rain and fog removing images:
wherein, carry out the fog processing of removing rain to nth initial image group, obtain nth and remove the rain fog image, specifically include:
step 2.1, obtaining the gray value of the ith pixel point of each initial image in the nth initial image group to obtain a gray value sequence, wherein I belongs to I, I is the total number of the pixel points in each initial image in the nth initial image group, and I is a positive integer;
step 2.2, taking the minimum value of the gray values in the gray value sequence as an initial seed point C 0 Taking the intermediate value of the gray values in the gray value sequence as the initial seed point C 1 Taking the maximum value of the gray values in the gray value sequence as an initial seed point C 2 ；
Step 2.3, performing iterative clustering on all gray values in the gray value sequence obtained in the step 2.1 by utilizing a K-means clustering algorithm to obtain three new clustering centers X 0 New cluster center X 1 And a new cluster center X 2 The new cluster center X 1 Greater than new cluster center X 0 And is smaller than the new cluster center X 2 ；
Step 2.4, new clustering center X 1 The gray value of the ith pixel point after rain and fog removal is taken as the gray value of the ith pixel point after rain and fog removal;
step 2.5, repeating the step 2.1 to the step 2.5 until I rain and fog removed gray values are obtained;
step 2.6, collecting the I rain and fog removed gray values to obtain an nth rain and fog removed image;
step 3, reconstructing the N rain and fog removing images by using a three-dimensional reconstruction method to obtain three-dimensional point cloud data of a soil erosion underlying surface to be modeled;
and 4, interpolating the three-dimensional point cloud data of the soil erosion underlying surface to be modeled to obtain the DEM of the soil erosion underlying surface.
2. The method for constructing the DEM of the soil erosion surface morphology in the rainfall process according to claim 1, wherein when the K-means clustering algorithm is used for iteratively clustering all gray values in the gray value sequence obtained in the step 2.1 in the step 2.3, the iteration stop condition is that the difference value between clustering centers obtained after two adjacent iterations is less than or equal to 10 -6 Or the iteration times are more than 2000, and three new clustering centers are obtained.
3. A soil erosion surface morphology observation method in a rainfall process is used for obtaining soil erosion amount between adjacent observation moments in the rainfall process, and is characterized by being executed according to the following steps:
a, collecting a DEM of a soil erosion underlying surface to be observed before rainfall;
step B, respectively obtaining the DEM of the soil erosion underlying surface at a plurality of observation moments in the rainfall process by adopting the method for constructing the soil erosion surface morphology DEM in the rainfall process according to any one of claims 1-2;
step C, obtaining the volume difference between the DEM of the soil erosion underlying surface at each observation time and the DEM of the soil erosion underlying surface to be observed before rainfall, which is obtained in the step A, and obtaining the accumulated soil erosion amount at each observation time;
and D, obtaining a difference value between the accumulated soil erosion amounts at the adjacent moments, and obtaining the soil erosion amount between the adjacent observation moments.
4. The utility model provides a soil erosion surface morphology DEM construction equipment in rainfall process, includes the image acquisition module:
the image acquisition module is used for simultaneously acquiring a plurality of images of the soil erosion underlying surface to be modeled to obtain N initial image groups, wherein the size of each initial image in the initial image groups is the same;
the system is characterized by further comprising an image rain and fog removing module, a three-dimensional data obtaining module and a modeling module;
the image rain and fog removing module is used for respectively removing rain and fog from the N initial image groups to obtain N rain and fog removing images;
the image rain and fog removing module comprises a gray value sequence obtaining submodule, an initial seed point selecting submodule, a clustering submodule, a gray value updating submodule and an image reconstruction submodule;
the gray value obtaining submodule is used for obtaining the gray value of the ith pixel point of each image in the nth initial image group to obtain a gray value sequence, I belongs to I, I is the total number of the pixel points in each initial image in the nth initial image group, and I is a positive integer;
the initial seed point selection submodule is used for taking the minimum value of the gray values in the gray value sequence as an initial seed point C 0 Taking the intermediate value of the gray values in the gray value sequence as the initial seed point C 1 Taking the maximum value of the gray values in the gray value sequence as an initial seed point C 2 ；
The clustering submodule is used for carrying out iterative clustering on all the gray values in the obtained gray value sequence by utilizing a K-means clustering algorithm to obtain three new clustering centers, namely a new clustering center X 0 New cluster center X 1 And a new clustering center X 2 The new cluster center X 1 Greater than new cluster center X 0 And is smaller than the new cluster center X 2 ；
The grey value updating submodule is used for updating a new clustering center X 1 The gray value of the ith pixel point after rain and fog removal is taken as the gray value of the ith pixel point after rain and fog removal;
the image reconstruction submodule is used for collecting the I gray values subjected to rain and fog removal to obtain an nth rain and fog removal image;
the three-dimensional data acquisition module is used for reconstructing the N rain and fog removing images by using a three-dimensional reconstruction method to acquire three-dimensional point cloud data of a soil erosion underlying surface to be modeled;
and the modeling module is used for interpolating the three-dimensional point cloud data of the soil erosion underlying surface to be modeled to obtain the DEM of the soil erosion underlying surface.
5. The soil erosion surface morphology DEM construction apparatus as claimed in claim 4, wherein when iterative clustering is performed on all gray values in the obtained gray value sequence by using K-means clustering algorithm in the clustering submodule, the iteration stop condition is that the difference between clustering centers obtained after two adjacent iterations is less than or equal to 10 -6 Or the iteration times are more than 2000, and three new clustering centers are obtained.
6. The utility model provides a rainfall in-process soil erosion earth's surface form observation device which characterized in that includes benchmark model acquisition module, observation model acquisition module, accumulative total erosion volume acquisition module and soil erosion volume acquisition module:
the reference model acquisition module is used for acquiring a DEM of the soil erosion underlying surface to be observed before rainfall;
the observation model obtaining module is also used for respectively obtaining the DEM of the soil erosion underlying surface at a plurality of observation moments in the rainfall process by adopting the soil erosion surface form DEM constructing device in the rainfall process according to any one of claims 4 to 5;
the accumulated erosion amount obtaining module is used for obtaining the difference between the DEM of the soil erosion underlying surface at each observation moment and the obtained volume of the DEM of the soil erosion underlying surface to be observed before rainfall to obtain the accumulated soil erosion amount at each observation moment;
the soil erosion amount obtaining module is used for obtaining a difference value between the accumulated soil erosion amounts at adjacent moments and obtaining the soil erosion amount between the adjacent observation moments.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910515164.0A CN110363857B (en) | 2019-06-14 | 2019-06-14 | Method and device for constructing and observing soil erosion surface morphology DEM in rainfall process |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910515164.0A CN110363857B (en) | 2019-06-14 | 2019-06-14 | Method and device for constructing and observing soil erosion surface morphology DEM in rainfall process |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110363857A CN110363857A (en) | 2019-10-22 |
| CN110363857B true CN110363857B (en) | 2022-09-20 |
Family
ID=68216136
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910515164.0A Active CN110363857B (en) | 2019-06-14 | 2019-06-14 | Method and device for constructing and observing soil erosion surface morphology DEM in rainfall process |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110363857B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113740332A (en) * | 2021-07-28 | 2021-12-03 | 南昌工程学院 | Water and soil conservation test system and method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015039375A1 (en) * | 2013-09-17 | 2015-03-26 | 中国科学院深圳先进技术研究院 | Method and system for automatically optimizing quality of point cloud data |
| CN105205855A (en) * | 2015-09-14 | 2015-12-30 | 西北农林科技大学 | Method for measuring water and soil loss conditions |
| CN108180897A (en) * | 2018-01-06 | 2018-06-19 | 中国科学院、水利部成都山地灾害与环境研究所 | Sloping upland soil water reservoir capacity rate Method of fast estimating |
-
2019
- 2019-06-14 CN CN201910515164.0A patent/CN110363857B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015039375A1 (en) * | 2013-09-17 | 2015-03-26 | 中国科学院深圳先进技术研究院 | Method and system for automatically optimizing quality of point cloud data |
| CN105205855A (en) * | 2015-09-14 | 2015-12-30 | 西北农林科技大学 | Method for measuring water and soil loss conditions |
| CN108180897A (en) * | 2018-01-06 | 2018-06-19 | 中国科学院、水利部成都山地灾害与环境研究所 | Sloping upland soil water reservoir capacity rate Method of fast estimating |
Non-Patent Citations (1)
| Title |
|---|
| 基于三维激光扫描技术的花岗岩风化土体侵蚀表面特征研究;徐加盼等;《水土保持学报》;20160430;第30卷(第02期);第14-19页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110363857A (en) | 2019-10-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hodge et al. | In situ characterization of grain‐scale fluvial morphology using Terrestrial Laser Scanning | |
| Stojic et al. | The assessment of sediment transport rates by automated digital photogrammetry | |
| Woodget et al. | From manned to unmanned aircraft: Adapting airborne particle size mapping methodologies to the characteristics of sUAS and SfM | |
| CN110390687B (en) | Three-dimensional laser scanning-based river channel erosion and deposition measuring method | |
| CN109100719B (en) | Terrain map joint mapping method based on satellite-borne SAR (synthetic aperture radar) image and optical image | |
| Cândido et al. | High-resolution monitoring of diffuse (sheet or interrill) erosion using structure-from-motion | |
| CN105205855B (en) | A kind of soil erosion situation method for measuring | |
| Guo et al. | Digital close range photogrammetry for the study of rill development at flume scale | |
| Di Stefano et al. | Testing the use of an image‐based technique to measure gully erosion at Sparacia experimental area | |
| Báčová et al. | A GIS method for volumetric assessments of erosion rills from digital surface models | |
| CN111507375B (en) | Urban waterlogging risk rapid assessment method and system | |
| Mali et al. | Assessing the accuracy of high-resolution topographic data generated using freely available packages based on SfM-MVS approach | |
| CN110363857B (en) | Method and device for constructing and observing soil erosion surface morphology DEM in rainfall process | |
| Detert et al. | On the way to airborne gravelometry based on 3D spatial data derived from images | |
| Sun et al. | Characterizing erosion processes on a convex slope based on 3D reconstruction method | |
| CN108956392B (en) | Unmanned aerial vehicle identification method for tidal flat sediment types | |
| Rodríguez et al. | Potential of mobile application based on structure from motion (SfM) photogrammetry to monitor slope fast erosion by runoff water | |
| Merel et al. | The monitoring of soil surface development using analytical photogrammetry | |
| CN110555538B (en) | Wind power plant wind speed prediction method and prediction system | |
| He et al. | Quantifying spatial distribution of interrill and rill erosion in a loess at different slopes using structure from motion (SfM) photogrammetry | |
| CN111127613B (en) | Image sequence three-dimensional reconstruction method and system based on scanning electron microscope | |
| CN113538668A (en) | Geological model construction method, geological model construction system, storage medium and electronic equipment | |
| CN114998530B (en) | Water body monitoring method and device based on real-scene three-dimensional terrain | |
| Moritani et al. | Using digital photogrammetry to monitor soil erosion under conditions of simulated rainfall and wind | |
| Marino et al. | Measuring rock slope damage on rubble mound breakwater through digital photogrammetry |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |