KR101406010B1 - Classification method for levee composition using lidar data and orthophotograph - Google Patents

Abstract

The present invention relates to a method for classifying levee components using LiDAR data and orthoimages. More particularly, provided is a method for precisely classifying a top plate, a rear berm, and an erosion area which are the flat area of a levee by determining whether each berm candidate plate is a berm or the erosion area by extracting the top plate, the berm candidate plate, and an erosion area plate using the light detection and ranging (LiDAR) material and applying the extracted result to the orthogimages.

Description

BACKGROUND OF THE INVENTION Field of the Invention [0001] The present invention relates to a method of classifying elements of an embankment using a Lada data and an orthoimage image,

The present invention relates to a method of classifying a levee component using a Lada data and an orthoimage, and more particularly, to a method of classifying a levee floor plate, a rear jaw candidate plate, And to apply multispectral orthoimages to the multispectral orthoimages to discriminate whether each of the dorsal plate candidates is a ducal or an erosion area, thereby more precisely classifying the dunnage, dorsal and erosion areas as the flat areas of the embankment.

Water is an indispensable resource that is indispensable to the birth of all life, including mankind. As many countries in the world are expected to become water-scarce countries in the future, 20% of the world's population (about 1.1 billion people) can not drink clean water due to the failure of water management policies.

Especially, rainfall is biased year by year, seasonally and regionally, and seasonal river flow fluctuation is very characteristic due to the characteristics of mountainous terrain where slopes are urgent, and natural conditions are very unfavorable to water resource management. In the middle of the year, there is a problem of water supply due to shortage of river water. Unlike foreign countries, which have steady rainfall during the year, floods and droughts are frequent, and the necessity of river management will increase continuously.

Therefore, in order to properly perform the role of the river, the accurate status of the river and the management technology are increasing in importance, and the precise river surveying technology that provides the basic data of the river status is becoming important. Future river surveillance techniques will be more demanding for accuracy, speed and database, and the technology is expected to progress.

The embankment is a representative river system installed in the river basin, and is a workpiece installed along the river to maintain smooth communication and protect the flowing water. The structure of the embankment consists largely of the embankment site including the excavation, the front slope, the back slope, the dike floor, etc., and it can be used to adjust the shape if necessary. Means a line connected in the horizontal direction. Thus, the structure of the embankment is divided into several types, and it is necessary to accurately grasp the structure of the embankment so that it can be prepared in advance for disasters such as flood. However, in order to accurately measure and express the structure of the bank, there is a problem that it requires a lot of manpower, time and expenses, and it is difficult to acquire information related to the structure of the bank in real time.

Korean Unexamined Patent Publication [10-2001-0000443] discloses a building extraction system and method by aerial surveying image and radar data fusion, and a recording medium storing the program source.

Korean Published Patent [10-2001-0000443]

SUMMARY OF THE INVENTION Accordingly, the present invention has been made to solve the above-mentioned problems, and it is an object of the present invention to provide a dewar and erosion area, which is difficult to distinguish using elevation difference with a dike floor, , A method of classifying the elements of a bank using Lada data and orthoimage which classify the dorsal and erosion areas more precisely.

According to another aspect of the present invention, there is provided a method for classifying elements of a bank using ladder data and orthoimage, the method comprising the steps of: According to the mapping program, a light detection and ranging (LiDAR) point group data is received, and a binary image is generated to generate a binary image representing flat plates corresponding to a predetermined inclination range using altitude information and tilt information Generating step S10; A dike floor plate extracting step (S20) of extracting a top plate having a higher average altitude value than adjacent planes using the altitude information on the binary plane; A step of extracting a rear jaw candidate plate (S30) for extracting a rear jaw candidate plate of the embankment including a difference in average altitude value from the dike floor plate using the altitude information on the binary image; An erosion zone plate extraction step (S40) of extracting an erosion zone plate of a bank which does not belong to the dike floor plate and the rear jaw candidate plate using the altitude information on the binary image; And a step S50 of sorting the subpopulation area and the erosion area by using the altitude information and the clustering information to classify whether the hypothesis candidate plate is a duchy of a bank or an erosion area of a bank in a multispectral orthoimage image.

In addition, the step of generating the binary image (S10) may include a step (S11) of generating a digital surface model (DSM) which is a grid structure using the interpolation technique; A noise removing step (S12) of removing noise by using a noise removing technique on the digital surface model; A slope grid generation step (S13) of generating a grid of slope gradients using altitude values of pixels constituting the digital surface model; And a binary image extracting step (S14) of extracting a binary image extracting pixels having an inclination corresponding to a predetermined inclination range from the ladder inclination grid; And a control unit.

In addition, a linear interpolation method is used as the interpolation method of the digital surface model generation step (S11).

In addition, median filtering is used as the noise removal technique of the noise removal step S12.

In addition, the predetermined gradient range of the binary image generation step (S14) is 0 to 9 degrees.

Also, the predetermined range of the rear jaw candidate plate extracting step (S30) is 2.5 to 3.5 m.

The step S50 includes a band input step S51 for inputting the values of Red, Green, Blue and NIR (Near Infra-Red) bands of the multispectral orthoimage image to the binary image, ; An object classifying step (S52) of classifying objects using clustering algorithms for points located on the rear jaw candidate plate of the embankment; And a step S53 for dividing the ridge region into an erosion region of the embankment by using the clustering information and the predetermined criterion.

In addition, the object of the object classification step S52 is a plurality of objects selected from pavement road, unpaved road, vegetation and soil.

In addition, ISODATA (Interactive Self-Organizing Data Analysis) clustering algorithm is used as the clustering algorithm in the object classification step S52.

In addition, a predetermined criterion of the rear-jaw and erosion zone classifying step S53 is that when the rear jaw candidate plate includes road clusters, the plate is divided into the rear jaw of the embankment, and when the rear jaw candidate plate comprises only vegetation and clay clusters The plate is characterized by the erosion zone of the embankment.

According to the method of classifying the elements of the embankment using the Lada data and the orthoimage in accordance with the embodiment of the present invention, it is possible to classify the ducks and the erosion area of the embankment, which are difficult to classify using the difference in elevation value, Can be accurately classified using a multi-spectral orthoimage.

Also, it is possible to extract a dike floor plate with higher accuracy by using a linear interpolation method and median filtering based on the generated grid.

In addition, there is an effect of improving the reliability by applying the average slope of the dike floor of river embankment established by the Ministry of Land, Transport and Maritime Affairs.

Also, there is an effect that the rear jaw and erosion area of the bank can be accurately discriminated by using the object information of the rear jaw candidate plate of the bank.

In addition, it is possible to extract more accurate object information by using an ISODATA (Iterative Self-Organizing Data Analysis) clustering algorithm.

In addition, it is possible to classify dike floor, back jaw and naturally occurring erosion area composed of flat areas located on the bank, so that stability evaluation and efficient management of the bank can be effected.

Figure 1 is a topographic view of a generally distinct stream.
FIG. 2 through FIG. 4 are flowcharts illustrating a method line mapping method using a Lada data and an orthoimage image according to an embodiment of the present invention.
5 is a view showing a ladder grid.
Figure 6 is a view showing a grid of ladder inclination.
7 is a view showing a binary image.
8 is a view showing a result of extracting a dike floor plate on a binary image and a binary image.
9 is a view showing a result of inputting a band of a multi-spectral orthoimage image to a Lada point cloud data of a concrete type bank.
FIG. 10 is a diagram showing a classification result obtained by applying an ISODATA clustering algorithm based on the result of FIG. 9; FIG.
Fig. 11 is a view showing the result of dividing into the dorsal ridges and erosion zones of the embankment based on the result of Fig. 10; Fig.
12 is a view showing a result of inputting a band of a multi-spectral orthoimage image to a Lada point cloud data of a green embankment.
FIG. 13 is a diagram showing a result obtained by applying the ISODATA clustering algorithm based on the results of FIG. 12; FIG.
FIG. 14 is a view showing the result of dividing into the rear jaw and erosion area of the embankment based on the result of FIG. 13; FIG.

Hereinafter, the present invention will be described in more detail with reference to the drawings. It is to be noted that the same elements among the drawings are denoted by the same reference numerals whenever possible. In the following description, well-known functions or constructions are not described in detail since they would obscure the invention in unnecessary detail.

FIG. 1 is a topographical view of a generally divided stream, FIGS. 2 to 4 are flowcharts illustrating a method line mapping method using a Lada data and an orthoimage image according to an embodiment of the present invention, FIG. 7 is a view showing a binary image, FIG. 8 is a view showing a result of extracting a dike floor plate on a binary image and a binary image, FIG. 9 is a view showing a concrete image of a concrete FIG. 10 is a view showing a result of classifying by applying the ISODATA clustering algorithm based on the result of FIG. 9, and FIG. 11 Fig. 12 is a view showing the result of dividing into the ridge of the embankment and the erosion area based on the result of Fig. 10, and Fig. FIG. 13 is a diagram showing a classification result obtained by applying an ISODATA clustering algorithm based on the result of FIG. 12, and FIG. 14 is a graph showing a result of classification based on the result of FIG. The back of the embankment and the erosion area.

As shown in FIG. 1, the embankment is an artificial structure that is installed along the river to prevent the overflow of flood water from being overflowed. The embankment is composed of various elements such as a dike floor, a back ridge, a front slope, .

Among the various components of the embankment, the dike floor and the rear jaw are made up of flat plates, and the average height difference between the dike floor and the back jaw is about 3m. Because of the need for continuous maintenance and reinforcement to maintain the effective management and stability of the embankment, the flat areas of the dike floor and the mapping of the duck area are very important for effective management of the embankment. In Fig. 1, the dike floor of the embankment is a flat plate existing on the top of the embankment, and is used as a road or the like due to the surrounding roads and the like. The backrest of the embankment is a flat plate provided on the rear slope of the embankment And the erosion area of the embankment is defined as a flat plate that is naturally formed on the front slope or back slope of the embankment due to continuous erosion. The dike floor of the embankment is classified by calculating the altitude difference with the plate such as the dorsal surface and the erosion area due to the high altitude value compared to the dorsal surface or the erosion area, whereas the erosion area and the ducus are formed at similar altitude , It is difficult to distinguish the difference in the method of calculating the altitude difference.

According to the embodiment of the present invention, the method of classifying the elements of the embankment using the Lada data and the orthoimage is based on the following facts: the dug area of the embankment (mainly used as plates here) Classification.

Basis 1. The dorsal plate of the embankment must contain a concrete / asphalt road area, since the dorsum of the embankment is mainly used as a passage for the hydraulic work.

Grounds 2. Since the erosion area located on the front slope or the back slope of the bank is naturally occurring, most elements of the erosion area plate are composed of natural soil or vegetation.

Basis 3. Generally, in river design criteria since 2003, the levee no longer has a back jaw, because of the possibility that rainwater penetrates into the bottom of the levee and penetrates into the levee of the levee, which may reduce the stability of the levee.

Because of the above three reasons, it is not possible to classify the erosion area and the duck area of the embankment only by calculation of the difference value of the elevation value. Using the spectral information, the main components of the erosion area and the duck area candidate plate are identified, It is necessary to classify the dorsal area.

As shown in FIG. 2, according to an embodiment of the present invention, a method of classifying elements of a bank using ladder data and orthoimage is classified into a banking mapping program, which is implemented by a program executed by an arithmetic processing means including a computer, A dike floor plate extracting step S30, an erosion area plate extracting step S40, and a back jaw and erosion area sorting step S50 do.

Lada data has the advantage of providing 3D information of the terrain, but it has a disadvantage that the horizontal accuracy is relatively low compared to the image data and it is composed of point cloud data. Since the orthoimage data is composed of pixels that have higher horizontal accuracy and are connected unbroken than the Lada data, the edge information extracted from the orthoimage data is more accurate than the lines composed of Lada points It has the advantage that it can be described accurately.

The binary image generation step S 10 receives a light detection and ranging (LiDAR) point cloud data and generates a binary image representing flat plates corresponding to a predetermined gradient range using altitude information and gradient information do.

3, the digital image generation step S10 includes a digital surface model generation step S11, a noise removal step S12, a slope grid generation step S13, and a binary image extraction step S13. A binary image extracting step (S14); And a control unit.

In the digital surface model generation step S11, a RIDO digital surface model (DSM) is generated using the RDA point group data by interpolation. At this time, the linear interpolation method may be used as the interpolation method of the digital surface model generation step (S11).

In other words, the embankment has a characteristic that the altitude suddenly changes from the ground to the front slope and the rear slope surface, so that the Lada point cloud data are converted into the grid data (grid grid) using the linear interpolation method.

5, the value of each pixel represents the altitude value of the corresponding region. The embankment is composed of steeply sloping elements such as the front slope and the back slope, and sloping elements such as the dike parquets.

The noise removal step (S12) removes the noise using the noise cancellation technique on the digital surface model. In this case, median filtering may be used as the noise removal technique of the noise removal step S12.

The slope grid generation step S13 generates a slope grid using the elevation values of the pixels constituting the digital surface model.

In order to effectively represent the erosion areas of the dike floor, back jaws and bank, the ladder grid is converted to a ladder tilt grid, as shown in Fig. In the slope grid, each pixel is a slope calculated considering the difference between the altitude of the surrounding central pixel and the highest altitude value among the eight surrounding pixels.

A pixel having a high inclination in the slope grid of FIG. 6 indicates a front slope or a back slope of a bank, an outer wall of a building, and a pixel having a low slope indicates a dike floor, a back wall, and an erosion area of a bank.

A binary image extracting step (S14) for extracting a binary image extracts pixels having an inclination corresponding to a predetermined inclination range from the ladder inclination grid. The extracted binary images represent flat plates. In this case, the predetermined gradient range of the binary image generation step S14 may be 0 to 9 degrees. This is because, when a binary image representing flat plates is generated by applying the average slope of the dike floor and the back jaw of the river bank defined by the Ministry of Land, Transport and Maritime Affairs, as shown in Fig. 7, the dike floor plate, It can be seen that several flat plates are extracted.

In other words, it is possible to generate more accurate dike floor plates by using the linear interpolation method and median filtering based on the generation of the slope grid of Rada.

As shown in FIG. 8, the boundary between the ground and the front slope and the back slope of the bank is determined using a vector-based method and is called a bank boundary (red line). Here, the vector-based method can be confirmed by the Korean registered patent [10-1078238], and a detailed explanation will be omitted.

After defining the flat plates included in the bank boundary in Fig. 8 (left drawing in Fig. 8), plates having a higher average altitude value than the adjacent plates are selected (right drawing in Fig. 8) .

As shown in the right drawing of FIG. 8, after defining the plate having the highest altitude value as the dike floor, it is necessary to judge whether the low altitude plate is the ditch of the bank or the erosion area of the bank.

The dike floor plate extracting step (S20) extracts a top plate having a higher average altitude value than adjacent plates using the altitude information on the binary image.

In the step of extracting the rear jaw plate (S30), the rear jaw candidate plate including the difference of the average altitude values from the dike floor plate within a certain range is extracted using the altitude information on the binary image. At this time, the predetermined range of the rear jaw candidate plate extracting step S30 may be 2.5 to 3.5 m.

The erosion area plate extraction step (S40) extracts the erosion area plate of the bank that does not belong to the dike floor plate and the rear jukebox plate using the altitude information on the binocular.

In the step S50 of classifying the rear and top erosion regions, whether the rear projection plate of the multispectral orthoimage is the back of the embankment or the erosion area of the embankment is classified using the altitude information and the clustering information.

As shown in FIG. 4, the step S50 may include a band input step S51, an object classification step S52, and a rear jaw and erosion zone classification step S53. have.

In other words, classify the ditches and erosion areas of the embankment on a bank of different materials.

The levee treated in Figs. 9 to 11 is a levee in which the front slope area is formed of concrete or a stone block (here, it is defined as a concrete levee), and the levee treated in Figs. 12 to 14 is a levee Green levee).

The band input step S51 inputs the values of the red, green, blue and NIR (Near Infra-Red) bands of the multispectral orthoimage image to the binary image. In other words, the values of Red, Green, Blue and Near Infra-Red (bands) of the multispectral orthoimage image are input to the Lada point cloud data (see FIGS. 9 and 12)

In Fig. 9, it is necessary to determine whether plates (blues) having a certain altitude difference from the dike floor of the embankment are the rear jaw of the embankment or the erosion area of the embankment.

For example, plates with a mean height difference of about 3 m (2.5 to 3.5 m) from the dike floorboards of the bank are extracted and defined as the dorsal candidate plate of the embankment (blue in Fig. 9 and green in Fig. 12) . Plates with a difference in average altitude value of about 3 m or less can be defined first by the erosion area of the embankment (yellow in Fig. 9 and yellow in Fig. 12).

The object classification step S52 classifies the objects using the clustering algorithm of the points located on the dorsal candidate plate of the embankment. At this time, the object of the object sorting step S52 may be a plurality of objects selected from pavement roads, unpaved roads, vegetation, and soil. In addition, ISODATA (Iterative Self Organizing Data Analysis) clustering algorithm may be used as the clustering algorithm in the object classification step S52. We apply the ISODATA clustering algorithm, which is one of the clustering algorithms, to each of the Lada point cloud data to classify the points of the Lada point cloud data located at the dorsal candidate plate (the blue color in Fig. 9 and the green dots in Fig. 12) .

The step S53 of dividing the ridge and erosion zone is divided into the erosion zones of the dike candidate plate and the embankment using the altitude difference, clustering information, and predetermined criteria. At this time, the predetermined criterion in the step S53 for sorting the rear jaw and the erosion area is that if the rear jaw candidate plate includes the road clusters, the plate is divided into the jaws of the embankment, and if the rear jaw candidate plate comprises only the vegetation and the soil clusters The plate may be characterized as being divided into erosion areas on the bank.

When a cluster including the concrete road (red) is included as shown in FIG. 10 when the ISODATA clustering algorithm is applied to the points located on the dorsal candidate plate of the embankment, the plate is defined as the ditch of the embankment.

By the stream design standards established since 2003, the levees no longer fall back. However, if the dike's dorsal candidate plate contains concrete or asphalt road clusters, it is not constructed recently but is defined as the dorsal of the existing remaining dike. Therefore, as shown in FIG. 10, since a part of the bank sidewall candidate plate is constituted by concrete road clusters, the plate is defined as the sidewall of the bank. Thus, the embankment area of Figure 11 is divided into the erosion area (yellow) of some of the dykes and the dorsal area (red) of many of the embankments.

The green levee can be classified into a ridge candidate plate (green) and an erosion area (yellow) of a bank using a difference in altitude value, as shown in Fig.

FIG. 13 shows an example of the objects of the point cloud groups positioned on the dorsal candidate plate of the embankment by applying the ISODATA clustering technique to the Raida point group data located on the dorsal candidate plate of the embankment of FIG. 12 (FIG. 13). In Fig. 13, since the dorsal candidate plate of the embankment is composed of the soil (yellow) and vegetation (green) regions, the dorsal candidate plate of the embankment defined in Fig. 12 is defined as a simple embankment erosion area . Therefore, as can be seen from Fig. 14, it can be seen that all the flat plates shown in Fig. 12 are classified as the erosion zone of the bank.

In other words, more accurate object information can be extracted using ISODATA (Iterative Self-Organizing Data Analysis) clustering algorithm, and it is possible to accurately distinguish the dorsal and erosion areas of the embankment using the object information of the dike candidate plate.

According to the method of classifying the elements of the embankment using the Lada data and the orthoimage in accordance with the embodiment of the present invention, it is possible to classify the ducks and the erosion area of the embankment, which are difficult to classify using the difference in elevation value, Can be accurately classified using a multispectral orthoimage. In addition, it is possible to classify the dike floor, the back jaw and the naturally occurring erosion area composed of the flat area located on the bank, so as to evaluate the stability of the bank and to manage efficiently.

It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.

S10: Binary image generation step
S11: Digital surface model generation step
S12: Noise removal step
S13: Creation of slope grid
S14: Binary image extraction step
S20: Dike floor plate extraction step
S30: Rat chin plate extraction step
S40: Erosion local plate extraction step
S50: Classification of the occipital area and the erosion area
S51: Band input step
S52: Object classification step
S53: Tear and erosion area classification step

Claims (10)

By the embankment mapping program embodied in the form of a program executed by an arithmetic processing means including a computer,
(S10) for receiving a light detection and ranging (LiDAR) point cloud data and generating a binary image representing flat plates corresponding to a predetermined inclination range using altitude information and tilt information, ;
A dike floor plate extracting step (S20) of extracting a top plate having a higher average altitude value than adjacent planes using the altitude information on the binary plane;
A step of extracting a rear jaw candidate plate (S30) for extracting a rear jaw candidate plate of the embankment including a difference in average altitude value from the dike floor plate using the altitude information on the binary image;
An erosion zone plate extraction step (S40) of extracting an erosion zone plate of a bank which does not belong to the dike floor plate and the rear jaw candidate plate using the altitude information on the binary image;
A step S50 of classifying the multi-spectral orthoimage using the altitude information and the clustering information to classify whether the dorsal surface candidate plate is the dorsal surface of the embankment or the erosion area of the embankment;
/ RTI >
The rear jaw and erosion area classification step S50
A band input step (S51) of inputting values of Red, Green, Blue and NIR (Near Infra-Red) bands of the multispectral orthoimage image to the binary image;
An object classifying step (S52) of classifying objects using clustering algorithms for points located on the rear jaw candidate plate of the embankment; And
(Step S53) of dividing the tear candidate plate and the erosion area of the embankment using the difference in altitude value, clustering information, and predetermined criteria;
The method comprising the steps of: a) classifying the elements of the embankment using the Lada data and the orthoimage.
The method according to claim 1,
The binary image generation step (S10)
A digital surface model generation step (S11) of generating a Rada digital surface model (DSM) as a grid structure using the interpolation technique;
A noise removing step (S12) of removing noise by using a noise removing technique on the digital surface model;
A slope grid generation step (S13) of generating a grid of slope gradients using altitude values of pixels constituting the digital surface model; And
A binary image extracting step (S14) of extracting a binary image extracting pixels having an inclination corresponding to a predetermined inclination range from the ladder inclination grid;
The method comprising the steps of: a) classifying the elements of the embankment using the Lada data and the orthoimage.
3. The method of claim 2,
As an interpolation technique of the digital surface model generation step (S11)
A method of classifying elements of a bank using ladder data and orthoimage, characterized by using a linear interpolation method.
3. The method of claim 2,
As the noise cancellation technique of the noise cancellation step S12
A method of classifying elements of a bank using lidar data and orthoimage, characterized by using median filtering.
3. The method of claim 2,
The predetermined gradient range of the binary image generation step (S14)
The method according to any one of claims 1 to 3,
The method according to claim 1,
A certain range of the rear jaw candidate plate extracting step (S30)
The method of claim 1, further comprising the steps of:
delete The method according to claim 1,
The object of the object classification step S52 is
Wherein the object is a plurality of objects selected from a pavement road, an unpaved road, vegetation, and soil.
The method according to claim 1,
As the clustering algorithm of the object classification step S51
ISODATA (Iterative Self-Organizing Data Analysis) clustering algorithm is used.
The method according to claim 1,
The predetermined criterion of the dorsal and erosion area classification step S53 is
If a road cluster is included in the rear jaw candidate plate, the plate is divided into a rear jaw of the embankment, and if the rear jaw candidate plate comprises only vegetation and clay clusters, the plate is divided into erosion areas of the embankment A Method of Line Mapping Using Data and Orthographic Image.

Images