KR100836196B1 - Method for precise topograhic information extraction of lidar using transmittance property - Google Patents

Abstract

An LIDAR(Light Detection and Ranging) topographic information extraction method using the transmittance property of vegetation is provided to form a digital elevation model in the forest region where topographic information is not extracted well, by classifying non-ground points in stages on the basis of elevation difference between a first response signal and the last response signal by using LIDAR point data. An LIDAR topographic information extraction method(1) using the transmittance property of vegetation is composed of an initial separating step for separating a singular response(33) repeatedly recorded in ASCII(American Standard Code for Information Interchange) type data outputted by storing binary data obtained by reflecting an LIDAR signal in the forest region, while separating the binary data into a first response and a second response(30), a first separating step for filtering non-ground points(73) by separating candidates(51) for ground points(71) and non-ground points that are responses of the crowns of trees, in the first and last responses(50), and a second separating step for filtering the candidates for ground points that are objects approaching the ground over the predetermined distance on the basis of the surface maximally approaching the ground(70).

Description

Method for Precise Topograhic Information Extraction of Lidar Using Transmittance Property

1 is a schematic block flow diagram according to a method for extracting Lidar terrain information using vegetation permeation characteristics of the present invention.

Figure 2 is a graph showing the first separation step according to the Lidar terrain information extraction method using the vegetation permeation characteristics of the present invention.

Figure 3 is a graph of the first separation step according to the Lidar terrain information extraction method using the vegetation permeation characteristics of the present invention.

4 is a flow chart schematically showing a method for extracting Lidar terrain information using vegetation permeation characteristics of the present invention.

5 is a graph according to the initial separation step of the Lidar terrain information extraction method using vegetation permeation characteristics according to an embodiment of the present invention.

6 is a view applied to the study area 1 in the Lidar terrain information extraction method using vegetation permeation characteristics according to an embodiment of the present invention.

7 is a view applied to the study area 2 in the Lidar terrain information extraction method using vegetation permeation characteristics according to an embodiment of the present invention.

       <Brief description of reference numerals for the main parts of the drawings>

1: Lidar Terrain Information Extraction Method Using Vegetation Transmission Characteristics

10: Lidar Source Data 30: Initial Separation Steps

31: pure initial response 33: single response

35: pure terminal response 50: primary separation stage

51: ground point candidate 53: non-ground point

70: secondary separation step 71: ground point

73: free ground point

The present invention relates to a method of extracting Lidar terrain information using vegetation permeation characteristics, and more particularly, in the vegetation region by using the ASCII type Lidar data which is stored in the forest area by reflecting the Lidar signal and separating it into an initial response and a late response. Lidar topographic information extraction method using vegetation permeation characteristics that extracts topographic information in forest areas where it is difficult to extract topographical information by using height difference of appearing early and late response signal.

In general, a digital elevation model (3D) is a three-dimensional coordinate used for the construction of a geographic information system (GIS), a digital terrain model (DTM) representing a terrain, and a DTD. (Digital Terrain Data), DTED (Digital Terrain Elevation Data), and the like.

In addition, Lidar fires laser beams and uses their reflections and absorptions to provide accurate and dense information about three-dimensional geospatial spaces, as well as laser signals in areas where vegetation such as forests exist. It is transmitted and reflected on the ground.

However, in order to construct the precise terrain data using the Lidar, filtering, which is a technique of separating only the signal reflected from the surface of the laser signal, has been developed based on the urban area, so that a method of converting to a lattice or using only the terminal reflected signal is used. As a result, interpolation method reduces the quality of precise data provided by Lidar, and in the case of extracting effort from the forest area, the separation rate is reduced by separating the ground points using only the terminal reflection signal, and the vertical interference factor Due to the tree planting and lower vegetation, there was a problem that the pure surface could not be represented accurately.

The present invention has been made to solve the above problems, and it is intended to extract high-precision topographic information from forest areas where it is difficult to extract topographic information based on the vegetation permeation characteristics of Lidar using Lidar point data. Lidar's vegetation permeation characteristics are suitable for effectively constructing precision high-altitude data showing pure land surface in forest terrain where vertical interference factors such as trees and lower vegetation exist. An object of the present invention is to provide a method of extracting Lidar terrain information using vegetation permeation characteristics that can increase the precision.

In order to achieve the object as described above, the present invention is to separate a single response that is duplicated in the ASCII format Lidar data stored by separating the binary data obtained by reflecting the Lidar signal in the forest area into an initial response and a late response Initial separation step; A first separation step of filtering the non-floor point by separating the non-ground point, which is the response of the ground point candidate and the crown portion, from the initial response and the late response; A second separation step of filtering a ground point candidate, which is a ground proximity object exceeding a predetermined distance based on a surface closest to the ground among the ground point candidates; It includes.

Here, the first step of separating may include setting a point close to each other between an elevation of the highest value of the initial response recorded in the water pipe and an elevation of the lowest value of the terminal response reflected in the sub-water area; Setting as a ground point candidate if the height difference between the point close to the initial response maximum is less than the height difference between the point close to the terminal response minimum; If the height difference between the point close to the initial response maximum is greater than the height difference between the point close to the terminal response lowest value, filtering the non-ground point; Characterized in that consists of.

And, the data of the height exceeding the ground point in the first separation step is characterized in that the filtering.

In addition, the secondary separation may include setting, as a reference plane, a plane composed of data closest to the ground among the ground point candidates; Classifying a ground point exceeding a predetermined distance from the reference plane as a non-ground point; Filtering a ground proximity proximate to the ground according to the filtering; Characterized in that it comprises a.

In addition, the number of ground points is adjusted according to the predetermined distance.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a schematic block flow diagram illustrating a Lidar terrain information extraction method using vegetation permeation characteristics of the present invention. As shown in the figure, Lidar terrain information extraction method using the vegetation transmission characteristics according to the present invention receives a signal reflected by transmitting a signal on the air, according to the stored data is Lidar original data 10, When the data is obtained, the binary data is separated into an initial response and a last response, and stored as ASCII data.

As described above, the Lidar original data 10 includes the binary data including an initial response and a terminal response in the form of an ASCII code.

Here, the received signals such as the initial response or the late response are recorded by changing the position of each received signal in a transmissive area such as a vegetation area, but in an area interspersed with artificial structures such as buildings or roads that are not transmissive. Or the same because of an artificial structure in which the terminal response is not transmitted.

The regional distribution characteristic of the single response (Singular Response, 33) in which the initial response and the late response are recorded at the same time is recorded in the vegetation area such as the safari area as well as the area where the artificial structure exists. 33) occurs, but is less than in areas where there are many artificial structures.

Therefore, Lidar provides information about the height value of the trees in the area as the difference between the elevation values represented by the initial response and the late response except for the single response in the forest area. To this end, the initial response and the final response are separated into a single response 33 which is recorded in duplicate with the pure initial response 31 and the pure late response 35 in the initial separation step 30.

In addition, when the initial response and terminal response of the Lidar original data 10 are classified into the pure initial response 31, the single response 33, and the pure final response 35 in the initial separation step 30, respectively. In the filtering process of the data, by removing the data for the single response 33 which is duplicated, the efficiency of processing the high-density Lidar data can be brought, Lidar terrain information using the vegetation transmission characteristics according to the present invention As a step before applying the extraction method, the single response 33 is first separated.

Then, from the pure initial response 31 and the pure late response 35 of the initial separation step 30 in which the single response 33 is separated, the ground point candidate 51 and the non-ground surface such as the water pipe part closest to the ground point. Point 53 is separated as the primary separation step.

Here, the primary separation step 50, such as removing the response of the water pipe part, separates the Lidar data recorded per fixed unit area into the water pipe part of the tree and the area near the ground reflected and recorded in the vicinity of the ground into respective groups. It means the process.

2 is a graph illustrating a first separation step according to the Lidar terrain information extraction method using the vegetation permeation characteristics of the present invention, Figure 3 is a first separation step according to the Lidar terrain information extraction method using the vegetation permeation characteristics of the present invention Is a graph. As shown in the drawings, in the first separation step of the Lidar terrain information extraction method using the vegetation permeation characteristics according to the present invention, the Lidar data recorded per fixed unit area is reflected by the water pipe part of the tree and the ground recorded near the ground This is done to separate point candidates.

Here, in order to separate the ground point candidate, it plays a role of extracting data that can be confirmed as a signal reflected from the water pipe part, and the area to be executed is 10m x 10m, and the area indicates the vegetation transmission characteristics according to the present invention. A space in which 100 to 150 received signals of Lidar data used in the used Lidar terrain information extraction method (1) is recorded, which is preferably used in terms of data processing and time since the initial response and the late response are properly distributed.

As shown in FIG. 2, the first response (FR) and the last response (LR) show characteristics of the vertical distribution over the forest area, and the initial response (FR) is reflected from the water pipe and recorded. The terminal response LR is reflected and recorded in a portion of the water pipe and the middle portion, but most are reflected and recorded in the ground.

Based on this vertical distribution, the primary separation is performed. Lidar data recorded in a certain area is recorded at the elevation (Z) of the highest recorded initial response (FR) reflected by the water pipe and reflected below the water pipe layer. The lowest end of the final response (LR) is determined as the closest part of the data to determine the i-th data.

The following formula represents the primary separation step according to the invention.

Here, the first stage of separation is based on the elevation data in the forest area where the trees exist, so that the first stage is separated into the non-ground point, which is the data corresponding to the crown, and the ground point candidate, which is the data distributed near the ground point. It is not separated by ground points alone.

Among the ground point candidates 51 separated in the first separation step 50, a plane according to Equation [3] constituted by most ground point candidates is set as a reference plane, and the equation is expressed by Equation [4] from the reference plane. Ground points exceeding a defined distance are classified as non-ground points 53 and the ground proximity close to the ground is filtered according to the filtering.

This can be expressed by the following formula.

That is, in the plane consisting of the ground point candidates 51 of FIG. 3, which are ground planes of the ground separation candidates of the first separation stage located in an area adjacent to the ground, a distance to an arbitrary point i is denoted as di, This may be represented by Equations 3 and 4 above.

Here, the number of the ground point 71 is adjusted according to the distance of di. In the present invention, the ground point candidate 51 having the distance of 0.3 [m] or more is removed.

Since the vertical error of the ALTM 3070 is 15 cm when the flight altitude is 1200 m, it is preferable to set it to 30 cm in consideration of the point where the flight altitude is 1500 m and the loss of the ground point, which can be changed.

In addition, the secondary separation step 71 is to determine the non-ground point 73, which is the ground surface, which is records reflected in the area close to the ground, from the data separated by the ground point 71 as a result of the primary separation step 73. By searching and removing, we can select only the data that are appropriate to represent the continuous space in the numerical elevation model.

4 is a flowchart schematically illustrating a method of extracting Lidar terrain information using vegetation transmission characteristics of the present invention. As shown in the figure, the Lidar terrain information extraction method using the vegetation permeation characteristics according to the present invention separates a single response that is duplicated and recorded in Lidar, which stores binary data at the time of data acquisition into an initial response and a late response. To begin with the initial separation step (S10).

The non-ground point, which is the response of the ground point candidate and the crown portion, is separated from the initial response and the final response, and a first separation step S20 of removing the non-ground point is performed, and the ground point candidate is maximized on the ground. Based on the adjacent surfaces, the ground point candidate, which is a ground proximity exceeding a certain distance, is subjected to the secondary separation step S30 of removing.

In other words, in the first separation step, only the response of the crown portion is removed from the initial data to filter out the ground point candidates, and in the first separation step, in the case of an exception or the surface consisting of the ground point candidates for perfect terrain measurement, By removing ground point candidates that exceed a certain distance from the surface adjacent to the ground, it removes the elements that hinder accurate representation of the terrain due to the forest area, and can be applied to various terrains such as water pipes and structures with artificial structures. have.

5 is a graph according to an initial separation step of the Lidar terrain information extraction method using the vegetation permeation characteristics according to an embodiment of the present invention, and will be described with reference to FIG. 3. As shown in the figure, the vegetation according to the present invention Lidar terrain information extraction method using the permeation characteristics used the height characteristics of the tree in the region that forms a high density of trees, which is a characteristic of Lidar data recorded in the forest area.

Here, by separating the Lidar original data into pure initial response, single response, pure late response, etc., the characteristics of a single response are retrieved. Artificial structures such as buildings or roads are mainly recorded as a single response, where the vegetation is concentrated. Many single responses are reflectively recorded in layers, and most Lidar data are recorded in a single response, even where there are no trees present or where the surface is exposed.

Therefore, the amount of single response in a certain area indicates the density of trees, and in the case of a single response, it can be seen that it is distributed in a large part of the crown and partially exists in the lower part of the crown. have.

The data of the initial classification stage classified into pure initial response, single response, and pure final response are reclassified into two types of ground point candidate and non-ground point in the primary separation stage.

In addition, Figure 3 shows the result after performing the primary separation step for the data shown in Figure 5, after the primary separation step is performed, the secondary separation to select only the final ground point, Table 1 shows the information on Lidar data for the study area and the results for the final separation results.

Research Area Sum(#) FR LR PFR SR PLR Water pipe Ground point Free Ground Point One 50,574 38,593 38,589 11,985 26,608 11,981 32,066 13,613 36,961 2 85,736 72,632 72,595 13,138 59,594 13,104 40,675 32,992 52,744

As described above, the study area 1 and the study area 2 both have a large number of single responses, and in particular, the study area 2 in which the road exists has many single areas that are reflected and received on the road.

In addition, the number of separated ground points in study area 1 is 13.613, which corresponds to 26.92% of the total data, and the number of paper points in study area 2 is 32.992, which corresponds to 38.48%.

In addition, in study area 2, roads were separated by ground points, and due to the smaller forest distribution area than study area 1, many ground points appeared, and the number of water pipe layers was reflected from the water pipes. Refer to the number of materials.

FIG. 6 is a diagram illustrating Lidar terrain information extraction using vegetation permeation characteristics according to an embodiment of the present invention, and FIG. 7 is a diagram illustrating Lidar terrain information extraction using vegetation permeation characteristics according to an embodiment of the present invention. It is a figure applied to the study area 2 by the method.

As shown in the drawings, the Lidar terrain information extraction method using the vegetation permeation characteristics according to the present invention is the ground point data separated for the study area 1 and study area 2 in Table 1 and the numerical values produced using the same Show the altitude model.

Here, 7 (a) and 8 (a) is a superposition of the ground point separated from the Lidar data with the digital aerial image, (b) shows a digital altitude model was produced using the ground point of (a) , (c) shows the superposition of numerical elevation models and ground level data together.

As a result of separating the ground points, since there are not many ground points, there is an area (o-zone) that is not easy to interpolate in constructing a numerical altitude model. When there is a high density of pine trees, the vertical distribution of Lidar data is present. As can be seen, the Lidar signal transmitted and reflected deeply below the water pipe is significantly reduced.

Although the preferred embodiments of the present invention have been described above by way of example, the scope of the present invention is not limited to such specific embodiments, and those skilled in the art are appropriate within the scope described in the claims of the present invention. It will be possible to change.

As described above, according to the present invention having the configuration as described above, it is not easy to extract terrain information by classifying non-ground points step by step based on the elevation difference between the initial response signal and the terminal response signal using Lidar point data. It is easy to build local numerical altitude models, and can show pure surface in forest terrain where vertical interference factors such as trees and lower vegetation exist, and can increase precision.

Claims (5)

An initial separation step of dividing the binary response obtained by reflecting the Lidar signal in the forest area into an initial response and a late response to separate a single response that is repeatedly recorded from the ASCII type Lidar data stored therein; A first separation step of filtering the non-floor point by separating the non-ground point, which is the response of the ground point candidate and the crown portion, from the initial response and the late response; A second separation step of filtering a ground point candidate, which is a ground proximity object exceeding a predetermined distance based on a surface closest to the ground among the ground point candidates; Lidar terrain information extraction method using vegetation permeation characteristics comprising a. The method of claim 1, The primary separation step is Comparing the elevation of each response recorded in the area using the elevation of the highest reflected reflexive response and the lowest of the late response; Setting a non-floor point when one response elevation is close to an elevation of the initial response maximum; Setting as a ground point candidate if one response elevation approaches a terminal response minimum; Lidar terrain information extraction method using the vegetation permeation characteristics, characterized in that consisting of. The method of claim 1, Lidar terrain information extraction method using vegetation transmission characteristics, characterized in that for filtering the initial response recorded by using the elevation difference between the initial response and the terminal response in the first separation step. The method of claim 1, The secondary separation step Setting a plane consisting of data closest to the ground among the ground point candidates as a reference plane; Classifying a ground point exceeding a predetermined distance from the reference plane as a non-ground point; Filtering a ground proximity proximate to the ground according to the filtering; Lidar terrain information extraction method using a vegetation permeation characteristics comprising a. The method of claim 4, wherein Lidar terrain information extraction method using the vegetation transmission characteristics, characterized in that the number of ground points is adjusted according to the predetermined distance.

Images