KR101349375B1 - Method of automatic three dimensional plotting of building for three dimensional geographic information system - Google Patents

Abstract

Disclosed is an automatic drawing method using an automatic drawing system which automatically draws a three-dimensional polyhedron or polygon of a building at the middle of geographic features in order to draw an automatic three-dimensional space of the building for a three-dimensional geographic information system. On the premise that the target of predetermined data is installed at multiple inflection points on the three-dimensional contour of a building, and aerial image information of a target-detectable form is acquired through an aerial photography, the disclosed automatic three-dimensional space drawing method of a building includes the following steps of: storing the aerial image information in an aerial image data acquiring unit by correlating a standard point measurement result with an aerial image; performing the mapping of a target about the aerial image data in a target mapping unit; composing a projection image which includes three-dimensional information by using the target mapping product of the aerial image data and digital elevation data in a projection image composing unit; and converting corresponding raster data into three-dimensional vector data by automatically confirming a target in the projection image in a three-dimensional vector converting unit and producing a three-dimensional polygon about the building of the projection image. [Reference numerals] (S1) Step for storing target data; (S2) Step for photographing aerial image information; (S4) Step for mapping a target; (S5) Step for storing digital evaluation data; (S6) Step for producing a projection image; (S7) Step for a three-dimensional vector; (S71) Step for reading a three-dimensional building shape; (S711) Step for extracting a band; (S712) Step for extracting a target; (S713) Step for producing a three-dimensional polygon about a building

Description

Automatic 3D spatial image automatic drawing of building for 3D geographic information system {method of automatic three dimensional plotting of building for three dimensional geographic information system}

The present invention relates to a three-dimensional spatial image drawing of a ground building, and more particularly, it is possible to automatically read and extract a three-dimensional outline of a ground building on which a target is installed from an aerial photographed image and automatically reflect the spatial image drawing data. The present invention relates to an automatic three-dimensional spatial image automatic drawing method of a building for a three-dimensional geographic information system that can generate three-dimensional geographic information data that can prevent errors while quickly and efficiently performing three-dimensional spatial image drawing.

With the recent development of computers and software, the development of precision optics and laser measuring devices, digital map production has become possible, and the development of the mapping paradigm is rapidly changing from conventional analog mapping to digital mapping.

Image drawing is a technology in the field of processing aerial photographed image with corresponding data and making it into a map.A digital map including coordinate information such as longitude, latitude, and sea level at each point of the map is a two-dimensional digital map and three-dimensional map. There may be a three-dimensional numerical map of.

A digital map is a digitized map that is produced in a form capable of computerized processing of various geospatial information to be displayed on a feature or other map, and can be analyzed and edited using a computer.

Digital maps are the geographic information that is the basis of all digital maps. They are very useful for use, freely scale, free from distortion, and integrated with various GIS data. Although it is economical, it still requires a lot of cost, effort, and time, and thus efforts are being made for a more economic way of producing digital maps.

On the other hand, the demand for a map model that can provide more realistic and realistic information in the use of geographic information system is increasing. In the case of two-dimensional digital maps, many parts of spatial information cannot be expressed, but the three-dimensional digital maps can be included to represent these parts, so that people who are not familiar with maps can easily recognize and express them. There is an increasing demand for 3D digital maps, which have a large amount of information and are highly applicable to various fields.

For example, 3D digital maps eliminate the need for reference materials to identify building heights in urban planning, establish predictions and preventive plans related to the environment and disasters, utilize them in civil design such as road construction, and build infrastructure in the ICT sector. It can be used very conveniently in providing information for a vehicle, providing information of a vehicle navigation system, and designing an aerodrome.

However, the existing digital map production is mostly made on the basis of the two-dimensional digital map production process, and there are many parts in which the development of the tool for the production of the three-dimensional digital map has not been sufficiently made yet. In addition, since the amount of work for mapping is much greater in order to express more extensive information than the two-dimensional digital map, the three-dimensional digital map is fully utilized. In addition, there is a need for an automatic drawing method that can automatically create a three-dimensional form of a feature represented on a screen or paper map.

Even in the production of a three-dimensional digital map, the pre-analysis step for a two-dimensional digital map and a substantial part of the analysis map can be performed as it is. For example, when using aerial photographs obtained from aerial photography, aerial photographs are taken at some overlapping angles and analyzed and processed to obtain three-dimensional coordinates for each point in the target area, based on which two-dimensional or You will create a 3D digital map. Therefore, it can be said that the two-dimensional digital map and the three-dimensional digital map are distinguished through the drawing analysis.

A drawing is a work that describes various features of a target area. A drawing is a work that can be done on a monitor, but it is a conceptual work because it is not a line drawing on paper but is actually processing data about an object. .

The three-dimensional digital map expresses each layer of the terrain and features three-dimensionally, and the building stereoscopic drawing is also performed for the building, which is a representative layer. Building three-dimensional drawing means the three-dimensional drawing of a building among artificial landmarks. Building shape data for building three-dimensional image is one of the important data to be able to grasp the information necessary for real life most intuitively, especially in the city where office space and residential high-rises are made high.

In the drawing operation, a vectorization operation is usually performed to convert raster data obtained by aerial photographing results into vector data commonly used in digital maps using various programs.

In the past, when drawing a building plan for a two-dimensional digital map, many methods of extracting and drawing building outlines or semi-automatically determining building outlines using a program have been used. This manual work or semi-automatic work requires the effort and time of the worker, which increases the cost and time of digital mapping. In building stereoscopic drawings for 3D digital maps, the amount of outlines to be drawn is much greater, so the manual work is more expensive and time consuming.

Recently, a program that reflects the progress of the image processing development and automatically performs the building plan drawing using the characteristics of each layer is used. However, the automation by image processing has its own limitations, and it is still difficult to plan buildings in complex urban areas, and in building three-dimensional drawings, errors are more likely to occur in automation tasks due to the complexity of the work itself.

Therefore, there is a need for an automatic drawing method of a building stereoscopic which can prevent the occurrence of an error of the existing automatic drawing while automatically extracting an outline for building stereoscopic drawing.

Republic of Korea Patent No. 10-0545358, Name: restoration method of three-dimensional building using digital map and aerial laser survey data. Korean Unexamined Patent Publication No. 2001-0000443, Name: Recording medium storing a building extraction system and method by fusion of aerial survey image and rider data, and a program source thereof. Republic of Korea Patent No. 10-1020915, Name: Boundary filtering method of a two-dimensional urban building object using RGB band emphasis and IHS color transform.

The present invention can compensate for the problems of the conventional building automatic drawing method in the automatic three-dimensional spatial drawing of the building, while automatically extracting the outline for the three-dimensional representation of the building while preventing the occurrence of errors in the existing automatic drawing. An object of the present invention is to provide an automatic three-dimensional spatial diagramming method for buildings.

The automatic three-dimensional spatial drawing method of the building to achieve the above object,

An automatic drawing system that automatically draws three-dimensional (three-dimensional) polyhedrons or three-dimensional (three-dimensional) polygons of a building among three features to form a three-dimensional digital map. Automatic 3D spatial drawing method of building using

The automatic drawing system includes a target specification definition unit which stores the target specification, an aerial image data acquisition unit which stores the result by correlating the drawing coordinate data obtained by performing a point surveying operation to the image obtained through aerial photography, and a figure for the feature. Numerical elevation data storage unit for receiving the elevation data received from the outside of the automatic drawing system, the expression unit for performing the expression on the aerial image data received from the aerial image data acquisition unit receives the information on the expression element, the table Building inflection points (including bend points) are clearly defined as the reference point of the building's three-dimensional polygons, including targets. An orthoimage to determine the altitude of the point) And a three-dimensional vector conversion unit for converting and reconstructing three-dimensional raster data included in the ortho-image information created by the orthoimager to three-dimensional vector data, wherein the three-dimensional vector conversion unit defines the target specification in the orthoimage. It is equipped with a three-dimensional building shape reading unit that recognizes the building part by using the target specifications obtained from the department, and processes the boundary line or contour that represents the three-dimensional outline of the building part to form a three-dimensional polygon.

An expression step of performing an expression using an expression element in the expression unit and aerial image data from the aerial image data acquisition unit;

The orthoimage comprises an orthoimage including the altitude information at all positions of the digital map production area by using the result of the facial expression by the facial expression unit and the numerical elevation data obtained from the numerical elevation data storage unit. Configuration steps,

And a three-dimensional vectorization step of converting the three-dimensional raster data of the orthoimage including the altitude information into three-dimensional vector data in the three-dimensional vector conversion unit.

The three-dimensional vectorization step automatically three-dimensional polygons of the building image of the orthoimage including the altitude information using the orthoimage including the altitude information in the three-dimensional building shape reading unit and the target specifications of the target specification definition unit. It is provided with a three-dimensional building shape reading step of processing to achieve.

At this time, the three-dimensional building shape reading step is made by a three-dimensional building shape reading unit provided with a band extraction unit, a target extraction unit, a building extraction unit,

A band extraction step of extracting an image of a band suitable for detecting the target from the orthogonal image by the band extracting unit;

Calculating, by the target extracting unit, a pixel group indicating a target specification characteristic among pixels constituting the building candidate region in the band image, comparing the area of the target specification with the area converted from the image of the pixel group; Identifying targets by checking whether the value falls within the tolerance range. Vectorized targets by automatically converting raster data about targets that can be obtained through orthoimage information including altitude which is the work result of the orthoimage component. Obtaining pattern data,

Using the target pattern data vectorized by the building extracting unit, extracting the 3D raster data of the inflection point of the building using the orthoimage information that is the result of the orthogonal image construction unit, and the target including the branch direction of the target around the inflection point. Depending on the nature of the specification, two or more extension lines, including extension lines in the branch direction, may be created, such as two extension lines perpendicular to each other in the horizontal plane, and perpendicular or inclined extension lines, but not from another target or the target, One of the individual polygon faces made by connecting the other targets of the building and extends until they meet substantially, taking into account the effective error, and extracting the effective extension lines, which are these substantially encountered extensions, in other words Gun surrounded by effective extension Determine the three-dimensional polygon to be made by having a conversion building a three-dimensional polygon screen and storing a three-dimensional vector data.

According to the data processing that is converted into three-dimensional vector data, the building can be recognized as a three-dimensional (three-dimensional) polygon when the digital map is visually implemented on the ground or the screen.

In the present invention, after the step of automatically drawing the three-dimensional polygon of the building three-dimensional (stereoscopic) contour is made, the user sees the three-dimensional digital map created based on the display or the paper map. Unlike the 2D digital map, an implementation program is needed to show the target area on the map according to the direction and altitude.

In the present invention, the image or image data of each step means a collective data or data file that has information for displaying a visual image on a monitor screen or the ground through an appropriate program and hardware. It can be said that.

The step of automatically drawing three-dimensional polygons is implemented on a working monitor screen in a computer operator's computer system that embeds an ortho-image as a target image in vector mapping or vector editing in digital mapping, so that three-dimensional polygons are automatically drawn. You can check the drawing process.

In the present invention, a target having a predetermined specification is installed at a plurality of places forming an inflection point of a three-dimensional shape of a building, and obtaining aerial image information in a form capable of detecting the target through aerial inspection is a process that must be made for the entire drawing operation. It may be performed by a general subject or an ordinary aerial photographing subject and photographic equipment.

In the present invention, the expression unit or the orthogonal image unit for constructing the facial expression or the orthoimage and the orthogonal image to obtain the altitude and elevation of each position, as in the conventional numerical drawing operation, the analysis device and the computer device coupled thereto It can be said that the functional part of the total body composed of combined programs. In some cases, it is made in the state of inputting the pre-written numerical elevation model. The division of the functional part is mainly made by the partial program of the software constituting the system. And can be seen.

In the present invention, the three-dimensional vector conversion unit is provided with the object data obtained as a result of the normal image construction work, the drawing program, and a computer device embedded therein, the three-dimensional building shape reading unit and the three-dimensional vector conversion unit The band extractor, target extractor, and building extractor are also conceptually divided by the partial programs constituting the entire drawing program.

For example, using the information on the target to recognize the building area and the overall shape of the building, and to determine and create the three-dimensional polygons to be implemented on the digital map, the partial program in the drawing analysis and vectorization program for the digital drawing production in the computer system It can be performed automatically using a building three-dimensional image processing program.

In the present invention, the position where the target is installed is an inflection point or an important inflection point that is important for forming a three-dimensional contour in the building area of the orthoimage, in addition to the position of forming a bending point or inflection point that is a boundary line with the outside of the area occupied by the building in the orthogonal image. Make sure the reference points are included. The exposed surface of the target may be made of a material and shape having optical properties such as color and reflectivity that can be remarkably recognized in a predetermined band. Selection of a band suitable for recognizing a target may be performed by selecting an R band if the target is a red (R) color, for example.

The target can be installed in an appropriate number according to the shape of the building, and can be installed on the inflection point which is the convex angled part or the concave angled part of the building outline, and the different surface characteristics, to distinguish the convex angled part and the concave angled part, It can be installed to have band recognition characteristics.

The shape of the target installed at the inflection point is composed of two points that point to the two directions in which the building outline extends from the inflection point or the angular point on the outline of the building, and the angle formed by the two is based on 90 degrees. Increased or decreased targets are also possible, but different targets are possible.

In particular, the target for three-dimensional stereoscopic drawing has a three-dimensional configuration having two branches in addition to the inclined or vertical branches with respect to the horizontal plane, or instead of the inclined or vertical branches off the horizontal features of the inflection point You may want to have a color or other form of marker that can be displayed. The computer program can identify the characteristics of these target specifications and draw horizontal extension lines in the vertical, vertical, and inclined directions when drawing extension lines, and determine which targets are connected to which of different altitudes and reflect them in the automatic drawing process. have.

The target can be additionally installed at a point within the individual polygons of the building to indicate that it is an area within a single polygonal plane of the building's three-dimensional polygons. Targets of this attribute also install surface characteristics or shapes at angled parts. Can be formed differently from the target to be installed.

According to the present invention, since the three-dimensional contour recognition of the building can be clear, the building three-dimensional drawing operation is made automatically, and the drawing operation can be done quickly and efficiently and economically, and at the same time, the drawing operation can be made accurately.

The present invention is additionally made in combination with the existing automatic drawing method to supplement the problems of the conventional automatic drawing method, it is possible to automatically check the three-dimensional contour of the building, error detection and correction, and replace the existing automatic drawing method It is possible to draw three-dimensional outlines of buildings automatically while being made independently.

1 is a conceptual diagram showing the schematic configuration of an automatic drawing system for the automatic drawing method of the present invention;
2 is a flowchart showing an embodiment of an automatic drawing method according to the present invention;
Figures 3a to 3c is a view illustrating the form and installation form of the target used in the automatic drawing method of the digital mapping building plane using the target,
4 is a flowchart showing an example of the detailed step configuration that a possible building shape reading step in one embodiment of the method of the present invention may have.
5 is a diagram illustrating each step of the process in order to explain the process of building polygonization.

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

As shown in FIG. 1, an automatic drawing system according to an embodiment of the present invention is largely divided into an automatic drawing computer system 100 and an analysis drawing device 200, and the computer system includes information and data unit 110. And the program unit 120.

The information and data unit 110 stores the target specification definition unit 111 storing the target specification, the aerial image data obtaining unit 113 storing the aerial image data, the numerical elevation data storing unit 115, and the numerical map. The digital map storage unit 117 is provided.

The program unit 120 receives the information about the facial expression elements, the facial expression unit 121 for performing facial expressions on the aerial image data, the result of the facial expressions and orthoimages including the altitude data of each position using the numerical elevation data. An orthoimage image forming unit 123 and a 3D vector conversion unit 125 for converting and reconstructing raster data forming orthogonal image information into 3D vector data. A three-dimensional building shape reading unit 1251 is provided which recognizes the portion and processes the boundary line or contour of the corresponding portion of the building to form a three-dimensional polygon. The three-dimensional building shape reading unit 1251 includes a band extracting unit 1251a and a target. And an extraction part 1251b and a building extraction part 1251c.

However, in the configuration of FIG. 1, the program unit substantially executes the program for each step in the computer system, and connects the information by connecting with other parts such as data, data interpreter, or analysis part other than the part where the program is stored in the computer storage device. It can be seen as a concept that includes a processing device for transmitting and receiving.

A target specification storage step S1 and an aerial image data storage step S2 shown in FIG. 2 by such an automatic drawing system; Facial expression step S4; A digital elevation data storage step of storing digital elevation data such as terrain digital elevation model (DEM) data in the digital elevation data storage unit (S5); Ortho-image production step (S6); A three-dimensional vectorization step (S7) for the building of the ortho-image is made, the three-dimensional vectorization step includes a three-dimensional building shape reading step (S71), and the three-dimensional building shape reading step includes a band extraction step ( S711), the target extraction step (S712) and the building three-dimensional polygonalization step (S713).

At this time, the target specification storage step (S1) is made on the premise to define and determine the target specifications, to produce a target, and to install the target in the target building,

In the three-dimensional vectorization step, after the three-dimensional building shape reading step, the automatic three-dimensional spatial mapping method of the building can be completed by registering the building layer property in the digital map for the three-dimensional building polygon.

In addition, when such three-dimensional space automation is usually made, the final completion is achieved by performing an error check.

As a result, in the completed 3D digital map, these building 3D polygons can be visually realized on the display screen or on the ground, and the expression program that implements the 3D digital map on the screen or ground is different from the 2D digital map. The location, the reference altitude, and the direction viewed are displayed to display a three-dimensional numerical map. This representation has a form similar to that displayed on the monitor according to the view program of the existing three-dimensional navigation.

Looking more specifically at the steps of the embodiment of the present invention, for the present invention, first, the specifications of the target should be determined to manufacture the target, and the manufactured target should be installed in the building.

Target specifications may include shape and size, color, surface reflectivity for specific bands, and the like. The most commonly used form is the L-shaped marker. The L-shaped cover is usually formed so that the two extending from the center located at the inflection point of the building's outline form an angle of 90 ° to each other, but the inside angle may be formed at an angle other than 90 °. Marks may be installed with different specifications to recognize the difference. In addition to the planar position at the inflection point of the outline of the building identified by the target, the elevation of the target and the elevation of the installed point are identified.

The target needs to be formed in a size that can be sufficiently recognized in consideration of the recognition in the aerial shooting and the picture analysis step, but if it is formed too large, the manufacturing and installation costs increase. The target can be formed mainly by using metal panels or synthetic resin panels in consideration of durability and surface characteristics, and can be manufactured separately and attached to the building, but it is also possible to think of the method of forming the target itself in the form of paint coating and installing it integrally with the building. have. The surface of the target makes it possible to be noticeably recognized generally in a conventional aerial shooting environment, in which case it is possible to reduce the size of the target in terms of specifications.

The targets can be installed in an appropriate number according to the shape of the building.The targets can be installed around the vertices that are convex angles or concave angles of the building outline, and the targets that are installed on them so as to distinguish the convex angles and the concave angles. Can be installed to have different specifications, for example different surface characteristics, recognition characteristics on a band.

In particular, the target for the three-dimensional stereoscopic drawing has a three-dimensional configuration that has a branch inclined or perpendicular to the horizontal plane in addition to the two branches, or a color that can characterize the inflection point in place of the off-branched branches. You can also have a cover.

A target can be additionally installed at a point within this polygon to indicate that it is within the area of the roof or topmost polygon of a number of individual polygons representing the building geometry. Unlike the target, for example, it can be formed in the form of X or ㅁ. If the target of this property is located outside the area of the roof or top of individual polygons representing the top surface of a three-dimensional building in the image obtained as a result of the drawing, the drawing of the building polygon may be wrong and error correction may be made.

3A to 3C illustrate the shape and installation form of the target used in the automatic drawing method of the digital mapping building plane using the target. Here, the shape of the target 21 of FIG. 3a is an L-shape, and the cabinet is manufactured to be applicable to a polygonal shape by adding or subtracting an angle between branches at an inflection point in accordance with the outer shape of the upper part of the building. 3B is a plan view of the case where the target 21 is provided in the building 22, and FIG. 3C is a perspective view of the building in which the target is installed.

After the target is installed, operations such as aerial photography and facial expressions, which are normally performed for creating a digital map, are performed, and ground reference point surveying or photographic reference point surveying for facial expressions before facial expressions can be performed in a similar manner to a normal digital map production process. have.

For the aerial image data storage step (S2), a preliminary operation of acquiring photographs and satellite images taken by aircraft or satellites is premised. In the case of producing an orthographic image using such aerial photographs, in order to interpret two-dimensional information on a photographed photograph three-dimensionally with respect to a photographed subject region, at least two different angles of the entire subject region are applied. Photo must be taken.

Photographing consists of 60% Endoverlap where each image is overlapped in the direction of the aircraft with overlapping images for map editing, and 30% Sidelap to connect between one flight path and the next. Will be taken.

Films of aerial photographs are scanned and saved as files. In the scanning process, an original image may be obtained by dividing an image of each band corresponding to an R, G, or B or near infrared wavelength. Since a specific band may have a property to identify a specific feature well, it is to use such a point to obtain a band-specific image.

Since aerial photographs do not have coordinates as image data, local ground surveys should be conducted at points that can be easily identified in the image. Is done through.

Since surveying is time-consuming and expensive, ground reference points are extracted from points that can be easily identified from digital topographic maps. In general, the ground reference point is selected from the point of easy photographic identification such as the intersection of roads, bridges, waterways or road bends and mountain peaks, and image processing is required to more accurately select the position of the reference point.

For example, histogram equalization for dark images, contrast techniques based on thresholding, and filtering to make the boundary visible clearly may be used. The ground control point should be selected to be distributed evenly within the target area, and the position correction can be performed even with a small ground control point using the software applying the interpolation method.

This reference point surveying operation consists of inputting a reference point, editing a reference point, and arranging the reference point, and the inputting of the reference point may be converted into an internal format by inputting an input format in the form of ASCII of reference point ID, X, Y, and Z. .

In the editing of the reference point, the reference point is divided into a plane point and an elevation point, and if there is an error in inputting coordinates into an ASCII file, an edit is required, and the reference point is added or deleted.

The arrangement of the reference point may divide the input reference point into a plane point and an elevation point so that the plane point is positioned in the edge region as the target area, and the elevation point may be evenly distributed inside the target area. In addition, in order to improve the accuracy, a certain number of conjugate points (Tie Points) are arranged for each picture and positioned in an overlapping area so that an error according to a photographing path does not occur.

Regarding the facial expression step S4, the facial expression element must be input for the facial expression, and is usually divided into an internal facial expression and an external facial expression, and the external facial expression is divided into a mutual expression, a geometric expression, or an absolute expression. Internal expression corrects distortions in photographing and distortions in scanning, and converts the coordinates of an object observed in the image into a coordinate system whose origin is the camera projection center. This distortion can be found in the lens manufacturing process and the digitization of analog photographs.

Since the lens cannot be realized with the ideal shape that was planned at the time of manufacture, the light is refracted, unlike the original point, as the light passes through the lens, causing distortion in the image. In addition, in the process of digitizing aerial photographs, distortion occurs in the image due to various factors such as physical changes in the film of the aerial photographs, and the camera focal length, the coordinates of the tavern, the fiducial marks, and the radial distortion value. You can perform internal expression as an input value.

External expression refers to reproducing the geometry at the time of photographing aerial photographs and satellite images. Here, the elements constituting the geometry are called the external expression 6 elements {Ω, Φ (Phi), Κ (Kappa), X, Y, Z}, and the expression elements are determined by various methods.

Cross-expression and absolute expression are performed to remove the time difference and establish the relationship between the model and the object. A parallax is a visual error of a vertical and horizontal direction occurring at a point where a store to the same point must meet vertically in a pair of photographs, and the parallax of the flying direction (x direction) is transverse parallax and the direction perpendicular to the flying direction ( The component in the y direction is referred to as the longitudinal difference.

Relative Orientation sets the relative inclination relationship between a pair of photographs at different angles, eliminating the time difference by matching the points determined to be the same point in the overlapping area with respect to one object. do. Six markers per object are used, three points per picture, and the facial expression points are set to spread as evenly as possible in the central area with less distortion for the picture.

Absolut Orientaion is the process of correcting the slope and scale to make the terrain and the model different, and the relationship between the model and the target space is established during the process of absolute expression. For the image, the observation of the reference point (or world coordinate) for the corresponding object space is required.

The data indicating the state of the terrain by the numerical value is called numerical elevation data. The numerical elevation data is a numerical record of the height value of the points distributed on the surface at regular intervals, and is used for the analysis of the terrain. This is the basis for the creation of.

The ways in which numerical elevation data are created and stored can be summarized in several ways.The Elevation Digtal Model, which is stored as a grid, by contour lines created by connecting points of the same height continuously, fault The profile method and the triangular Irregular Network (Triangular Irregular Network).

In the storage of numerical elevation data such as terrain DEM data (S5), an elevation digtal model is produced by extracting contours and elevation points from a topographic map of a certain scale in order to use them as data to correct the relief of aerial photographs. After processing the surface by using the TIN method, it is possible to extract and input the elevation data of a certain lattice, or input the separately produced terrain DEM data.

Ortho-image production step (S6) is an ortho-projection by modifying the deviation caused by the center projection or ground ups and downs of the characteristics of the camera at the time of taking the picture, using the aerial image data facial expression results and numerical elevation data Can be produced.

Therefore, for each position of the orthoimage obtained in this way, the numerical value at the reference position can be obtained by applying the neighboring measurement value and the interpolation method to the elevation of the surface and the upper elevation of the facility such as the building.

Next, while the three-dimensional vectorization step is performed for the building in the produced orthoimage (S7), the three-dimensional building shape reading step (S71) is performed among the features in the process.

In this three-dimensional building shape reading step, a target extraction step (S712) for recognizing the target in the orthoimage is performed, and the altitude and the surface elevation at the installation position (plane coordinates) of each target are grasped and extracted by these targets. Building 3D polygonalization step S713 is performed by connecting the building inflection points to each other to define an outline of the building or a three-dimensional polygon and store it as three-dimensional vector data.

Recognition of the target in the orthoimage may be accomplished through a kind of image processing program, in which the data is pre-populated in a computer mapping system composed of computer hardware and software to recognize the target by the specification of the target. By using the input data, the target is recognized in the orthoimage through the comparison and the inflection point and the branch direction are known.

For example, in a simple cuboid building on a flat surface, if you install two L-shaped targets at diagonal vertices in the top plane of the building, the three-dimensional image processing program that forms the building extractor will be located at two vertices in two directions on each of the targets. Draw extension lines, identify the reference points corresponding to the targets at the two points where the extension lines meet, define four reference points on the top plane of the building, recognize the rectangles that make up these reference points as the building plane, and place vertical lines on the ground surface vertically downward from these reference points. The four points where the ground surface and the vertical line meet each other are considered as the bottom of the building on the ground, and the four corner lines (extension lines) on the top of the building, which form a rectangle, and the four corners on the side of the building formed by the vertical line, and the four points on the ground are connected. The four edges of the square to form the solid polygon of the cuboid do.

In the case where the building is formed by stacking a number of cuboids having a step width in the height direction and overlapping, the targets are installed at each level so that the targets form a planar polygon at each level, and the planar polygons are contours of a two-dimensional digital map. It is possible to grasp the actual shape of the building and represent the rectangular parallelepiped that forms each stage as 12 edges, and convert the edges of the whole building into vector data and express them according to the direction of the map.

In order to express the shape of these buildings more accurately, it is necessary to increase the number of targets installed and the types of targets given different three-dimensional processing methods, and to refine the program for processing them, but the cost may be increased. The present invention is limited to and may be used as an auxiliary means of the existing three-dimensional automatic drawing method.

Within the program, building outlines may be drawn directly by targets, but for some typical building types, a standard model is set up, and the program extends the standard model to one direction of the building if it determines that the object is a standard model. In addition, the vector data can be changed and processed by replacing the shape of the building with the generalized model through the three-dimensional rotation.

In this step, in order to recognize the target, for example, if the band most suitable for reading the target is a near infrared band, a band extraction step (S711) of extracting a near infrared band (image of the near infrared band) is performed first.

Then, pixels having brightness values corresponding to candidates for the building area are extracted from the image by band operation, and pixels having the same area as the target area (here, the area can be recognized through other specifications or properties of the target). Recognizing the target may be accomplished by extracting the group.

The representation of the building in three-dimensional polygons is primarily a drawing rather than a recognition, but can also be accomplished through some kind of image processing program.

The drawing program is one of the tools for the overall drawing for the drawing analysis step for the digital mapping. The drawing program is a function part for the automatic drawing plan of the building plane, and the sub-program such as the image processing program is added to the existing drawing program. In this case, if the operator selects to execute this sub-program, for example, the internal and external expressions are performed at the drawing analysis stage, and the target image or the target raster-type data obtained as the orthoimage is performed. Apply subprograms to the

When the sub-program is executed, targets matching the predefined specifications and attributes are recognized in the target image or the target data, and at the bending points of the recognized targets, they extend in the direction of the target to simply connect a group of targets to each other, If there is a point that extends the branch of the target and crosses each other and is not a target, a polygonalization operation of connecting the targets to each other to form a closed curve including the points is performed as one kind of conversion into vector data.

The above is a simple example, but the types of targets are classified into several categories, and the image processing program treats different types of targets differently and processes data according to predetermined rules to form polygons. May be done until noon is confirmed.

Fig. 4 is a flowchart showing an example of the detailed step configuration that a possible building shape reading step S71 in one embodiment of the method of the present invention can have.

In this case, first, the band extraction step (S711) and the target extraction step (S712), the pixel representing the target specification characteristics using the target specification obtained from the target specification definition unit among the pixels of the image of the building candidate region in the entire image obtained through the band operation The group is calculated as a target candidate (S7121), and the targets are identified by checking whether the comparison difference falls within an allowable error range by comparing the area of the target specification and the converted image (S7122), and using the target, the building inflection part The raster data including the target location and the altitude are extracted, and the raster data is automatically converted into the vector data for the targets to obtain a vectorized target pattern (S7123). Thereafter, extracting the inflection point raster data of the building using the target pattern (S7131), generating an extension line in consideration of the direction of the target branch around the building inflection point (S7132), and considering the point where the extension line meets the building 3D The building 3D polygonalization step S713 may be performed by including a polygon determination and data conversion step S7133.

Fig. 5 is a diagram showing each step of the process for explaining one exemplary process of building three-dimensional polygonization, in which identification number 401 shows an image of a building imprinted in an aerial photo, and 402 is a vector of raster data. In the state where the target in the image is converted from the raster data to the vector data by converting the data into the data, the image is highlighted. 403 extracts the center points of the inflection from the target pattern of the polygon. In the state connected by the line, 404 is the extraction of only the, line, 405 is the scale conversion method to obtain the building inflection point by matching the 과 line with the outline of the building area (building exterior wall), and 406 is the scale conversion. Extract the four inflection points a (Xa, Ya, Za), b (Xb, Yb, Zb), c (Xc, Yc, Zc), d (Xd, Yd, Zd) at the top of the building, Each opened down Deriving a point by drawing a line of intersection with the surface and create a three-dimensional polygon is a rectangular parallelepiped shape of the building is drawn by the eight points derived this way indicates a state shown.

The transformed data is integrated with the result of the classification of various layers representing roads and other features, and when the vectorization is completed, the result of the previous stage of the three-dimensional digital map corresponding to the three-dimensional drawing circle is obtained in the converted data form. Lose.

As described above through the embodiment, most ideally, the upper end of each building is formed in the same shape as the lower end, and the upper end forms the same level plane, and has a rectangular structure to facilitate the installation and reading operation of the target. Although suitable, the real building may not be made of the same level plane at the top and may be made of a complicated polygon, so that the target needs to be manufactured and installed in consideration of the recognition possibility at the time of shooting and the correct readability of the orthoimage.

That is, the target may be installed at different levels if it is recognizable on aerial photography. In the case of non-cuboid, the target is installed at the reference point (inflection point) on the three-dimensional coordinates that can best display the characteristics of the building geometry.

3D vectorization of buildings by registering the properties of building layers in digital maps for 3D polygons following the 3D vectorization step of drawing analysis process for digital map or reading 3D building shape in vector editing process The step of carrying out can be completed. In addition, when such a three-dimensional space automatic painting is usually made, the error check is subsequently performed to complete the three-dimensional automatic painting work.

While the present invention has been described with reference to only a few parts of the present invention in detail, for those skilled in the art, these specific techniques are only preferred embodiments, and thus the scope of the present invention is limited. The point is obvious. Therefore, the practical scope of the present invention will be defined by the appended claims and equivalents thereof.

21: target 22: building
100: computer system 110: information, data section
111: target specification definition unit 113: aerial image acquisition unit
115: numerical elevation data storage 117: numerical map storage
120: program unit 121: facial expression unit
123: an orthoimager 125: a vector converter
1251: building image reader 200: analysis

Claims (1)

An automatic 3D spatial drawing method of buildings by an automatic drawing system that automatically draws 3D (three-dimensional) polygons of a building among features to create a digital map for a 3D geographic information system.
The automatic drawing system includes a target specification definition unit in which a target specification is input and stored, an aerial image data acquisition unit storing a result by associating drawing coordinate data obtained by performing a reference point survey operation to an image obtained through aerial photography, and a feature A digital elevation data storage unit for receiving and storing the digital elevation data from the outside of the automatic drawing system, an expression unit for performing an expression using information on the aerial image data and expression elements obtained from the aerial image data acquisition unit, An orthoimage component which constitutes an orthoimage including altitude information at all points of the digital map preparation area by applying the numerical elevation data obtained from the numerical elevation data storage unit to the result of the expression, and the raster forming the orthoimage. A three-dimensional vector conversion unit for converting and reconstructing the data into three-dimensional vector data; The circle vector converting unit includes a three-dimensional building shape reading unit which recognizes the building part in the orthoimage and processes the boundary line or contour of the building part to form a three-dimensional polygon by using the target specification received from the target specification defining unit. ,
A target specification storage step of inputting and storing a target specification from the outside of the automatic drawing system by the target specification definition unit;
An aerial image data storing step of storing a result obtained by associating the aerial photographed image by the aerial image data acquiring unit with the drawing coordinate data secured by the reference point surveying operation;
An expression step of performing an expression with the aerial image data and an expression element inputted from the aerial image data acquisition unit by the expression unit;
A numerical elevation data storage step of inputting the numerical elevation data by the numerical elevation data storage unit and storing the numerical elevation data in an allocated area;
It includes altitude information at all points of the target area to create a digital map using the numerical elevation data received from the digital elevation data storage unit and the facial expression result of the aerial image data received from the facial expression unit by the orthoimage composition unit. Ortho-image production stage consisting of ortho-images,
And a three-dimensional vectorization step of converting raster data of an orthoimage including the altitude information into three-dimensional vector data by the three-dimensional vector conversion unit.
The three-dimensional vectorization step includes an orthoimage including the altitude information and the altitude information using the target specification by the three-dimensional building shape reading unit having a band extracting unit, a target extracting unit, and a building extracting unit. It is equipped with a three-dimensional building shape reading step of automatically processing the building shape of the image into a three-dimensional polygon,
In the 3D building shape reading step, the band extracting unit includes a band capable of detecting a target in an orthogonal image including the altitude information among a red band (R), a green band (G), a blue band (B), and a near infrared band. A band extraction step of extracting an image,
The target extracting unit calculates, as a candidate for the target, a group of pixels representing the specification characteristics of the target among the pixels forming the building candidate region in the band image based on the target specification information extracted by the target specification defining unit. Orthoimage information including the altitude information of the ortho-image component to identify the target by checking whether a difference value falls within an allowable error range by comparing an area on a specification with an area obtained by converting an image of the pixel group. A target extraction step of automatically converting raster data for the target obtained through the vector data into vector data to obtain a vectorized target pattern;
And a building three-dimensional polygonalization step of determining a three-dimensional polygon of the building by the building extraction unit and storing the three-dimensional polygon of the building in an allocated area.
The building three-dimensional polygonization step
A building inflection point raster data extraction step of extracting three-dimensional raster data of an inflection point of a building using the orthoimage information, which is the result of the orthogonal image construction unit, by the building extraction unit using the target pattern data;
The building extracting unit generates two or more extension lines among extension lines according to the properties of the target specification including branch direction extension lines of the target from the inflection point, but the other one of the two or more extension lines, an extension line from another target, and one of the ground surfaces. Generating an extension line around the inflection point of the building to identify an effective extension line
The building extracting unit comprises a building 3D polygon determination and data conversion step of determining and storing the 3D polygon of the building with the effective extension line as a 3D vector data for the 3D geographic information system. Automatic 3D spatial image automatic drawing method of building.

Images