WO2006134962A1

WO2006134962A1 - Projection diagram generation system

Info

Publication number: WO2006134962A1
Application number: PCT/JP2006/311918
Authority: WO
Inventors: Shigeo Takahashi; Kenichi Yoshida; Tomoyuki Nishita; Kenji Shimada
Original assignee: The University Of Tokyo
Priority date: 2005-06-15
Filing date: 2006-06-14
Publication date: 2006-12-21
Also published as: JPWO2006134962A1

Abstract

[PROBLEMS] To provide a system capable of comparatively rapidly generating a projection diagram such as a route while eliminating shield of geographical characteristics. [MEANS FOR SOLVING PROBLEMS] A processing unit (1) performs the following processes: (1) A process for extracting a characteristic point expressing geographical characteristics in a part which may be associated with a shield from a 3D geographical model. (2) A process for calculating an optimal arrangement of the characteristic point on a 2D projection plane when viewed from a certain point of view. (3) A process for generating a 3D geographical shape satisfying the optimal arrangement of the characteristic point on the 2D projection plane. (4) A process for generating a 2D projection diagram when the 3D shape is viewed from the point of view or from its vicinity.

Description

Specification

Projection diagram generation system

Technical field

[0001] The present invention relates to a projection map generation system.

Background art

[0002] The car navigation system has made remarkable progress in recent years, and it has a 3D view as a function as a bird's-eye view, making it easy for users to associate with the actual landscape. On the other hand, if the 3D terrain shape is displayed three-dimensionally using perspective projection, the searched route will be shielded by buildings and terrain ups and downs, and the information necessary for driving cannot be conveyed.

[0003] In commercial car navigation systems, the problem is that, in situations where there are multiple overlapping force shields, such as displaying the shielding objects transparently, the route and geographical features that are of interest also vary. They overlap, making it difficult to convey their relative positional relationship. Therefore, in order to drastically solve such a shielding problem, it is necessary to properly arrange roads and geographical features on the projection map and to present them to the user by breaking the law of perspective projection. Such a projection that has lost the perspective projection is referred to herein as a non-perspective projection.

[0004] In the research of non-perspective projection, several methods have been proposed. A method of creating a projection that combines the effects of projection from different viewpoints by distorting the perspective projection image using two-dimensional transformation [Non-Patent Documents 4 and 8] and a method of breaking the perspective projection with a curved projection line [Non-Patent Document 2], a method of creating a panorama image used for the background image of a cell animation as seen from a specified camera path [Non-Patent Document 7], rendering individual independent objects with different viewpoint powers. In addition, a method of synthesizing on a two-dimensional screen in consideration of depth [Non-Patent Document 1] has been proposed. Among them, a technique for improving the expression ability of non-perspective projection by transforming a three-dimensional model has been devised in recent years [Non-Patent Documents 3, 5, 6]. A variety of non-perspective projections can be expressed by using the deformation of the 3D model.

[0005] However, even with the techniques described in Non-Patent Documents 3, 5, and 6, it is difficult to directly grasp the behavior on the two-dimensional projection plane from the deformation of the model in the three-dimensional space. others Therefore, it is difficult to confirm whether the shielding of the path is avoided in the actually obtained 2D projection plane. If recalculation is performed when occlusion is occurring, the amount of computation and computation time will increase. In particular, if an animation is generated by continuously displaying a two-dimensional projection plane as a frame, this problem becomes more noticeable because a large number of images must be generated continuously.

Non-patent 乂 ffl ^ l: M. Agrawala, D. Zorin, and T. unzner. Artistic multiprojection rena ering. In Eurographics Rendering Workshop 2000, pages 125-136, 2000.

Non-patent document 2: ϊ ". Kurzion and R. Yagel. Interactive space deformation with hardware -assisted rendering. IEEE Computer Graphics & Applications, 17 (5): 66-77, 1997. Non-patent document 3: P. Rademacher. View -dependent geometry.In Computer Graphics (Pr oceedings of Siggraph '99), pages 439-446, 1999.

Non-Patent Document 4: S. M. Seitzand C. R. Dyer. View morphing. In Computer Graphics (Proceedings of Siggraph '96), pages 2 丄 _30, 1996.

Non-Patent Document 5: K. Singh. A fresh perspective. In Proceedings of Graphics Interface 2 002, pages 17-24, 2002.

Non-Patent Document 6: S. Takahashi, N. Ohta, H. Nakamura, Y. Takeshima, and I. Fujishiro. Modeling surperspective projection of landscapes for geographical guide-map generation. Computer Graphics Forum, 21 (3): 259- 268, 2002.

Non-Patent Document 7: D. N. Wood, A. Finkelstein, J. F. Hughes, S. E. Thayer, and D. H. Sales.Multiperspective panoramas for eel animation.In Computer Graphics (Procee dings of biggraph '97), pages 243-250, 1997.

Patent Document 8: D. Zorin and A. H. Barr. And orrection of geometric perceptual distortio ns in pictures.In Computer Graphics (Proceedings of Siggraph '95), pages 257-264, 1995.

Disclosure of the invention

Problems to be solved by the invention

The present invention has been made in view of the above circumstances, and is a system, method, or computer that can generate a projection map that avoids obstruction of geographical features such as a route at a relatively high speed. It is intended to provide a program. Another object of the present invention is to provide a system for displaying a generated two-dimensional projection diagram.

Means for solving the problem

[0007] A projection map generation system according to the present invention includes a processing unit. The processing unit is configured to perform the following processing:

(1) Processing to extract feature points representing geographic features in a part that may be related to occlusion from a 3D terrain model;

(2) A process for calculating the optimum arrangement of the feature points on the two-dimensional projection plane when viewed from a certain viewpoint;

(3) A process of generating a 3D terrain shape that satisfies the optimal arrangement of the feature points on the 2D projection plane;

(4) A process of generating a two-dimensional projection when the three-dimensional shape is viewed from the viewpoint or the vicinity thereof.

[0008] The projection map generation system of the present invention may further include a storage unit. The storage unit stores the three-dimensional terrain model. In this case, the processing unit is configured to acquire the three-dimensional terrain model from the storage unit. Further, the processing unit is configured to store the generated two-dimensional projection view in the storage unit.

[0009] A projection map display system according to the present invention includes the above-described projection map generation system and a display unit. The display unit is configured to display the two-dimensional projection view.

[0010] In the projection map generation system, the processing unit generates a two-dimensional projection map at different time points by performing the processes (1) to (4) in response to the movement of the viewpoint. It may be. Further, when generating a two-dimensional projection map at a certain point in time, the processing unit uses a feature point that satisfies the following condition as a feature point in the processing of (1). May be:

(Condition)

Feature points used in the 2D projection generated before the certain time point, and exist in the view volume in the 2D projection map at the certain time point.

The optimal arrangement of feature points in the projection map generation system is, for example, the certain viewpoint On the two-dimensional projection plane when looking at the force, shielding of the line connecting the feature points on the road is avoided, and the relative positional relationship of each feature point is maintained.

[0012] The extraction of the feature points in the process (1) can be performed, for example, by the following process:

(a) The process of placing the viewpoint in a direction inclined at a certain angle with respect to the horizontal plane passing through the reference point;

(b) When the position of the viewpoint relative to the three-dimensional terrain shape obtained from the three-dimensional terrain model is relatively rotated around the reference point and in a horizontal direction, the three-dimensional Processing to extract all the vertices on the outline of the topographic shape as feature points.

[0013] A projection map generation method according to the present invention includes the following steps:

(1) extracting a feature point representing a geographic feature in a part that may be related to occlusion from a 3D terrain model;

(2) calculating an optimum arrangement of the feature points on a two-dimensional projection plane when viewed from a certain viewpoint;

(3) generating a 3D terrain shape that satisfies an optimal arrangement of the feature points on the 2D projection plane;

(4) A step of generating a two-dimensional projection view when the three-dimensional shape is viewed from the viewpoint or the vicinity thereof.

[0014] A computer program according to the present invention is for causing a computer to execute the steps in the above-described projection drawing generation method.

The invention's effect

[0015] According to the present invention, it is possible to provide a system, method, or computer program capable of generating a projection map in which the obstruction of geographical features such as a route is avoided at a relatively high speed. Best form for

Hereinafter, a first embodiment of the present invention will be described with reference to the accompanying drawings.

[0017] (Configuration of First Embodiment)

As shown in FIG. 1, the projection map generation system according to this embodiment includes a processing unit 1 and a storage unit. A unit 2, a display unit 3, and a communication path 4 are provided.

The processing unit 1 can be configured by a CPU, for example.

The processing unit 1 is configured to perform the following processing.

(4) A process of generating a two-dimensional projection view when the three-dimensional shape is viewed at the viewpoint or its nearby force.

[0020] Details of these processes will be described later as operations of the present embodiment. In addition, these processes can be performed by computer software executed on a computer (or CPU).

[0021] The storage unit 2 includes a 3D terrain model used for processing in the processing unit 1, computer software necessary for operating the processing unit 1, and a 3D terrain shape generated by the processing unit 1.・ Necessary data (information) such as 2D projection maps can be stored. The storage unit 2 can be configured by an internal storage device or an external storage device.

The processing unit 1 is configured to acquire a 3D terrain model from the storage unit 2. The processing unit 1 is configured to store the generated 3D topographic shape and 2D projection map in the storage unit 2.

[0023] The display unit 3 is configured to display the two-dimensional projection view generated by the processing unit 1. An example of the display unit 3 is a display.

[0024] The communication path 4 is a medium that enables exchange of information among the processing unit 1, the storage unit 2, and the output unit 3. The communication path 4 may be, for example, a network such as a power LAN that is a bus line in a computer or the Internet. That is, the processing unit 1, the storage unit 2, and the output unit 3 may be physically separated from each other and connected by a network. Means for connecting to the network, such as interface configuration and Since the protocol is well known, detailed description is omitted.

[0025] (Operation of First Embodiment)

Next, a method for generating a two-dimensional projection map using the above-described system will be described with reference to the flowchart of FIG. This system is implemented as a car navigation system. It is assumed that the 3D terrain model used in this system is represented by a monovalent function. In this embodiment, the monovalent function refers to a function in which, for example, the altitude is uniquely determined when the latitude and longitude are determined.

[0026] The system of this embodiment generates a bird's-eye view animation as shown in Fig. 3 when a route to be taken is given. In this system, the depression angle of the viewpoint is a force that can take a value from 20 degrees to 70 degrees, as in the case of a general force Navi, where the depression angle is set to 30 degrees so that the effect of shielding avoidance can be best confirmed. An example of setting is shown.

[0027] In this system, the portion where the path is blocked in each frame to be generated is the force that transforms the projection map so that it is avoided. Otherwise, the state of the perspective projection map is kept as much as possible. Create a non-perspective projection. Figure 3 is a snapshot taken by the implemented system, where the sphere is the current location, the dark line (thick line) is the road as the route, and the light line is not the route (that is, it is not the way to go). Represents. Fig. (A) is a perspective projection diagram in which a part of the route is shielded, and Fig. (B) is a display example of a non-perspective projection diagram in which shielding of the route is avoided.

[0028] The idea of avoiding the shielding of the driving route in the method of the present embodiment is briefly as follows. If, on the projection plane, the mountain outline 5 is on the road 6 as shown in Fig. 4 (a) and the shielding occurs, the mountain outline 5 as shown in Fig. 4 (b). Generates a two-dimensional projection that does not intersect with road 6. Such a change can be made by extracting the feature points of geological features such as mountains and roads, changing their arrangement on the projection plane, and reflecting the transformation on the topographic shape. This will be described in detail below. In Fig. 4, symbol F indicates road feature points, and symbol F indicates topographic feature points.

1 2

[0029] (Step Sa_ l in Figure 2)

In this step, the processing unit 1 extracts feature points that represent geographic features in parts that may be related to occlusion from the 3D terrain model. Where specifically features The number of feature points to be extracted is determined by the accuracy and usage of the 2D projection figure to be obtained, and computer restrictions. In general, in the case of a car navigation system, a point near the top of a mountain or a point for reproducing a road shape is selected. Furthermore, a point (vertex or bending point) for reproducing the outline of the building may be used as the feature point.

[0030] In this embodiment, from the contour line of the topographic curved surface, the peak F of the mountain, the bottom of the valley, and the end points thereof.

twenty one

F, intersection F between contour lines (see Fig. 5 (a)) is extracted as a feature point. This extraction

22 23

It can be easily calculated using viewpoint information and elevation information of the 3D terrain model. Since such a calculation method is conventionally known, detailed description thereof is omitted.

[0031] Furthermore, in this embodiment, the road force also has a point F, an end point F, and an intersection F at which the curvature is maximized.

11 12

Then, the inflection point F (Fig. 5 (b)) is taken out as a feature point. By connecting these feature points, the road

13 14

The shape can be roughly approximated. Note that a normal car navigation system often has road information as two-dimensional information separately from terrain curved surface information. In this embodiment, this two-dimensional information is also converted into a three-dimensional terrain model. It will be explained as conceptually included. In other words, the 3D terrain model may be a collection of a plurality of information including 2D information.

Here, in the present embodiment, it is not necessary to use the feature points on the terrain and the road in the entire 3D terrain model, and only the feature points in the view volume may be selected and used.

[0033] Furthermore, the feature points on the road illustrated in FIG. 5 (b) may be limited to feature points on the driving route and feature points on the road around the intersection. In order to avoid occlusion, only the feature points on the road as a route (points that can approximate the road by connecting them with straight lines) are sufficient. However, by extracting feature points on the road around the intersection, the road shape around the intersection can be prevented from being distorted.

An example of feature points automatically extracted by the system is shown in FIG. Black points (square points) in the figure are feature points. In the present embodiment, the feature point is specified by coordinate information. Information on the extracted feature points is also stored in the storage unit 2.

[0035] (Step Sa_ 2 in Figure 2)

In this step, the optimal feature points on the binary projection plane when viewed from a certain viewpoint The placement is determined using a spring model. The optimal arrangement here means an arrangement that keeps the original perspective projection as much as possible while avoiding the shielding of the road. To avoid the shielding of the road is to avoid the shielding of the line connecting the feature points on the road in this embodiment. Retaining the original perspective projection as much as possible means maintaining the relative positional relationship of the feature points in the present embodiment.

[0036] Such an optimal arrangement is obtained by obtaining the positional relationship of feature points in which occlusion is completely avoided, and approaching the arrangement of feature points in perspective projection using, for example, a spring model, The power to seek is S. The processing in this step is performed on the premise of the display state on the two-dimensional projection plane.

[0037] First, as a positional relationship in which shielding is completely avoided, a position {^} (ί = 1, ..., η) (see Fig. 7 (c)) where feature points are projected on a horizontal plane is used (n Is the number of feature points. This is because it is assumed that the topographic shape is a monovalent function, so that if it is projected onto the horizontal plane, it is guaranteed that no shielding will occur.

[0038] Next, in order to make the arrangement as close as possible to the arrangement of feature points {p} (i = l, ..., n) (see Fig. 7 (a)) in perspective projection, the following The spring model is constructed as follows. The meaning of the spring used in this model is only a virtual one for calculation by a computer.

[0039] First, Delaunay triangulation is applied to the feature point arrangement in Fig. 7 (c) (see Fig. 8 (a)), and the relative positional relationship of each feature point is maintained at the end point of each side. A spring is attached. In addition, a spring is also placed between p and q so that each feature point q approaches the corresponding perspective projection position pi. As an equilibrium state of this spring model, the optimum arrangement of feature points (Fig. 7 (b)) is obtained. Here, during the optimization process, check that the edges connecting the feature points do not intersect so that the relative positional relationship between the feature points on the road and the feature points on the terrain connected to it does not change. Therefore, it is possible to obtain a feature point arrangement that avoids occlusion (see Fig. 8 (b)).

[0040] In the implementation of the present embodiment, the spring force acting on each feature point is formulated as follows.

One ^-xj \-\ _qi -τ —Γΐ j A ^{Xt 1}

― (1) i

i €

[0041] In Equation (1), is the force acting on the feature point with index i, A is the set of feature point indexes adjacent to each other in the Delaunay triangulation, R is on the road, and T is the feature on the topography A set of point indices. In Equation (1), X is the position of the feature point, and when the spring model reaches the equilibrium state, its value converges to r in Fig. 7 (b). The first term on the right side is the force that approaches the perspective projection, and the second term is the force that maintains the relative positional relationship. Furthermore, the third term acts as a repulsive force between the feature points on the road and the feature points on the terrain. This repulsive force is introduced so that the geographical features such as roads and mountains are separated from each other so that they do not interfere with each other on the projection.

[0042] (Step Sa_ 3)

In this step, the position of the feature point (position in the three-dimensional space) is obtained so as to satisfy the optimum arrangement of the feature point on the two-dimensional projection plane obtained in step Sa_2. Using the position of this feature point as a constraint in the 3D space, the 3D terrain model is transformed. As a result, it is possible to generate a three-dimensional terrain shape that satisfies the feature point arrangement obtained on the two-dimensional projection plane.

[0043] However, the position of the feature point position in the three-dimensional space in the depth direction cannot be determined only from the feature point arrangement on the two-dimensional projection plane. In the present embodiment, as shown in FIG. 9, “the distance from the screen S (projection plane) to the feature point F on the 3D terrain model does not change” before and after the deformation. It has been converted. To do this

twenty four

Then, the position F (position in 3D space) of the feature point on the terrain after deformation is determined.

241

Deform the terrain model using the position as a geometric constraint. Here, the symbol F and the symbol F are attached to the feature points projected on the screen S.

24S 241S In the present embodiment, more specifically, the difference between the terrain shape before and after the deformation is expressed by a linear sum of unimodal basis functions using a Gaussian function or the like, and the position of the feature point on the terrain. The differential shape can be obtained by solving the simultaneous equations obtained from the constraints. By considering the differential shape in this way, deformation can be performed while maintaining the undulations of the high-frequency components that originally exist on the topographic curved surface. Such a deformation operation using a basis function is well known as such and will not be described in detail.

[0045] In this step as well, as in the case of feature point extraction, efficiency can be improved by limiting the deformation area. Actually, in order to cope with a sudden pan of the camera, it is preferable to perform deformation within the range of the view volume when the projection plane is about twice the size. FIG. 10 is viewed from a distance from the normal viewpoint, and the range indicated by the black dots shows the area that is actually deformed.

[0046] (Step Sa_4)

By looking at the 3D terrain shape generated as described above from the viewpoint, it is possible to generate individual static non-perspective projections (2D projections). In other words, it is possible to obtain a two-dimensional projection diagram in which shielding is avoided. In the obtained 2D projection map, shielding is basically avoided, so recalculation for shielding avoidance is basically unnecessary. Therefore, in the present embodiment, it is possible to obtain a two-dimensional projection diagram in which shielding is avoided at a relatively high speed. The obtained static non-perspective projection (two-dimensional projection) can be displayed on the display unit 3.

[0047] Furthermore, the processing unit 1 can generate two-dimensional projection diagrams at different points in time by performing the processing of each step described above corresponding to the movement of the viewpoint. If these figures are displayed in succession, an animation (movie) can be displayed as the display changes along the path (that is, in response to a change in viewpoint). This display can be performed by the display unit 3 in the example of the present embodiment.

[0048] However, even if the diagram generated in the above step is frame-by-frame, animation that retains coherence with respect to time cannot be generated. In order to maintain such coherence, two processes are performed in this embodiment.

First, feature points that are as identical as possible are selected between frames. This is This is realized by using the same feature point as it is if it exists in the view volume even in the current frame.

[0050] That is, when generating a two-dimensional projection map at a certain point in time, if there is a feature point that satisfies the following condition, it is used as a feature point in the process of step Sa_l.

(Condition)

A feature point used in a 2D projection created before a certain point in time, and exists in the view volume in the 2D projection at a certain point in time.

FIG. 11 shows a state in which feature points are taken in consecutive frames. Figure (a) shows that feature points are not reused, and it can be confirmed that a completely different feature point is selected as the frame advances by one. In this figure, the time axis goes from top to bottom in the figure. As can be seen from Figure (b), the reuse of feature points eliminates the difference in deformation caused by non-uniform feature points.

[0052] Further, by smoothing the differential shape representing the deformation required in a plurality of frames before and after, inter-frame coherence can be further maintained. For example, smoothing can be performed for each of the four frames before and after the weight of each frame obtained from a Gaussian function centered on the current frame.

[0053] (Experiment 1)

Based on the actual terrain and road data, a non-perspective projection animation was generated by the method of this embodiment described above. The calculation time per frame is about 3 seconds with CPU 3.0GHz Pentium (registered trademark) 4, memory 2GB RAM.

[0054] Fig. 3 (b) shows a part of the results. FIG. 4A is a perspective projection view.

According to the present embodiment, it is possible to avoid unnecessary shielding S by appropriately arranging the path on the projection view.

It should be noted that the description of each of the above embodiments is merely an example, and does not indicate a configuration essential to the present invention. The configuration of each part is not limited to the above as long as the gist of the present invention can be achieved.

[0056] For example, in the embodiment, in step Sa_2, the optimum arrangement of feature points is obtained. The spring model was used as a method for this. However, there are various other methods such as particles that approximate the force acting between feature points to the force between atoms. The particles themselves are a well-known method. In short, the specific content is not limited as long as it is a method capable of obtaining the optimum arrangement of feature points.

In the above embodiment, the method for preventing the shielding of the road has been described. However, it is also possible to prevent the shielding of features other than the road, such as buildings and geographical landmarks. In this case, the feature points should be placed on the contour of the object, preventing occlusion.

(Configuration and operation of the second embodiment)

In step Sa-l (see Fig. 2) of the first embodiment described above, from the contour of the topographic curved surface, the peak F of the mountain, the bottom of the valley, these end points F, and the intersection F of the contours (Fig. 5 ( see a))

21 22 23

Were extracted as feature points. This extraction was performed using viewpoint information and elevation information of the 3D terrain model.

[0059] Extraction of feature points in the second embodiment is performed as follows. Also in the present embodiment, extracting the peak points of the mountains, the bottoms of the valleys, their end points, and the intersection points of the contour lines (see FIG. 12) as feature points from the contour line of the topographic curved surface is the first embodiment. The form is basically the same. However, the feature points in the second embodiment are extracted at a plurality of viewpoint positions by moving the viewpoint P so as to surround the target terrain curved surface. Symbol R is attached to the movement trajectory of viewpoint P. Then, the feature point set obtained at each viewpoint position is summed, and the entire sum set is used as a feature point set on the terrain curved surface as a set of feature points independent of the viewpoint position. Since a method for extracting a contour line and a feature point thereon from a topographic curved surface has been conventionally known, a detailed description thereof will be omitted. Also, such feature points on the terrain curved surface can be extracted in advance and read as input together with the terrain curved surface data.

That is, feature point extraction according to the present embodiment is performed by the following processing:

(a) A process of placing the viewpoint P in a direction inclined at a certain angle with respect to a horizontal plane passing through the reference point. Here, the certain angle is not particularly limited, but in the case of a general car navigation system, it is in a range of 20 ° to 70 °. The reference point is a point that is always referred from the viewpoint P. For example, in the case of a car navigation system, it is the position of the own vehicle. The

(b) The contour of the 3D terrain shape when the position of the viewpoint relative to the 3D terrain shape obtained from the 3D terrain model is rotated relative to the reference point in the horizontal direction. Processing to extract all the vertices on the line as feature points. Here, the vertex means the vertex of the polyhedron that composes the 3D terrain model. The relative rotation angle of the viewing point around the reference point is generally the entire circumference (360 °) (see Fig. 12). In addition, as a relative movement method of the visual point, it may be moved continuously, but usually 15 ° to 30 °. It is preferable to move discretely for each angle. In short, any method for moving the viewpoint may be used as long as it can extract substantially necessary vertices.

Also in the present embodiment, points, end points, and intersections (FIG. 5 (b)) at which the curvature is maximized are extracted from the road as feature points.

An example of feature points automatically extracted by the system of the second embodiment is shown in FIG. Black points (square points) in the figure are feature points.

[0063] Further, in step Sa-2 of the first embodiment described above, from the position {q} (i = l, ..., n) (see FIG. 7 (c)) where the feature points are projected onto the horizontal plane. In order to get as close as possible to the arrangement of feature points { _P } (i = l, ..., n) in perspective projection (see Fig. 7 (a)), a spring model was used. This point is basically the same in the second embodiment, but in the second embodiment, the following method is used.

[0064] Also in the second embodiment, from the positions { _qi Ki = l, ..., _n ) (see Fig. 7 (c)) where the feature points are projected onto the horizontal plane, it is possible to avoid seeing the roads. Projection Arrangement {½ = 1,..., 11) (Building a spring model as close as possible to Fig. 7 (see)) However, in the second embodiment, in the process of finding the optimum arrangement of feature points , Each feature point moves on a straight line connecting {q} (i = l, ..., n) and {p} (i = l, ..., n) on the projection plane, The calculation is simplified.

[0065] The calculation of the optimal arrangement of feature points is performed in two steps. First, for the feature point arrangement in Fig. 7 (c),

(See Fig. 14 (&) (b)), and attach a spring to the end point of each side to maintain the relative position of each feature point. In addition, a spring is also placed between p and q so that each feature point q approaches the corresponding perspective projection position P. This spring model

i i i

As an equilibrium state, find the optimal feature point arrangement (Fig. 7 (b)). At this time, in the second embodiment, the feature on the road is allowed in the state where the intersection of the edges connecting the feature points causing the shielding of the road is permitted. Only the optimal arrangement of points is determined first (see Fig. 15 (a) and (b)).

[0066] After that, with the position of the feature point of the road fixed, the position of the feature point on the terrain curved surface is moved so as to avoid the intersection of the edges connecting the feature points, and the intersection is avoided. In this way, the calculation using the spring model is performed again, and the optimal arrangement of the feature points on the topographic curved surface is calculated (see Figs. 16 (a) and 16 (b)). As a result, it is possible to obtain a feature point arrangement in which shielding is avoided.

[0067] Further, in step Sa_3 of the first embodiment described above, the difference in topographical shape before and after deformation is expressed by a linear sum of basal functions that form a single peak using a Gaussian function, etc. By solving the simultaneous equations obtained from the point position constraints, the difference shape is obtained.

[0068] In the second embodiment, the terrain shape difference before and after deformation is represented hierarchically by a linear sum of basis functions that form a single peak using a B-spline function, and feature points on the terrain. The position constraints are applied to the coarse constraint, B-spline approximation, and then the fine force trace and B-spline approximation. As a result, the difference shape that satisfies the position constraints of the feature points can be obtained at high speed. By considering the differential shape in this way, it is possible to perform deformation while maintaining the undulation of the high-frequency component that originally exists on the topographic curved surface. Such deformation operations using basis functions are already known as shown in the following papers, so detailed explanations are omitted.

S. Lee, G. Wolberg, and S. Y. Shin, "Scattered Data Interpolation with Multilevel B -Splines," IEEE Transactions on Visualization and Computer Graphics, Vol. 3, No. 3, pp. 228—244, 1997.

[0069] Further, in step Sa-4 of the first embodiment, processing such as selecting the same feature point as much as possible between frames was performed in order to maintain coherence with respect to time. However, in the second embodiment, such processing is not necessary.

Since the configuration and advantages other than those described above in the second embodiment are the same as those in the first embodiment, description thereof will be omitted.

[0071] (Experimental example 2)

Based on the actual terrain and road data, a non-perspective projection animation was generated by the method of the second embodiment described above. The experimental conditions are the same as in Experimental Example 1, CPU 3.0GHz Pentium (registered trademark) 4, memory 2GB RAM. In this case, the calculation time per frame was approximately 0.5-1.0 seconds. That is, it can be seen that the processing speed can be increased by the method of the second embodiment. In the second embodiment, unlike the first embodiment, feature points can be extracted without depending on the viewpoint. For this reason, it is not necessary to perform recalculation for feature point extraction when the viewpoint is moved. Thereby, it is considered that the processing is speeded up.

[0072] As specific means of each unit (including functional blocks) for realizing each embodiment, hardware, software, network, a combination thereof, and any other means can be used. This is obvious to those skilled in the art. It is also possible to combine functional blocks into a single functional block. Further, the functional block may be realized by cooperation of a plurality of hardware or software.

Brief Description of Drawings

FIG. 1 is a block diagram showing a schematic configuration of a projection map generation system according to a first embodiment of the present invention.

FIG. 2 is a flowchart for explaining a projection map generation method according to the first embodiment of the present invention.

[Fig. 3] Fig. 3 (a) is an example of a perspective projection view in which a part of the route is hidden by a mountain. Figure (b) is an example of a non-perspective projection view in a state where the shielding of the route is avoided.

[FIG. 4] FIG. 4 (a) is a diagram schematically showing the arrangement of feature points where a route (road) is blocked. Figure (b) is a diagram schematically showing a state in which shielding is avoided by changing the arrangement of feature points.

[FIG. 5] FIG. 5 (a) is an explanatory diagram showing feature points on the topography. Figure (b) is an explanatory diagram showing feature points on the road.

[Fig. 6] An example of feature points extracted by the system is shown. Fig. (A) shows feature points on the terrain, and Fig. (B) shows feature points on the road.

[Fig. 7] An explanatory diagram for explaining the arrangement of feature points. Fig. (A) shows the arrangement on the perspective view, Fig. (B) shows the optimum arrangement, and Fig. (C) shows the arrangement projected on the horizontal plane. ing.

[Fig. 8] An explanatory diagram for explaining Delaunay triangulation of feature points. Fig. (A) shows the state before optimization, and Fig. (B) shows the state after optimization. FIG. 9 is an explanatory diagram for explaining the movement of the feature points in the three-dimensional space so as to satisfy the optimal arrangement on the projection plane.

[FIG. 10] An explanatory diagram showing the limited range of the deformation area (range represented by black dots), where FIG. (A) is an example of wire frame display and FIG. (B) is an example of surface display.

[Fig. 11] An explanatory diagram for explaining how feature points are taken. Fig. (A) shows no reuse of feature points, and Fig. (B) shows an example of reuse.

FIG. 12 is an explanatory diagram showing feature points on the topography in the second embodiment.

FIG. 13 (a) is a diagram showing an example of feature points on the terrain extracted by the system of the second embodiment.

FIG. 13 (b) is a diagram showing an example of feature points on the road extracted by the system of the second embodiment.

[FIG. 14 (a)] An explanatory diagram for explaining Delaunay triangulation of feature points, and FIG. (A) shows a state before optimization.

FIG. 14 (b) is an explanatory diagram for explaining Delaunay triangulation of feature points, and FIG. 14 (b) is an enlarged view of FIG.

FIG. 15 (a) is an explanatory diagram for explaining Delaunay triangulation of feature points, and FIG. 15 (a) shows an example of calculating the optimum arrangement of feature points on the road.

[FIG. 15 (b)] An explanatory diagram for explaining Delaunay triangulation of feature points, and FIG. 15 (b) is an enlarged view of FIG.

FIG. 16 (a) is an explanatory diagram for explaining Delaunay triangulation of feature points, and FIG. 16 (a) shows a state after optimization.

FIG. 16 (b) is an explanatory diagram for explaining Delaunay triangulation of feature points, and FIG. 16 (b) shows an enlarged view of FIG.

Explanation of symbols

1 Processing section

pLfe p | 5

3 Display section

4 communication path Mountain contour road

Claims

Claims [1] A projection map generation system comprising a processing unit, wherein the processing unit performs the following processing:

[2] The apparatus further includes a storage unit, the storage unit stores the 3D terrain model, and the processing unit is configured to acquire the 3D terrain model from the storage unit. Is configured to store the generated two-dimensional projection map in the storage unit

The projection map generation system according to claim 1, wherein:

[3] A projection diagram display system comprising: the projection diagram generation system according to claim 1 or 2; and a display unit, wherein the display unit is configured to display the two-dimensional projection diagram.

[4] The processing unit is configured to generate two-dimensional projection diagrams at different points in time by performing the processes (1) to (4) in response to the movement of the viewpoint.

When generating a two-dimensional projection map at a certain point in time, if there is a feature point that satisfies the following condition, it is used as a feature point in the processing of (1) above. :! ~ 3, displacement force Projection map generation system according to item 1:

(Condition)

[5] The optimal arrangement of the feature points means that on the two-dimensional projection plane when viewed from the certain viewpoint, shielding of lines connecting the feature points on the road is avoided, and the relative positions of the feature points are The projection map generation system according to any one of claims 1 to 4, wherein the relationship is maintained.

[6] The projection generation system according to claim 1, wherein the extraction of the feature points in the processing of (1) is performed by the following processing. :

[7] Projection map generation method characterized by comprising the following steps:

[8] A computer program for causing a computer to execute the steps according to claim 7.