High-speed super-resolution image stereoscopic visualization processing system and high-speed super-resolution image stereoscopic visualization processing program

Abstract

Provided is a super-resolution stereoscopic visualization image processing system having a high-speed processing function capable of obtaining a concave-convex image at a high speed with a detailed resolution. The present invention is provided with a base map database (110), a region definition unit (112), a super-resolution rasterizing processing unit (135), a moving average unit (134), a rectangular plane coordinate conversion unit (145), a super-resolution image generation unit (151), an X-direction adjustment unit (152), and a display processing unit (150). An interpolation process is performed, and after a specific number of moving averaging processes are applied to the post-interpolation elevation value (zri) of each super-resolution fine mesh (mbi) of the super-resolution square mesh (Mbi), a planar right-angle super-resolution mesh (Mdi) is generated and a square-adjusted super-resolution mesh (Mei) is generated. Thereafter, a super-resolution image is generated.

Description

Technical Field

[0001] The present invention relates to a high-speed super-resolution image stereoscopic visualization processing system. Background Art

[0002] In recent years, the Geospatial Information Authority of Japan (hereinafter referred to as the Geospatial Information Authority) has publicly released a Digital Elevation Model (DEM) on the Internet.

[0003] Using such a DEM, in recent years, a red stereoscopic map based on Patent Document 1 has been publicly released.

[0004] The outline of the red stereoscopic map is generated by obtaining the inclination, ground openness, and underground openness using a 5m DEM (Digital Elevation Model), calculating the peak-valley degree (also referred to as the floating and sinking degree) based on the ground openness, underground openness, and inclination, and then synthesizing by assigning red chroma to the inclination and brightness to the peak-valley degree.

[0005] However, since the red stereoscopic map is a raster image, when it is magnified to observe the terrain undulations in more detail, jagged edges (jagged borders) will occur.

[0006] That is, even when the red stereoscopic map is overlaid on the Geospatial Information Authority map and magnified, only jagged edges (jagged borders) can be seen, so the image is not clear.

[0007] A patent (Patent Document 2: Super-resolution Stereoscopic Visualization Processing System) that can solve this problem is disclosed.

[0008] The super-resolution stereoscopic visualization processing system of Patent Document 2 defines a mesh group of latitudes and longitudes in a specific area (for example, 1 km × 1 km) of the digital elevation model in plane rectangular coordinates (depending on the location, it will be a parallelogram, a trapezoid long in the longitudinal direction, or a rectangle).

[0009] After that, the division distance that equally divides each X-direction side of the mesh group in the plane rectangular coordinates into an odd number (excluding 1) is obtained.

[0010] After that, the two-dimensional plane (X-Y) corresponding to a specific area (for example, 1 km × 1 km) is divided by the division distance to define a super-resolution fine mesh (about 55 cm) of the size of the division distance in the two-dimensional plane (X-Y).

[0011] After that, a mesh group (5m × 5m) defining a rectangular coordinate system is defined in a two-dimensional plane (X-Y), and an interpolated elevation value (example of 9 divisions) obtained by interpolating the elevation values of the super-resolution fine meshes (about 55 cm) is extracted. A grid of this size is set as a smoothing grid, and a square moving average filter (smoothing mesh (5m × 5m)) composed of a smoothing grid group arranged in the vertical and horizontal directions by the odd number of such smoothing grids is generated.

[0012] After that, the super-resolution fine meshes (about 55 cm) defined in the two-dimensional plane (X-Y) are sequentially specified, and for each of the specified super-resolution fine meshes, the central smoothing grid of the square moving average filter (smoothing mesh (5m × 5m)) is set at the position of the super-resolution fine mesh, and a moving average filter (smoothing mesh (5m × 5m)) is defined in the two-dimensional plane (X-Y).

[0013] After that, a smoothed elevation value obtained by smoothing based on the interpolated elevation value group of the super-resolution fine mesh group at this moving average filter (smoothing mesh (5m × 5m)) is extracted, and this smoothed elevation value is assigned to the specified super-resolution fine mesh.

[0014] After that, each time the smoothed elevation value is assigned to the super-resolution fine meshes of the two-dimensional plane (X-Y), this super-resolution fine mesh is used as a point of interest, and for each of these points of interest, a consideration distance from this point of interest is defined by the number of super-resolution fine meshes corresponding to the division distance, and the floating and sinking degree within the number of super-resolution fine meshes is extracted, and a red three-dimensional visual processing for displaying this floating and sinking degree in gray scale (for example, a color in the red system) is performed.

[0015] Prior Art Documents

[0016] Patent Documents

[0017] Patent Document 1: Japanese Patent Publication No. 3670274

[0018] Patent Document 2: Japanese Patent Publication No. 6692984 Summary of the Invention

[0019] -Problems to be Solved by the Invention-

[0020] However, in the super-resolution visualization processing system of Patent Document 2, after converting the 5m DEM (square grid) defined by latitude and longitude into a plane rectangular coordinate (becoming a trapezoid, rectangle, etc.), for the grid defined by this plane rectangular coordinate, TIN bilinear interpolation is performed. Then, moving averaging is performed using a square moving average (smoothing) filter for smoothing, and a red stereoscopic image generation process is performed.

[0021] That is, the square grid (DEM) defined by latitude and longitude is converted into a plane rectangular coordinate (becoming a trapezoid, rectangle, etc.), and a square moving average filter is applied to the grid of this trapezoid, rectangle, etc.

[0022] In order to convert the latitude and longitude coordinates into a plane rectangular coordinate, complex calculations are required. However, the super-resolution visualization processing system of the prior art initially converts the 5m DEM of latitude and longitude into a plane rectangular coordinate, then performs interpolation (TIN bilinear interpolation), and performs moving average processing to obtain a super-resolution image.

[0023] Therefore, due to errors, as a result, it takes time (slow processing speed) until a super-resolution image is obtained. Especially if the area becomes larger, it will take more time.

[0024] The present invention is made in view of the above problems, and its object is to obtain a super-resolution stereoscopic visualization image processing system having a high-speed processing function capable of obtaining an uneven image at a detailed resolution at high speed.

[0025] -Means for Solving the Problem-

[0026] The high-speed super-resolution image stereoscopic visualization processing system according to the present invention includes: (A) a unit that, for each of a plurality of square grid groups in a specific area of a digital elevation model, obtains a super-resolved square grid in which the square grid is defined by a fine square super-resolution fine grid group; and

[0027] (B) a unit that, for each of the super-resolved square grids, performs interpolation processing and assigns an interpolated elevation value to each super-resolution fine grid of the super-resolved square grid; and

[0028] (C) a unit that, for each of the super-resolved square grids, performs a specific number of moving average processes on each super-resolution fine grid and updates the interpolated elevation value to a smoothed elevation value; and

[0029] (D) a unit that generates a planar rectangular super-resolution mesh in which the super-resolution square mesh after the unit (C) is defined by rectangular coordinates; and

[0030] (E) a unit that generates a square super-resolution stereoscopic image based on the planar rectangular super-resolution fine mesh of the planar rectangular super-resolution mesh.

[0031] -Advantages of the Invention-

[0032] As described above, according to the present invention, a super-resolution image can be obtained at high speed. In addition, even if the super-resolution image using DEM is enlarged, no sawtooth (jagged edge) will be seen, and the unevenness can be observed stereoscopically with detailed resolution. In addition, no lattice-like artifacts will be generated. Brief Description of the Drawings

[0033] Figure 1 It is a flowchart for explaining the outline of the high-speed super-resolution image stereoscopic visualization processing system of Embodiment 1.

[0034] Figure 2 It is an explanatory diagram of an image obtained by the high-speed super-resolution image stereoscopic visualization processing system of Embodiment 1.

[0035] Figure 3 It is a program block diagram of the high-speed super-resolution image stereoscopic visualization processing system of Embodiment 1.

[0036] Figure 4 It is a detailed flowchart (1) of the high-speed super-resolution image stereoscopic visualization processing system of Embodiment 1.

[0037] Figure 5 It is a detailed flowchart (2) of the high-speed super-resolution image stereoscopic visualization processing system of Embodiment 1.

[0038] Figure 6 It is an explanatory diagram of an image in which 5mDEM is downloaded by the display processing unit 150 and colors are added to the inclination for display.

[0039] Figure 7 It is an explanatory diagram of the 9×9 division of the super-resolution square mesh Mbi.

[0040] Figure 8 It is an explanatory diagram of the virtual super-resolution mesh Mbbi.

[0041] Figure 9 It is an explanatory diagram of TIN bilinear interpolation.

[0042] Figure 10It is an explanatory diagram of points during TIN bilinear interpolation processing.

[0043] Figure 11 It is an explanatory diagram of an example image of the result after bilinear interpolation.

[0044] Figure 12 It is an enlarged diagram explaining before and after bilinear interpolation.

[0045] Figure 13 It is an explanatory diagram of the reason for performing moving average processing.

[0046] Figure 14 It is an explanatory diagram of the moving average mesh.

[0047] Figure 15 It is an explanatory diagram (1) of the effect caused by moving average processing.

[0048] Figure 16 It is an explanatory diagram (2) of the effect caused by moving average processing.

[0049] Figure 17 It is an explanatory diagram of the DEM data after super-resolution smoothing processing.

[0050] Figure 18 It is an explanatory diagram of the elevation trajectory when not performing super-resolution image smoothing processing (moving average) and when performing it.

[0051] Figure 19 It is an explanatory diagram of the image after smoothing processing (9×9 box average).

[0052] Figure 20 It is an enlarged diagram explaining the effect of the first moving average.

[0053] Figure 21 It is an enlarged diagram explaining the effect of the second moving average.

[0054] Figure 22 It is an explanatory diagram of plane rectangular projection conversion.

[0055] Figure 23 It is an explanatory diagram of the image before plane rectangular coordinate conversion and after projection conversion.

[0056] Figure 24 It is an explanatory diagram of X-direction adjustment.

[0057] Figure 25 It is an explanatory diagram of the input screen for performing X-direction adjustment.

[0058] Figure 26Explanation diagram (1) of the square-adjusted super-resolution halftone Mei caused by the X-direction adjustment.

[0059] Figure 27 Explanation diagram (2) of the square-adjusted super-resolution halftone Mei caused by the X-direction adjustment.

[0060] Figure 28 Explanation diagram of the red stereoscopic image generated from the square-adjusted super-resolution halftone Mei.

[0061] Figure 29 Explanation diagram of the arrangement of the inclination after the smoothing process.

[0062] Figure 30 Explanation diagram (1) of the resampling in the projection conversion process of the present embodiment.

[0063] Figure 31 Explanation diagram of the resampling in the projection conversion process of the present embodiment.

[0064] Figure 32 Explanation diagram (3) of the resampling in the projection conversion process of the present embodiment.

[0065] Figure 33 Explanation diagram of the overall generation process of the red stereoscopic image.

[0066] Figure 34 Explanation diagram (1) of the generation process of the red stereoscopic image.

[0067] Figure 35 Explanation diagram (2) of the generation process of the red stereoscopic image.

[0068] Figure 36 Explanation diagram of the overall generation process of the red stereoscopic image of the mountain.

[0069] Figure 37 Block diagram of the program of the super-resolution image generation unit 151.

[0070] Figure 38 Schematic configuration diagram for explaining the convex part emphasizing image creation unit 11 and the concave part emphasizing image creation unit 12.

[0071] Figure 39 Schematic configuration diagram for explaining the inclination emphasizing unit 13.

[0072] Figure 40 Explanation diagram of the gray scale.

[0073] Figure 41 Explanation diagram of the calculation method of the super-resolution underground openness and the above-ground openness.

[0074] Figure 42 It is an explanatory diagram of the data structure of the super-resolution DEM.

[0075] Figure 43 It is an explanatory diagram of the super-resolution red stereo image generated by the super-resolution image visual processing system based on the prior art.

[0076] Figure 44 It is an explanatory diagram of the super-resolution image generated by the high-speed super-resolution image stereo visualization processing system according to this embodiment.

[0077] Figure 45 It is an explanatory diagram of the usage example.

[0078] Figure 46 It is a schematic configuration diagram of Embodiment 2.

[0079] Figure 47 It is an explanatory diagram of the generation of the smoothed contour information Ji.

[0080] Figure 48 It is an explanatory diagram of an image of an example in which the contour lines (vectors) of the smoothed contour information Ji are overlapped with the red image without smoothing processing.

[0081] Figure 49 It is Figure 48 an enlarged view of.

[0082] Figure 50 It is an explanatory diagram of an image of the result after smoothing the contour lines using the elevation value zhi.

[0083] Figure 51 It is a diagram obtained by synthesizing the contour lines of the 1:25,000 map with the red image generated based on the 10m DEM.

[0084] Figure 52 It is an explanatory diagram of an image synthesized with the super-resolution red image generated by the high-speed super-resolution image stereo visualization processing system according to Embodiment 1 of the present invention.

[0085] Figure 53 It is a schematic configuration diagram of the Lab color-added high-speed super-resolution image stereo visualization processing system of other embodiments.

[0086] Figure 54 It is the flowchart (1) of the Lab color-added high-speed super-resolution image stereo visualization processing system of other embodiments.

[0087] Figure 55 It is the flowchart (2) of the Lab color-added high-speed super-resolution image stereo visualization processing system of other embodiments.

[0088] Figure 56 It is a flowchart (3) of a Lab color - assigned high - speed super - resolution image three - dimensional visualization processing system of other embodiments.

[0089] Figure 57 It is an explanatory diagram caused by an image of the process of obtaining the Lab color red super - resolution image KLi.

[0090] Figure 58 It is a schematic configuration diagram of the Lab colorization unit 320.

[0091] Figure 59 It is an explanatory diagram of the spectral distribution.

[0092] Figure 60 It is a process diagram of Lab colorization.

[0093] Figure 61 It is a scatter diagram showing the relationship between the ground openness and the underground openness.

[0094] Figure 62 It is a diagram showing an example of a screen of the super - resolution Lab color image Li.

[0095] Figure 63 It is a diagram of an example (1) of a screen of the Lab color red super - resolution image KLi.

[0096] Figure 64 It is a diagram of an example (2) of a screen of the Lab color red super - resolution image KLi.

[0097] Figure 65 It is a schematic configuration diagram of other Embodiment 2.

[0098] Figure 66 It is an explanatory diagram when considering the geodetic system by reducing the resolution of the DEM. Specific Embodiment

[0099] In the present embodiment shown below, devices and methods for embodying the technical idea (structure, configuration) of the invention are exemplified. The technical idea of the present invention is not limited to the following examples. The technical idea of the present invention can be variously modified within the scope of the matters described in the claims. In addition, it should be noted that the drawings are schematic, and the configurations of devices, systems, etc. may be different from the actual ones.

[0100] In this embodiment, a process of obtaining a super-resolution stereoscopic visualization image Ki at high speed using, as an example, a base map (hereinafter referred to as the 5mDEM base map Fa) of a numerical elevation model of 5m DEM (A: A-series representative laser) of the Geospatial Information Authority of Japan will be described (alternatively, it may be 10m DEM, 20m DEM, 50m DEM, or 1m DEM).

[0101] The super-resolution stereoscopic visualization image Ki (also referred to as the super-resolution red stereoscopic map) also varies depending on the area, season, etc. as the object (blue, green, yellow-green, etc.), but in this embodiment, colors in the red system (red, purple, vermilion, orange, yellow, green, etc.) will be used for description. Additionally, in the case of the sea, lakes, rivers, etc., it is preferable to use colors in the blue system, brown system, or green system.

[0102] <Embodiment 1>

[0103] An overview of this Embodiment 1 will be described.

[0104] (1) Perform oversampling (super-resolution) by dividing the points of 5m DEM (0.2 seconds of equal latitude and longitude: also referred to as the 5m DEM grid) of the base map (DEM: numerical elevation model) of the Geospatial Information Authority of Japan into 9 equal parts (odd numbers, excluding 1). (The number of points becomes 81 times (the super-resolution fine grid is 8×8 = 64)).

[0105] After that, obtain the interpolated elevation value after interpolation of the super-resolution fine grid (square) by bilinear interpolation.

[0106] (2) Next, for all super-resolution fine grids, obtain a 9×9 (the super-resolution fine grid is 8×8 = 64) Box average (two-dimensional moving average process) to smooth it. This operation may be repeated multiple times as needed (ideally, until the jagged feeling disappears).

[0107] After that, (3) perform a projection transformation to a plane rectangular coordinate system and adjust it for the X direction to create a super-resolution red stereoscopic map. Additionally, considering the distance of the openness, adjust it in such a way that it becomes the same as the initial general 5m DEM grid.

[0108] In addition, DEM (Digital Elevation Model) is defined by allocating latitude, longitude, elevation, etc. to the grid, but in this embodiment, it is simply referred to as the grid, and the divided fine grid (also referred to as the fine lattice) is called the super-resolution fine grid.

[0109] In addition, the meaning of "oversampling (minimization) with odd numbers" varies depending on the setting method of representative points, resulting in different definitions.

[0110] For example, when representative points are assigned to one of the corners of a mesh, points between two points (in the latitude direction and longitude direction) are included and divided into odd numbers.

[0111] In addition, when the center of the mesh is used as the representative point, it is divided in such a way that the number of super-resolution fine meshes becomes odd. The case of assigning representative points to one of the corners of the mesh will be mainly described.

[0112] Figure 1 It is a flowchart for explaining the outline of the high-speed super-resolution image stereo visualization processing system of Embodiment 1. The high-speed super-resolution image stereo visualization processing system is also called a super-resolution image stereo visualization processing system with a high-speed processing function.

[0113] As shown in Figure 1 , the base map (5m DEM (A)) defined by latitude and longitude stored in the memory of the Geospatial Information Authority of Japan is extracted (S10).

[0114] 5m DEM (square) is a digital elevation model that divides the ground surface into squares (frames) at equal intervals of 5m (specifically, 5.5×10 -5 : 5.5E-5), and elevation values (Z) and other data are assigned at the center of each square.

[0115] After that, an arbitrary area Ei (for example, 1 km × 1 km) is specified (S20), and this area Ei is maintained in the state of latitude and longitude and defined by the square super-resolution fine mesh mbi, and for each super-resolution fine mesh mbi, a fine rasterization process of the elevation value obtained by bilinear interpolation is performed (S30).

[0116] This fine rasterization process is to obtain a group of super-resolution fine meshes mbi that divide the 5m DEM mesh into an even number of parts, and divide them with the number of division points DKi (3×3, 5×5, 7×7, or 9×9: also called the number of division points) that divides the two points (in the latitude direction and longitude direction) including the corners of this 5m DEM mesh into 9 parts (obtained while maintaining the state of latitude and longitude).

[0117] The width after being divided by the number of division points DKi is called the divided width da. For example, in the case of 9×9, it is 0.02 seconds in latitude and longitude, and it is equivalent to 0.55555 m (also called approximately 60 cm) in distance. It is obtained while maintaining the state of latitude and longitude.

[0118] After that, meshes of the size of 5 m DEM (hereinafter referred to as super-resolution square meshes Mbi) are sequentially defined starting from the reference point of the area Ei (for example, the origin or a corner of the area Ei).

[0119] After that, data such as latitude, longitude, and elevation values of the base map (5 m DEM (A)) are assigned to the four corners of these super-resolution square meshes Mbi (hereinafter collectively referred to as 5 m DEM points Mpij). The upper right corner is used as the representative value (it can also be the center of the mesh) for explanation.

[0120] After that, for each of the super-resolution square meshes Mbi, for each of the super-resolution fine meshes mbi in this super-resolution square mesh Mbi, the elevation value, latitude, and longitude of this super-resolution fine mesh mbi (hereinafter referred to as super-resolution fine mesh points Pij) are obtained by bilinear interpolation (also called interpolation) using the 5 m DEM points Mpij and assigned. The elevation value is called the elevation value zri after interpolation.

[0121] After that, a raster color addition process (S40) of "color addition based on the elevation value zri after interpolation" is performed for these super-resolution fine meshes mbi. In the present embodiment, the image in which the super-resolution fine mesh mbi has been color-added is also called a fine raster image mgi.

[0122] After that, for each of the super-resolution fine meshes mbi (fine raster images mgi), a moving average (S50) of the elevation value zri (zr1, zr2,...) after interpolation for the super-resolution fine mesh points Pij is respectively performed.

[0123] This moving average (Box Average) is applied using a moving average mesh Fmi (also called a moving average filter) defined by the number of division points DKi (3×3, 5×5, 7×7, or 9×9).

[0124] The elevation value zri of the super-resolution fine mesh points Pij of the super-resolution fine mesh mbi (fine raster image mgi) for which "the elevation value after this moving average is specified as the elevation value zhi after smoothing processing" is updated to this elevation value zhi after smoothing processing (referred to as super-resolution smoothing processing).

[0125] Thereafter, if moving average processing is applied to all the super-resolution fine meshes mbi (fine grid images mgi) of the region Ei, these groups are displayed on the screen as the moving average processed fine grid images GHi (S60). Details will be described later.

[0126] Thereafter, the operator determines whether the moving average processed fine grid image GHi on the screen has become smooth (whether the vignetting is appropriate). When it has not become smooth, an instruction to perform smoothing processing for super-resolution is input, and the processing of step S50 is performed again.

[0127] In addition, when it is determined that it is smooth, planar rectangular coordinate conversion processing is performed for resampling (S90). After adjusting the X direction to be square, resampling is performed to perform super-resolution red stereo image generation processing (S100), and it is displayed on the screen (S110).

[0128] That is to say, as an initial process, for a 5m DEM mesh of a square defined by latitude and longitude, without performing planar rectangular coordinate conversion (e.g., Mercator), a fine (super-resolution) square mesh (super-resolution fine mesh) is defined while maintaining the latitude and longitude values, and TIN (triangulated irregular network) bilinear interpolation and moving average processing are performed. Thereafter, planar rectangular projection conversion is performed, the X direction is adjusted to be square, resampling is performed, and red stereo image generation processing is performed.

[0129] That is to say, since the process of converting latitude and longitude coordinates to planar rectangular coordinates is not performed in the initial process, no error is generated. Therefore, the time consumed until a super-resolution image is obtained is reduced (the processing speed is fast).

[0130] By this, as shown in (a) of Figure 2 , if the image of the 5m DEM processed as in the prior art is enlarged, the fine mesh groups (mbi) are displayed in a jagged manner. However, as shown in (b) of Figure 2 , in the present embodiment, even when enlarged, it becomes a smooth image. In addition, the super-resolution red stereo image generation processing will be described later.

[0131] Figure 3 It is a program block diagram of the high-speed super-resolution image stereoscopic visualization processing system of Embodiment 1.

[0132] As in Figure 3As shown, the high-speed super-resolution image stereoscopic visualization processing system 300 of Embodiment 1 is composed of a computer main body unit 100, a display unit 200, etc.

[0133] The computer main body unit 100 includes: a base map database 110 that stores a 5m DEM base map Fa, a region definition unit 112, a super-resolution rasterization processing unit 135, a moving average unit 134, a consideration distance grid number calculation unit 148, a plane rectangular coordinate conversion unit 145, a super-resolution image generation unit 151, an X-direction adjustment unit 152, a display processing unit 150, etc.

[0134] The super-resolution rasterization processing unit 135 includes a 5m DEM odd-number division unit 115, a TIN bilinear interpolation unit 137, a raster color addition processing unit 132, etc.

[0135] (Explanation of each part)

[0136] The region definition unit 112 reads the 5m DEM grid Mai (latitude, longitude, elevation, 5m frame) corresponding to the region Ei (for example, 50m to 1500m in length and width) specified by the operator from the 5m DEM numerical model in the base map database 110 into the memory 118.

[0137] The 5m DEM odd-number division unit 115 of the super-resolution rasterization processing unit 135 divides the sides in the latitude direction (hereinafter referred to as the latitude direction) and the longitude direction (hereinafter referred to as the longitude direction) of the 5m DEM grid Mai (Mai: 5m or 10m) of the region Ei in the memory 118, which is a square grid, by an odd number (excluding 1: 9×9), and sequentially generates a super-resolution square grid Mbi having a group of super-resolution fine grids mbi that are square.

[0138] The TIN bilinear interpolation unit 137 copies the square super-resolution square grid Mbi to the memory 142.

[0139] After that, for each of these super-resolution square grids Mbi (latitude and longitude), TIN bilinear interpolation (interpolation processing) is performed, and an interpolated elevation value zri is assigned to each super-resolution fine grid mbi of the super-resolution square grid Mbi.

[0140] The raster color addition processing unit 132 assigns a color value based on the interpolated elevation value zri in the memory 142, displays it on the screen through the display processing unit 150 described later, and activates the moving average unit 134.

[0141] The moving average unit 134 performs a moving average process a specific number of times (using a 9×9 moving average mesh) for each of the super-resolution square meshes Mbi in the memory 142, and updates the interpolated elevation value zri to the elevation value zhi after this smoothing process.

[0142] The planar rectangular coordinate conversion unit 145 defines the super-resolution square mesh Mbi after moving average in the memory 142 by planar rectangular coordinates, and generates this as the planar rectangular super-resolution mesh Mdi in the memory 149.

[0143] This planar rectangular super-resolution mesh Mdi becomes a square, rectangle, trapezoid, etc. However, in this embodiment, it is mainly described as a square.

[0144] The X-direction adjustment unit 152 generates, in the memory 153, a square-adjusted super-resolution mesh Mei (also referred to as a super-resolution mesh after square conversion) obtained by adjusting the planar rectangular super-resolution mesh Mdi (in the memory 149) to a square.

[0145] The super-resolution image generation unit 151 designates the square-adjusted super-resolution mesh Mei (super-resolution mesh after square conversion) in the memory 153, and in each of the designated ones, sequentially designates the adjusted fine meshes mei of the square-adjusted super-resolution mesh Mei as the points of interest.

[0146] After that, based on the elevation value zhi after the smoothing process, the inclination between the adjusted fine meshes mei adjacent to the adjusted fine mesh mei as the point of interest is obtained and assigned to the adjusted fine mesh mei of the point of interest.

[0147] In addition, the number of super-resolution fine meshes (hereinafter referred to as the number of super-resolution fine meshes considering distance) from the calculation unit 148 for considering the distance grid number is read, and within this number of super-resolution fine meshes considering distance, the peak-valley degree (also referred to as the floating and sinking degree) between the adjusted fine meshes mei adjacent to the point of interest is obtained, and the gradation color value (red-based color) representing the combination of this peak-valley degree and the inclination is assigned to the adjusted fine mesh mei of the point of interest.

[0148] After that, these data are memorized in the memory 153. That is to say, in the memory 153, there is memorized a super-resolution DEM which is a set of "super-resolution DEM data composed of region Ei, super-resolution square mesh Mbi, super-resolution fine mesh mbi (number), division width da, elevation value zri after bilinear interpolation, elevation value zhi after smoothing process, individual inclination of each super-resolution fine mesh mbi, color value of the inclination, color value of the floating and sinking degree (ground opening degree, underground opening degree), etc.".

[0149] The consideration distance grid number calculation unit 148 calculates the number of super-resolution fine meshes converted within the consideration distance L (for example, 50 m) input as the point of interest. For example, it outputs the super-resolution fine meshes equivalent to L / da to the super-resolution image generation unit 151.

[0150] The display processing unit 150 has a display memory (not shown), reads in data corresponding to the input image type into the display memory, and displays an image (super-resolution stereoscopic visualization image) with the color value assigned to this data on the screen of the display unit.

[0151] In addition, the super-resolution image generation unit 151 can also attach a color value to the planar rectangular super-resolution fine mesh mdi of the planar rectangular super-resolution mesh Mdi in the memory 149 and display it as a super-resolution stereoscopic visualization image through the display processing unit 150.

[0152] (Action description)

[0153] Figure 4 And Figure 5 is a detailed flowchart of the high-speed super-resolution image stereoscopic visualization processing system of the first embodiment.

[0154] In the base map database 110, a 5m DEM base map Fa (terrain) is memorized (S200).

[0155] The 5m DEM of the 5m DEM base map Fa is a point cloud obtained by aerial laser (at intervals of several tens of centimeters), and the area of this point cloud is the whole of Japan (several tens of kilometers to several hundreds of kilometers).

[0156] These point clouds include latitude, longitude, elevation value, intensity, etc. In this embodiment, these are simply called 5m DEM points, and the four corners of the frame of the 5m DEM are called 5m DEM four-corner points Maq (q: a, b, c, d).

[0157] In addition, in this embodiment, the 5m DEM four-corner points Maq (q: a, b, c, d), 5m DEM points, and 5m mesh frames are collectively called 5m DEM meshes Mai (square).

[0158] The area definition unit 112 designates the area corresponding to the area Ei (for example, 50 m to 1500 m, 2000 m, ··· 5000 m, ··· 10000 m, ··· in length and width) input (designated) by the operator for the 5 m DEM numerical model of the base map database 110, and reads the 5 m DEM grid Mai (latitude, longitude, elevation, 5 m box) of the designated area Ei into the memory 118 (S210, refer to Figure 6 )

[0159] However, Figure 6 It is an image to which color is added to the inclination and displayed by the display processing unit 150. PMoi is an example in which a representative value is set at the center of the 5 m DEM grid Mai.

[0160] That is, in the memory 118, the 5 m DEM grid Mai (latitude, longitude, elevation, box) is defined. In addition, the memory 118 defines the X-axis by longitude and the Y-axis by latitude.

[0161] Specifically, latitude and longitude are output to the XY file. The latitude direction is the Y direction, and the longitude direction is the X direction. However, it is simply described as the latitude direction and the longitude direction. In addition, in order to distinguish from the plane rectangular coordinates, there are cases where it is shown in the figure as "「i」 represents the latitude direction (Y direction), and 「j」 represents the longitude direction (X direction)".

[0162] Next, the super-resolution rasterization processing unit 135 performs fine rasterization processing. (Super-resolution rasterization processing unit 135)

[0163] The 5 m DEM odd-number division unit 115 of the super-resolution rasterization processing unit 135 successively generates, in the memory 118, super-resolution square meshes Mbi divided by the number of division points Dki (3×3, 5×5, 7×7, or 9×9) for obtaining a group of super-resolution fine meshes mbi obtained by dividing the 5 m DEM mesh Mai in the memory 118 according to the input number of division points Dki (3×3, 5×5, 7×7, or 9×9) and the type of DEM (in this embodiment, it is described as 5 m DEM), etc. (S230).

[0164] However, Figure 6 It is an image (magnified image) to which color is added to the inclination and displayed by the display processing unit 150.

[0165] In addition, when described in terms of the divided wide width da, it is approximately 0.02 seconds in terms of latitude and longitude (for example, when it is 9×9, it is equivalent to 0.55555 m).

[0166] In the present embodiment, the 5 m DEM mesh Mai that has been super-resolved is referred to as the super-resolved square mesh Mbi, and the mesh of the da size is referred to as the super-resolution fine mesh mbi.

[0167] In Figure 4 S230 and Figure 7 the super-resolved square mesh Mbi is presented.

[0168] In Figure 7 the four corner points of this super-resolved square mesh Mbi are referred to as the super-resolved square mesh corner representative points Mpq (Mpa, Mpb, Mpc, Mpd).

[0169] After that, for each of the four corners of the super-resolution fine mesh mbi, points such as latitude, longitude, and elevation values are assigned using the virtual super-resolution mesh described later (see Figure 8 ).

[0170] In the present embodiment, this is referred to as the super-resolution fine mesh point Pij. "i" represents the latitude direction (X direction), and "j" represents the longitude direction (Y direction).

[0171] In Figure 8 the super-resolved square mesh corner representative points Mpq (Mpa, Mpb, Mpc, Mpd) are recorded as the super-resolution fine mesh points Pij ((P1, 1), (P1, 9), (P9, 1), (P9, 9)), and the four corner super-resolution fine mesh points Pij that define the super-resolution fine mesh mbi adjacent to (P1, 9) are recorded as (P1, 9), (P2, 9), (P1, 10), (P2, 10).

[0172] That is, Mpa is (P1, 1), Mpb is (P1, 9), Mpc is (P9, 1), and Mpd is (P9, 9). In addition, PMoi is an example where the center of the super-resolved square mesh Mbi is used as a representative (elevation value) (referred to as the super-resolved square mesh center representative point PMoi).

[0173] After that, a process of virtually generating a 10×10 mesh (hereinafter referred to as the virtual super-resolution mesh Mbbi) (the size is equivalent to 0.02 seconds) in the uncharted memory is performed (also referred to as the virtual 10×10 mesh generation process) (S240: see Figure 8 ).

[0174] The generation process of the virtual 10×10 mesh is described.

[0175] The representative value (elevation) of the super-resolved square mesh Mbi is obtained by averaging the points at the four corners of the mesh. Therefore, if the points (elevations) at the four corners are not known, it cannot be defined.

[0176] Therefore, the generation process of the virtual 10×10 mesh is performed.

[0177] The generation process of the virtual 10×10 mesh generates the virtual super-resolved fine mesh Mbbi (10×10 is the number of dividing lines, and the number of super-resolved fine meshes is 9×9) (indicated by dotted lines) shown in Figure 8 for each super-resolved square mesh Mbi.

[0178] In Figure 8 the representative points at the four corners of the virtual super-resolved mesh Mbbi are denoted as Mqa, Mqb, Mqc, and Mqd. In addition, the points of the virtual fine mesh mbbi of the virtual super-resolved mesh Mbbi are denoted as (Pa1,1), (Pa1,2), ··· (Pa1,11), ···, (Pa11,1), (Pa11,2), ··· (Pa11,11). These (Pa1,1), (Pa1,2), ··· (Pa1,11), ···, (Pa11,1), (Pa11,2), ··· (Pa11,11) are collectively referred to as virtual fine mesh points (Pai,j).

[0179] After that, the process of determining the representative value of the 5mDEM mesh after super-resolution (S250) is performed.

[0180] First, the super-resolved square mesh corner representative points Mpq (Mpa, Mpb, Mpc, Mpd) at the four corners of the super-resolved square mesh Mbi are determined (calculated).

[0181] Specifically, for example, Figure 8 the super-resolved square mesh point MPb (elevation) at the upper right corner of the super-resolved square mesh Mbi as shown in

[0182] is used to determine the super-resolved square mesh corner representative point Mpa based on the elevation values of the virtual fine mesh points (Pa1,10), (Pa1,11), and (Pa2,10) of mbb10 of the virtual super-resolved mesh Mbbi. Figure 5 After that, the TIN bilinear interpolation unit 137 performs the TIN bilinear interpolation process (S260) as shown in

[0183] The TIN bilinear interpolation process (S260) copies the super-resolution square mesh Mbi of the square in the memory 118 and the associated data to the memory 142. After that, the super-resolution fine meshes mbi of this super-resolution square mesh Mbi are sequentially specified.

[0184] After that, the super-resolution fine mesh representative point Pqij (elevation) of this specified super-resolution fine mesh mbi is calculated by TIN bilinear interpolation (interpolating the elevation value) based on the "super-resolution square mesh corner representative point Mpq of the virtual super-resolution mesh Mbbi", the "super-resolution fine mesh point Pij of the specified super-resolution fine mesh mbi of the super-resolution square mesh Mbi", and the "super-resolution square mesh central representative point PMoi", etc. (see Figure 9 ).

[0185] Figure 9 is an explanatory diagram of TIN bilinear interpolation.

[0186] In addition, Figure 9 is an example where the center of the super-resolution fine mesh mbi is set as the super-resolution fine mesh representative point Pqij. In addition, Figure 10 is an explanatory diagram of the points during the TIN bilinear interpolation process.

[0187] This Figure 10 shows the situation of the super-resolution square mesh Mbi and the virtual super-resolution mesh Mbbi during the TIN bilinear interpolation process. However, colors are added to the elevation.

[0188] In addition, the elevation value after TIN bilinear interpolation of the super-resolution fine mesh representative point Pqij is called the bilinear interpolation after elevation value zri (zr1, zr2,..., also called the interpolated elevation value).

[0189] In Figure 10 the frames of the super-resolution square mesh Mbi and mbi are recorded in the area of a part of the place (for example, 5m × 5m).

[0190] Figure 11 is an example image of the result after bilinear interpolation. If compared with Figure 10 as a whole, the colors (the darker the color, the darker the vermilion) are scattered.

[0191] Figure 12 is an enlarged view for explaining before and after bilinear interpolation. Figure 12 The (a) of Figure 12In (b), after bilinear interpolation, it is displayed with color added for elevation. As shown in Figure 12 (a), the mbi has a jagged feeling. However, as shown in Figure 12 (b), overall, the colors are scattered.

[0192] After that, when the bilinear interpolation process for all the super-resolution fine meshes mbi at all the super-resolution square meshes Mbi in the memory 142 has been performed, the rasterization color addition processing unit 132 is started.

[0193] The rasterization color addition processing unit 132 sequentially designates the super-resolution fine meshes mbi in the memory 142, and at each designated one, reads in the bilinear interpolation elevation values zri (zr1, zr2,...), and assigns the color values corresponding to these elevation values to the super-resolution fine mesh mbi (S270).

[0194] However, since the super-resolution fine mesh mbi is a fine mesh, it becomes an image with a jagged feeling and so-called noise. Therefore, the correlation between each data point is increased to make the data smoothly connected, and the influence of singularities or noise without correlation is removed, and a moving average process (for example, a Kalman filter) is performed (S280).

[0195] For example, when bilinear interpolation is performed on the 5m DEM shown in Figure 13 (a), as shown in Figure 13 (b), the bilinear interpolation elevation value zri will shoot up sharply (the part of hui) or the value will drop sharply at the valley part (the part of hdi).

[0196] (Moving average)

[0197] The moving average unit 134 generates a moving average mesh Fmi with the number of segmentation points DKi (for example, 3×3, 5×5, 7×7 or 9×9) input by the operator in the memory 117 (refer to Figure 14 ). In the present Embodiment 1, the number of segmentation points DKi is set to 9×9 for illustration.

[0198] In addition, Figure 14 it is described that the vertical row of the moving average mesh Fmi is set to "i: latitude" and the horizontal row is set to "j: longitude" for the mesh number fm(i, j).

[0199] In addition, as shown in Figure 14As shown in (b) thereof, the moving average mesh Fmi (also referred to as a filter) may divide the number of division points DKi into 11×11 (one mesh, the size of which corresponds to 0.02 seconds) (denoted as Fmb, dotted line).

[0200] In addition, the moving average value (weighted average) at the center mesh is referred to as the elevation value zfi after smoothing processing (also referred to as the elevation value zfi after moving average) (also referred to as the smoothed elevation value), and the value of the specified super-resolution fine mesh mbi is updated to this elevation value zfi after smoothing processing.

[0201] After that, the grid color addition processing unit 132 is started, and the color (caused by the color scale) corresponding to the elevation value zfi after smoothing processing of the memory 142 is assigned to the super-resolution fine mesh mbi of the memory 142, and is displayed on the screen of the display unit 200 through the display processing unit 150 (S290).

[0202] In the present embodiment, the image thus displayed on the screen is referred to as the super-resolution image GZi.

[0203] After that, the operator determines whether the super-resolution image GZi on the screen is the desired smooth image (S300).

[0204] When it is not smooth, a re-smoothing process instruction for super-resolution is input, and the moving average unit 134 performs the process of step S280 again through this re-smoothing instruction.

[0205] Through this moving average processing, as shown in (b) of Figure 15 the part where the bilinear interpolation elevation value zri spurts sharply (the hui part) disappears, and the part where the value of the valley part drops sharply (the hdi part) also disappears. That is, it becomes smooth.

[0206] Specifically, Figure 16 the bilinear interpolation elevation value zri in (a) of Figure 16 becomes 9, 9, 9,... 1320, 10, 10, 10,... 10 as shown in (b) of

[0207] That is to say, the memory 142 stores the super-resolution smoothed DEM data RGi that is the source of the super-resolution image GZi shown in Figure 17 The super-resolution smoothed DEM data RGi is as shown in

[0208] The super-resolution smoothed DEM data RGi is as shown in Figure 17As shown in , it is composed of the region Ei, the super-resolution square mesh Mbi, the super-resolution fine mesh mbi (number), the divided width (for example, the curved surface width corresponding to 0.555 m), the elevation value zri after bilinear interpolation, the first-order smoothed fine elevation value zfi, and the second-order smoothed fine elevation value zfi', etc.

[0209] In addition, the smoothed fine elevation value zfi and the second-order smoothed fine elevation value zfi' are also collectively referred to as the smoothed processing values.

[0210] Furthermore, in the present embodiment, zri, zfi, zfi',... are collectively referred to as the elevation values zhi after smoothed processing.

[0211] In Figure 18 the trajectory of the elevation in the case where the smoothed processing (moving average) for the super-resolution image is not performed is shown.

[0212] As shown in Figure 18 when the moving average is not performed, since the curvature maximization processing (spline curve, Bezier curve, etc.) is performed, for example, the trajectory connecting the point A1 and the vertices A2 and A3 becomes a straight line Lai (represented by a solid line). However, in the present embodiment, due to the moving averaging process, the line (Lbi) connecting the elevation values zhi after smoothed processing becomes a point Ap that does not pass through the vertex A2 (it will become lower if the moving average is repeatedly performed).

[0213] In addition, Figure 18 between A1 and A2, and between A2 and A3, it is described as the width of the 5m DEM. In addition, the super-resolution fine mesh mbi is recorded as mb1, mb2,... mb8,... mb16 starting from A1.

[0214] Furthermore, when it is determined that it is a smoothed image, the input smoothed image (the super-resolution smoothed processing, the fine grid image GHi after moving average) is "OK".

[0215] In Figure 19 an example (enlarged) of visualizing the smoothed processing (also called 9×9 box average processing) is shown. Figure 19 In a color corresponding to the inclination. Figure 20 It is an enlarged view for explaining the effect of the first moving average, Figure 21 and it is an enlarged view for explaining the effect of the second moving average.

[0216] As shown in Figure 20 (a) thereof, there is a sawtooth edge before the moving average, but after the first moving average, as shown inFigure 20 As shown in (b) thereof, it becomes a smooth image in which the serrated edges are suppressed.

[0217] Furthermore, at the second time, the image after the first moving average ( Figure 21 in (a)) becomes even smoother.

[0218] After that, when the smooth image (super-resolution smoothing process) is "OK" (judged by the operator), the moving average unit 134 starts the plane rectangular coordinate conversion process, and then the plane rectangular coordinate conversion unit 145 performs the projection conversion process (plane rectangular coordinate conversion) (S320).

[0219] The projection conversion process (S320) converts the super-resolution fine mesh point Pij at the super-resolution fine mesh mbi of the super-resolution square mesh Mbi (latitude and longitude) allocated to the memory 142 to the plane rectangular coordinates, and outputs it as the plane rectangular point Pbij to the XYZ point file for plane rectangular coordinates (stored in the memory 149, refer to Figure 22 ).

[0220] The details of this projection conversion process will be described later.

[0221] The plane rectangular coordinate conversion is the "equiangular cylindrical projection method" in which the earth is placed in a cylinder where only the equator of the earth touches, and the longitude and latitude lines are first projected onto the cylinder, and then the cylinder is opened. The interval between the latitudes becomes wider as it approaches the poles.

[0222] Therefore, when converting to the plane rectangular coordinates, due to the distortion, depending on the location, it will become an inclined rectangle or a rectangle without distortion (there may also be a square case).

[0223] That is to say, in Figure 22 the square super-resolution square mesh Mbi (latitude and longitude coordinates) shown in (a) becomes the plane rectangular super-resolution mesh Mdi shown in Figure 22 the (b) or Figure 22 the (c).

[0224] The plane rectangular super-resolution mesh Mdi is composed of the plane rectangular points Pbij of (P1, 1),..., (Pb9, 9), and its shape mostly becomes a trapezoid or a rectangle (depending on the location, it may also be a square). In addition, the super-resolution fine mesh of the plane rectangular super-resolution mesh Mdi is called the plane rectangular super-resolution fine mesh mdi.

[0225] Specifically, for example, it becomes as follows.

[0226]

[0227] Use Figure 23 for supplementary explanation. In Figure 23 (a) of Figure 23 , an image before the conversion of the rectangular coordinates (also called the projection conversion) is shown. In Figure 23 (a) of Figure 23 and Figure 23 (b) of Figure 23 , examples of color addition for the elevation value are shown. As shown in

[0228] Next, the super-resolution image generation unit 151 reads the rectangular super-resolution fine mesh mdi of the rectangular super-resolution mesh Mdi in the memory 149 (S330), performs super-resolution image stereoscopic visualization processing (S340), and reads this image into the display memory and displays it on the screen of the processing unit 200 (S350).

[0229] Before performing the above-mentioned super-resolution image stereoscopic processing (S340), although not shown in the flowchart, the X-direction adjustment unit 152 performs X-direction adjustment processing.

[0230] (X-direction adjustment processing)

[0231] The X-direction adjustment unit 152 generates a square-adjusted super-resolution mesh Mei in the memory 153 by adjusting the rectangular super-resolution mesh Mdi (e.g., rectangle, trapezoid) in the memory 149 to a square (refer to Figure 24 ). As shown in Figure 24 (d) of

[0232] That is, the width is adjusted so that the rectangular super-resolution mesh Mdi becomes a square. This is called the square-adjusted super-resolution mesh Mei.

[0233] Specifically, the width of the Y-direction (side) of the square-adjusted super-resolution mesh Mei in the memory 149 is set to be the same as the width of the X-direction (side). That is, the X-direction (side) of the rectangular super-resolution mesh Mdi is moved upward (+ direction) so that the Y-direction (side in the longitude direction) of the rectangular super-resolution mesh Mdi becomes equal to the width of the X-direction (side in the latitude direction) of the rectangular super-resolution mesh Mdi. This is called the adjustment in the X-direction.

[0234] In addition, through this adjustment, the X direction (the side in the latitude direction) of the planar rectangular super-resolution fine mesh mdi is moved upward (+ direction), and the side in the Y direction (longitude) of the planar rectangular super-resolution fine mesh mdi is made equal to the side in the X direction (latitude direction) of the planar rectangular super-resolution fine mesh mdi. That is to say, the planar rectangular super-resolution fine mesh mdi becomes a square. This is called the adjusted fine mesh mei.

[0235] That is to say, in a rectangular coordinate system where the width in the Y direction (j: longitude) of the planar rectangular super-resolution fine mesh mdi of the planar rectangular super-resolution mesh Mdi (for example, rectangle, trapezoid) is made to match the point interval (0.5555 m: about 60 cm) in the X direction (i: latitude) of the planar rectangular super-resolution fine mesh mdi, the data (elevation) of the planar rectangular super-resolution fine mesh DEM of the square adjusted super-resolution mesh Mei is resampled (re-sampled) through projection conversion processing (S320).

[0236] Specifically, a screen as shown in Figure 25 is displayed and adjusted. If the inputs are ga (X-axis: 0.561056071515711) and gb (Y-axis: 0.684808514623378), then through the X-direction adjustment processing of the computer, it becomes gaa (X-axis: 0.561056071515711) and gbb (Y-axis: 0.561056071515711) of the following screen.

[0237] That is to say, it becomes a square. As shown in Figure 26 , the square adjusted super-resolution mesh Mei is formed by 4 large red circles, and the adjusted fine mesh mei is formed by 4 small red circles (Pai, j), and it becomes a square.

[0238] The fine mesh of this square adjusted super-resolution mesh Mei is called the adjusted fine mesh mei.

[0239] Using Figure 27 and Figure 28 , the effect of the image obtained by the super-resolution image generation unit 151 described later using this square adjusted super-resolution mesh Mei is described.

[0240] As shown in Figure 27 , Figure 26 the large ● (large red circle) is removed (not displayed). Figure 28 It is an image obtained by the super-resolution image generation unit 151 and becomes an image without jagged edges and without jagged sides.

[0241] That is, in the memory 153, as the super-resolution pre-data RMi (not shown), the region Ei (number), the square-adjusted super-resolution mesh Mei number (Me1, Me2,...), the adjusted fine mesh mei number (me1, m2,...) that individually constitutes the square-adjusted super-resolution mesh Mei, and the smoothed elevation value of each adjusted fine mesh mei, etc. are memorized.

[0242] In addition, before performing the super-resolution image three-dimensional visualization process (S340), the plane rectangular coordinate conversion unit 145 activates the consideration distance grid number calculation unit 148. In order to perform the super-resolution image three-dimensional visualization process, the consideration distance L is necessary. The calculation of this consideration distance is not shown in the flowchart, but is performed by the consideration distance grid number calculation unit 148.

[0243] When the input consideration distance L is 50 m, the number of meshes corresponding to the case where the number of division points DKi is 9 × 9 is output to the super-resolution image generation unit 151 as the number of super-resolution fine meshes KLi corresponding to the consideration distance.

[0244] The inclination calculation process of the super-resolution image three-dimensional visualization process (S340) for the super-resolution image generation unit 151 will be described.

[0245] The inclination calculation process designates the super-resolution pre-data RMi (region Ei, square-adjusted super-resolution mesh Mei, adjusted fine mesh mei, smoothed elevation value, etc.) in the memory 153.

[0246] After that, the square-adjusted super-resolution mesh Mei included in this super-resolution DEM pre-data RMi is designated, and the adjusted fine mesh mei that is mutually associated with this is designated.

[0247] After that, the super-resolution DEM pre-data RMi having the adjusted fine mesh mei adjacent (for example, in four directions) to the designated adjusted fine mesh mei is designated.

[0248] Next, obtain the inclination between the "smoothed elevation value zhi of the adjusted fine mesh mei included in the specified super-resolution DEM pre-data RMi" and the "smoothed elevation value zhi of each adjusted fine mesh mei included in the super-resolution DEM pre-data RMi in each of the four adjacent directions", and associate the adjusted fine mesh mei for which the average inclination (hereinafter referred to as inclination αi (or tilt)) is specified with each other.

[0249] That is, associate with the specified super-resolution DEM pre-data RMi.

[0250] Perform this kind of processing individually for each of all the adjusted fine mesh mei.

[0251] If the inclination αi (α1, α2,...) is made to correspond to the distance axis, it becomes as shown in Figure 29 in (b) below.

[0252] In addition, when explaining (b) of Figure 29 it is described as (a) of Figure 18 as Figure 29 the following.

[0253] Figure 29 The solid line in (b) is called the average inclination drawing line SLi (solid line).

[0254] As shown in (a) of Figure 29 below, the smoothed elevation values zhi of the adjusted fine meshes me1, me2, me3, me4 between A1 and A2 increase from zh1 to zh5 in a substantially constant ratio.

[0255] Therefore, as shown in (b) of Figure 29 below, there is no change in the average inclinations α1, α2, α3, α4 of me1, me2, me3, me4 between A1 and A2.

[0256] However, as shown in (a) of Figure 29 below, between A1 and A2 for me5, me6, me7, me8, the height gradually increases gently from zh5 to zh9.

[0257] Therefore, as shown in (b) of Figure 29 below, the average inclinations α5, α6, α7, α8 of me5, me6, me7, me8 gradually decrease in average inclination.

[0258] In addition, as shown in Figure 29As shown in (a), me9, me10, me11, and me13 between A2 and A3 gradually and gently increase the height from zh9 to zh14.

[0259] Therefore, as in Figure 29 As shown in (b) of FIG. 1 , the average inclinations α9, ..., α13 of me9, me10, me11, and me13 also change while decreasing slightly.

[0260] The important thing here is that, as in Figure 29 As shown in (b) of FIG. 1 , if the inclination of Lai when no moving average (smoothing for super-resolution) is performed is plotted on Figure 29 In (b), A1 to A2 (me1 to me8) become like the dotted line Lsai (overlapping with the solid line between me1 to me5).

[0261] Furthermore, between A2 and A3 (me9 to me16), the line changes suddenly downward like the dotted line Lsbi (the line between me14 and me16 overlaps with the solid line). This change point is recorded as Dsi.

[0262] However, in the present embodiment, since moving average (super-resolution smoothing processing) is performed, the location of Dsi does not change suddenly, but becomes like the average slope plot line SLi (solid line). Therefore, jaggies are not generated.

[0263] Next, the super-resolution stereo visualization processing (S340) of the super-resolution image generating unit 151 sequentially specifies the super-resolution DEM pre-data RMi in the memory, and for each of the specified super-resolution DEM pre-data RMi, individually specifies the adjusted fine mesh mei contained therein as a focus point.

[0264] For each of these focus points, an adjusted fine mesh mei corresponding to the "number of super-resolution fine meshes KLi equivalent to the considered distance" is specified, and the adjusted fine mesh mei with the largest smoothed elevation value zhi between the specified adjusted fine meshes mei is searched.

[0265] Then, the adjusted fine mesh mei having the largest smoothed elevation value zhi and the adjusted fine mesh mei of the focus point are used to obtain the above-ground openness and underground openness to obtain the peak-to-valley degree (also called floating degree).

[0266] After that, the gradation color value (color in the red system) representing the combination of the peak valley degree and the inclination degree is assigned to the adjusted fine mesh mei of the attention point and imaged. In the present embodiment, this is referred to as a super-resolution red image (super-resolution three-dimensional visualization image Ki).

[0267] Here, use Figure 30 、 Figure 31 、 Figure 32 to explain resampling in the projection conversion process (S320) of the plane rectangular coordinate conversion unit 145.

[0268] Figure 30 The (a) of Figure 30 shows the super-resolution square mesh Mbi (9×9: number of lines) after moving average processing of the memory 142. The vertical axis is latitude and the horizontal axis is longitude.

[0269] In addition,[[]] Figure 30 The (a) of Figure 30 shows representative points (〇 marks) at the corners of mbi of the super-resolution square mesh Mbi (9×9) after moving average processing. In Figure 30 the (a) of Figure 30 , 〇 marks are shown as an example at the third horizontal line from the top of the super-resolution square mesh Mbi after moving average processing.

[0270] In addition,[[]] Figure 30 in the (a) of Figure 30 , the vertical axis represents the latitude direction and the horizontal axis represents the longitude direction.

[0271] Figure 30 The (b) of Figure 30 shows the plane rectangular super-resolution mesh Mdi (solid line) after converting the super-resolution square mesh Mbi (before plane rectangular conversion: 9×9) after moving average processing into plane rectangular coordinates.

[0272] For the plane rectangular super-resolution mesh Mdi (solid line) after converting into plane rectangular coordinates, the vertical axis is labeled as Y and the horizontal axis is labeled as X (the Z direction is not recorded).

[0273] In addition,[[]] Figure 30 the (b) of Figure 30 shows the case where the plane rectangular super-resolution mesh Mdi becomes a trapezoidal shape when converted into plane rectangular coordinates.

[0274] In addition,[[]] Figure 30 in (b), Figure 30 the super-resolution square mesh Mbi after moving average processing in the (a) of Figure 30 is overlapped and shown (dotted line).

[0275] Here,[[]] Figure 30The △ marks shown in (b) are resampling points of the representative points (〇 marks) of the angles of mbi (however, this is an example where x and y are the same). As an example, the third one from the top of the planar rectangular super-resolution mesh Mdi after conversion to planar rectangular coordinates is shown with △ marks.

[0276] As shown in Figure 30 (b), the 〇 marks and the △ marks deviate from each other.

[0277] Figure 31 This is Figure 30 an explanatory diagram when the Z direction (elevation) of (b) is set as the vertical axis and the X direction is set as the horizontal axis. That is, as shown in Figure 31 , the line (solid line) connecting the △ marks to each other is the locus of elevation.

[0278] In addition, in Figure 32 , a case is shown where the planar rectangular super-resolution mesh Mdi becomes rectangular when converted to planar rectangular coordinates. Figure 32 (a) is the super-resolution square mesh Mbi after moving average before planar rectangular conversion. Figure 32 (b) overlays (dotted line) the super-resolution square mesh Mbi after moving average of (a). Figure 32

[0279] Using such data, the above-mentioned super-resolution image stereoscopic visualization processing (S340) and distance consideration calculation are performed.

[0280] The above-mentioned super-resolution image stereoscopic visualization processing (S340) uses the technology of Japanese Patent Publication No. 3670274.

[0281] An explanation of its general content is given.

[0282] As shown in Figure 33 , based on the identification number Idn and the height difference of the two-component vector Vn of the nth (n = 1 to N) processing, its longitude xn, latitude yn, and altitude zn are calculated, and this value is associated with the corresponding coordinate point Qn = {Xn = xn, Yn = yn, Zn = zn} in the virtual three-dimensional (3D) X - Y - Z orthogonal three-dimensional coordinate space 80 stored in the memory (not shown) (planar rectangular coordinate conversion).

[0283] That is, by storing the identification number Idn of the vector Vn in the memory area corresponding to the coordinate point Qn in the memory, the vector Vn is mapped to the three-dimensional coordinate space 80, and by performing this operation for a total of N vectors, the vector field 70 is mapped to the three-dimensional coordinate space 80 (processing P1).

[0284] Furthermore, a surface S of a column {Qn: n ≤ N} of an appropriate number of coordinate points with Id added in the three-dimensional coordinate space 80, which are connected with necessary smoothness, is obtained by the method of least squares or the like, and this surface S is divided into a total of M {M ≤ N} minute surface regions {Sm: m ≤ M}, a point of interest Qm is determined respectively, and the relevant information is stored in the memory.

[0285] After that, for each surface region Sm, a local region Lm+ on the front side (Z+ side) of the surface S within a specific radius is confirmed from its point of interest Qm, and the opening degree (i.e., the grazing solid angle to the sky side or the second differential value equivalent to the solid angle) Ψm+ around the divided point of interest Qm is obtained (processing P2), and is memorized as the floating degree of the surface region Sm.

[0286] An image obtained by performing gray scale display of the floating degree Ψm+ on the entire surface S is used as the processing result A. This image A clearly shows the peak side of the terrain, that is, the convex part (of the surface S) as a convex part.

[0287] After that, for the above surface region Sm, a local region Lm- on the back side (Z- side) of the surface S within the above specific radius is confirmed from its point of interest Qm, and the opening degree (i.e., the grazing angle to the ground side or the second differential value equivalent to the solid angle) Ψm- around the divided point of interest Qm is obtained (processing P3), and is memorized as the sinking degree of the surface region Sm. An image obtained by performing gray scale display of the sinking degree Ψm- on the entire surface S is used as the processing result C.

[0288] This image C clearly shows the valley side of the terrain, that is, the concave part (of the surface S) as a concave part.

[0289] However, it should be noted that this image C does not become a simple inversion of the image A.

[0290] After that, for the above surface region Sm, the floating degree Ψm+ and the sinking degree Ψm- are weighted and synthesized (w+Ψm+ + w-Ψm-) with a purposefully determined distribution ratio w+: w- (w+ + w- = 0) (that is, depending on whether the peak or the valley is to be emphasized) to obtain the three-dimensional effect brought to the vicinity of the point of interest Qm by the front and back local regions Lm (Lm+, Lm-) of the surface S within a specific radius. Figure 17The processing of P4) is performed and memorized as the floating and sinking degree Ψm of the surface area sm.

[0291] An image in which the floating and sinking degree Ψm is gray-scale displayed over the entire surface S is used as the processing result B. This image B makes the peaks and valleys of the terrain more prominent and enhances the visual three-dimensional effect by clearly displaying the convex parts (like convex parts) and the concave parts (like concave parts) of the surface S. Additionally, for the above-mentioned synthesized weighting, w+ = -w- = 1 in image B.

[0292] After that, for the above-mentioned surface area Sm, the maximum inclination (or the first derivative value equivalent to this inclination) Gm of the surface area Sm is directly or indirectly obtained (processing P6) by the method of least squares and memorized as the inclination Gm of the surface area Sm.

[0293] The achromatic display image of the image in which the inclination Gm is gray-scale displayed in the color R of the red system over the entire surface S is used as the processing result D. This image D also has the effect of visually creating a three-dimensional sense of the terrain (that is, the surface S).

[0294] After that, by mapping the three-dimensional coordinate space 80 together with its related information (Ψm, Gm, R) onto the two-dimensional plane 90 (processing P5), on the region 90m on the two-dimensional plane 90 corresponding to the divided region Sm of the surface S where the column connecting the coordinate points Qm is located, the R-color tone display of the inclination Gm is performed, and for the brightness of this R-color tone, gray-scale display corresponding to the floating and sinking degree Ψm is performed.

[0295] This image (achromatic display image) is used as the processing result F. This image F gives a visual three-dimensional sense to the terrain (that is, the surface S).

[0296] Image E represents the result of mapping (processing P5) the information of the image D (that is, the R-color tone representing the inclination Gm) and the information of the floating and sinking degree corresponding to image A (that is, the floating degree Ψm+) onto the two-dimensional plane 90, and emphasizes the peak parts.

[0297] Image G represents the result of mapping (processing P5) the information of the image D (that is, the R-color tone representing the inclination Gm) and the information of the floating and sinking degree corresponding to image C (that is, the sinking degree Ψm-) onto the two-dimensional plane 90, and emphasizes the valley parts.

[0298] Find the property isopleth (in this embodiment, the contour line and the outline of the terrain) Ea that links the coordinate points Qn where "the property extracted from the components of the vector Vn of the vector field 70 (in this embodiment, the altitude zn) is equal in the column of the coordinate point Qn", memorize this, and output or display it as needed (process P7).

[0299] This result I is also helpful for grasping the three-dimensional shape of the terrain (that is, the surface S).

[0300] After that, on the two-dimensional plane 90, map or output and display the three-dimensional coordinate space 80 together with its related information (Ψm, Gm, R), and map or output and display the above-mentioned property isopleth Ea (process P8). Take the display image (the monochrome display image) as the processing result H. This image H also gives a visual three-dimensional sense to the terrain (that is, the surface S).

[0301] Therefore, after performing the first step (61) of mapping the vector field (70) to the three-dimensional three-dimensional coordinate space (80) to obtain the corresponding coordinate point sequence, perform: the second step of taking the openness around the attention point, which is delimited by "the area on the front side within the specific radius of the attention point", on the local area of the surface connecting the coordinate point sequence, as the buoyancy (floating and sinking degree) (A) of the local area; and

[0302] The third step of taking the openness around the attention point, which is delimited by "the area on the inner side within the specific radius of the attention point", on the local area of the surface connecting the coordinate point sequence, as the sinking degree (C) of the local area; and

[0303] The fourth step of weighted synthesizing the buoyancy (A) and the sinking degree (C) to take the openness brought by "the area on the front side and the area on the inner side within the specific radius" to the attention point around the local area of the surface connecting the coordinate point sequence, as the floating and sinking degree (B) of the local area; and

[0304] The fifth step of mapping the three-dimensional coordinate space (80) to the two-dimensional plane (90) and performing gray-scale display (F) corresponding to the floating and sinking degree of the local area on the area on the two-dimensional plane (90) corresponding to the local area of the surface connecting the coordinate point sequence.

[0305] Next, a more specific description will be given. Based on DEM (Digital Elevation Model) data, the inclination corresponding to the inclination Gm, the ground openness equivalent to the floating degree Ψm+ of the first embodiment, and the underground openness equivalent to the sinking degree Ψm- are obtained. These three parameters are stored as a grayscale image of the planar distribution.

[0306] The difference image between the ground openness and the underground openness is added to the gray channel, and the inclination is added to the red channel. By creating a virtual color image, peaks or mountaintop parts are represented in white, valleys or depressions are represented in black, and the steeper the inclination, the redder it is represented. Using this combination of representations, a three-dimensional image can be generated even with a single image.

[0307] That is to say, the three-dimensional representation method of the three-dimensional image of this embodiment meshes between contour lines, and represents the difference (that is, the inclination) between adjacent meshes in red tones. Regarding whether it is higher or lower than the surrounding area, it is represented in grayscale. Corresponding to the floating and sinking degree Ψm, in this embodiment, it is called the peak-valley degree, which means that the brighter part is higher (peak) than the surrounding area, and the darker part is lower (valley) than the surrounding area. The light and dark are synthesized by multiplication to generate a three-dimensional effect.

[0308] That is to say, in this embodiment, the concept of openness is used. Openness quantifies the degree to which the location protrudes above the ground and sinks below the ground compared to the surroundings. That is, the ground openness, as Figure 34 shown, represents the width of the air visible within the considered distance L from the specimen location of interest. In addition, the underground openness represents the width of the underground within the considered distance L when looking at the area in an inverted perspective.

[0309] Openness depends on the considered distance L and the surrounding terrain. Generally, the ground openness becomes larger at locations that protrude higher from the surroundings, taking larger values at mountaintops or peaks and smaller values at depressions or valley bottoms. On the contrary, the underground openness becomes larger at locations that sink deeper underground, taking larger values at depressions or valley bottoms and smaller values at mountaintops or peaks.

[0310] That is to say, on the super-resolution fine mesh mbi included within the range from the point of interest to a certain distance (considered distance L), topographic profiles are generated in 8 directions, and the maximum value of the inclination of the line connecting each location to the point of interest (when viewed from the vertical direction) is obtained. This kind of processing is carried out for 8 directions.

[0311] In addition, within a certain distance range starting from the point of interest of the smoothed micro elevation value of the super-resolution micro mesh that has been inverted, topographic profiles are generated in each of the eight directions, and the maximum value of the slope of the line connecting each location to the point of interest is obtained (the minimum value when observing L2 (not shown) from the vertical direction in the stereogram of the ground surface).

[0312] This kind of processing is carried out for the eight directions. That is, the ground openness and the underground openness, as shown in Figure 32 , consider two basic locations A (iA, jA, HA) and B (iB, jB, HB). Since the specimen interval is about 60 cm, the distance between A and B becomes

[0313]

[0314] Figure 32 Taking the elevation of 0 m as the reference, the relationship between A and B at the specimen locations is shown.

[0315] The elevation angle θ of specimen location A relative to specimen location B is given by θ = tan-1{(HB - HA) / P}. The sign of θ becomes positive when (1) HA < HB and negative when (2) HA > HB.

[0316] The set of specimen locations within the range of azimuth D and considering distance L starting from the specimen location of interest is denoted as DSL.

[0317] This is called the "D-L set of the specimen location of interest". Here, let

[0318] DβL: The maximum value among the elevation angles of the elements of DSL relative to the specimen location of interest

[0319] DδL: The minimum value among the elevation angles of the elements of DSL relative to the specimen location of interest

[0320] (see Figure 35 (a) of Figure 35 (b) of

[0321] ), and the following general definitions are made.

[0322]

[0323] And

[0324] .

[0325] DφL represents the maximum value of the zenith angle in the sky from the local point of interest to the direction D that can be seen within the distance L. The so-called horizon angle generally corresponds to the angle on the ground when L is set to infinity. In addition, DψL represents the maximum value of the nadir angle in the ground from the local point of interest to the direction D that can be seen within the distance L. If L is increased, the number of local points belonging to DSL will increase. Therefore, DδL has a non-increasing characteristic, while DβL has a non-decreasing characteristic.

[0326] Therefore, both DφL and DψL have a non-increasing characteristic with respect to L.

[0327] In surveying, the concept of high angle is defined based on the horizontal plane passing through the local point of interest, which is not strictly consistent with θ. In addition, if we want to discuss the above-ground angle and the below-ground angle strictly, we must consider the curvature of the earth, and Definition 1 is not an absolutely correct description. After all, Definition 1 is a concept defined only on the premise of using DEM for terrain analysis.

[0328] The above-ground angle and the below-ground angle are concepts related to the designated direction D, but the following definitions are introduced as an expansion of these concepts.

[0329] Definition II: The above-ground openness and underground openness at a distance L from the local point of interest are respectively

[0330]

[0331] as well as

[0332]

[0333] That is to say, as in Figure 36 As shown in the figure, a composite image Dh is generated by multiplying and synthesizing the above-ground openness image data Dp (the peak is emphasized in white: also called the above-ground openness image Dp) and the underground openness image data Dq (the bottom is emphasized in black: also called the underground openness image Dq), and an inclined emphasized image Dr is generated in which the red color is emphasized more as the inclination of the inclined image data Dra (also called the inclined image Dra) increases, and this inclined emphasized image Dr is synthesized with the composite image Dh.

[0334] That is to say, by Figure 36 By the processing shown in , the above-mentioned super-resolution stereoscopic visualization image Ki (also referred to as super-resolution red stereoscopic image) is obtained and displayed on the display unit.

[0335] Therefore, by combining such expressions, a three-dimensional image can be produced even with a single image, so that the degree of height of the concavity and convexity and the degree of inclination can be grasped at a glance.

[0336] Next, the processing of the super-resolution image generation unit 151 will be described in detail.

[0337] Figure 37 It is a block diagram of the program of the super-resolution image generation unit 151.

[0338] As shown in Figure 37 , the super-resolution image generation unit 151 includes: a ground openness data creation unit 9, an underground openness data creation unit 10, and an inclination calculation unit 8 that read the smoothed elevation value zhi included in the super-resolution DEM data in the memory 153 (layer). Furthermore, it further includes a convex portion emphasis image creation unit 11, a concave portion emphasis image creation unit 12, an inclination emphasis unit 13, a first synthesis unit 14, and a second synthesis unit 15.

[0339] Figure 38 It is a schematic configuration diagram for explaining the convex portion emphasis image creation unit 11 and the concave portion emphasis image creation unit 12. However, in Figure 38 , the first synthesis unit 14, etc. are illustrated.

[0340] Figure 39 It is a schematic configuration diagram for explaining the inclination emphasis unit 13. However, in Figure 39 , the first synthesis unit 14, the second synthesis unit 15, etc. are described.

[0341] The convex portion emphasis image creation unit 11, as shown in Figure 38 , includes a first gray level 11A, a tone correction unit 22, etc., and the concave portion emphasis image creation unit 12 includes a second gray level 11B, a color inverse conversion processing unit 27, etc.

[0342] The ground openness data creation unit 9 generates topographic profiles in each of 8 directions on the adjusted fine mesh mei within the range from the attention point to a certain distance (considering the distance L), and obtains the maximum value of the inclination of the line connecting each location to the attention point (when viewed from the vertical direction) (refer to Figure 41 ). This kind of processing is performed for 8 directions.

[0343] In addition, the underground openness data creation unit 10 generates topographic profiles in each of 8 directions within the range from the attention point to a certain distance of the smoothed elevation value zhi of the inverted adjusted fine mesh mei, and obtains the maximum value of the inclination of the line connecting each location to the attention point (the minimum value when viewed from the vertical direction in the three-dimensional view of the ground surface for L2 (not shown)) (refer to Figure 41 ). This kind of processing is performed for 8 directions. As shown in Figure 41Individually obtained for each da (e.g., 0.5555 m) as shown.

[0344] The inclination calculation unit 8 obtains the average inclination (inclination) of the surface of the square adjacent to the point of interest (adjusted fine mesh mei) as described above. The average inclination (inclination) is the inclination of the surface approximated from four points using the least squares method.

[0345] The convex part emphasizing image creation unit 11, as shown in Figure 38 is provided with a color assignment process 20 for emphasizing convex parts.

[0346] This color assignment process 20 for emphasizing convex parts, as shown in Figure 38 is provided with a first gray level 11A for expressing peaks and valley bottoms in terms of brightness. Each time the ground opening degree data creation unit 9 obtains the ground opening degree (average angle when looking at the range of L in 8 directions from the point of interest: an index for determining whether it is located at a high place), the brightness (lightness) corresponding to the value of this ground opening degree ψi is calculated.

[0347] For example, when the value of the ground opening degree falls within the range of about 40 degrees to 120 degrees, 50 degrees to 110 degrees is made to correspond to the first gray level 11A and assigned 255 gradations (refer to Figure 40 (a) of).

[0348] That is, since the value of the ground opening degree is larger in the part closer to the peak part (convex part), the color becomes white.

[0349] In addition, the color assignment process 20 for emphasizing convex parts of the convex part emphasizing image creation unit 11 reads the ground opening degree and assigns color data based on the first gray level 11A (refer to Figure 40 (b) of), and stores this in the ground opening degree file 21 (ground opening degree image data Dpa).

[0350] On the other hand, the gradation correction unit 22 of the convex part emphasizing image creation unit 11 stores the ground opening degree layer Dp, which is "an image in which the color gradation of this ground resolution data Dpa is inverted", in the memory 23. That is, the ground opening degree layer Dp (ground opening degree image Dp) adjusted so that the peak becomes white is obtained. The expression of "layer" is recorded as a layer because it is an image synthesized with other images.

[0351] The concave part emphasizing image creation unit 12, as shown in Figure 38As shown, it has a color assignment process 25 for emphasizing concave parts. This color assignment process 25 for emphasizing concave parts has a second gray level 11B (refer to Figure 40 for expressing the convex bottom and peak in terms of brightness. When the underground openness data production unit 10 calculates the underground openness ψi (average in 8 directions from the point of interest) each time, the brightness corresponding to the value of this underground openness ψi is calculated.

[0352] For example, when the value of the underground openness falls within the range of about 40 degrees to 120 degrees, 50 degrees to 110 degrees is corresponding to the second gray level 11B (refer to Figure 40 ), and it is assigned 255 gradations.

[0353] That is to say, since the value of the underground openness is larger in the part closer to the bottom (concave part), the color becomes darker.

[0354] After that, as shown in Figure 38 the concave part emphasizing image production unit 12 reads the underground openness, assigns color data based on the second gray level 11B to it, and stores this in the underground openness file 26. Then, the color inverse conversion processing unit 27 corrects the color gradation of the underground openness image data Dqa and stores it in the memory 28.

[0355] In the case where the color becomes too dark, it is set to the color at the level after correcting the tone curve. This is called the underground openness layer Dq (also called the underground openness image) and is saved.

[0356] The inclination emphasizing unit 13, as shown in Figure 39 has a color assignment process 30 for emphasizing inclination.

[0357] This color assignment process 30 for emphasizing inclination has a third gray level 11C (refer to Figure 40 for expressing the degree of inclination in terms of brightness. When the inclination calculation unit 8 calculates the inclination (average in 4 directions from the point of interest) each time, the brightness (lightness) of the third gray level 11C corresponding to the value of this inclination is calculated.

[0358] For example, when the value of the inclination αi falls within the range of about 0 degrees to 70 degrees, 0 degrees to 50 degrees is corresponding to the third gray level 11C, and it is assigned 255 gradations. That is to say, 0 degrees is white and above 50 degrees is black. The color becomes darker at the location where the inclination αi is larger.

[0359] After that, as shown in Figure 39As shown in [figure], the color assignment process 30 for emphasizing the inclination of the inclination emphasizing portion 13 reads in the inclination (tilt) and assigns color data based on the third gray scale 11C.

[0360] Next, the red processing 32 emphasizes R in the RGB color mode function (however, in the case where 50% emphasis can also be performed). That is, in the memory 33 (layer), an inclination emphasized image Dr (also simply called the inclination image Dr) is obtained in which the greater the inclination, the more red is emphasized.

[0361] The first synthesis unit 14 obtains a synthesized image Dh by multiplying and synthesizing the ground openness image Dp and the underground openness image Dq. At this time, the balance between the two is adjusted so that the valley portion is not damaged.

[0362] The so-called "multiplication" mentioned above is a term for the layer mode in Photoshop (registered trademark), and it becomes an OR operation in numerical processing.

[0363] For example, it is assumed to be a stable red constructed with a hue of 0°, a chroma of 50%, and a brightness of 80%.

[0364] When the RGB values are specified in the range of 0 to 255 for each color, RED is set to about "204", GREEN is set to about "102", and BLUE is set to about "102". The HEX value (hexadecimal WEB color, HTML color code) is set to #CC6666. Or, the CMYK values used in color printing are roughly set to indigo "C20%", magenta "M70%", yellow "Y50%", and black "K0%".

[0365] The second synthesis unit 15 synthesizes (multiplication synthesis) with the inclination emphasized image Dr in which the greater the inclination, the more red is emphasized, and the super-resolution three-dimensional visualization image Ki is displayed through the display processing unit.

[0366] That is, in the memory 153 of the super-resolution image generation unit 151, as shown in Figure 42 As the super-resolution DEM data, the area Ei (number), the square-adjusted super-resolution mesh Mei (number), the adjusted fine mesh mei (number), the division width da, zri, the elevation value zhi after smoothing processing, the inclination αi, the color value of the inclination, the color value of the floating and sinking degree (not shown: ground openness, underground openness), etc. are stored. The set of this super-resolution DEM data is also simply called the super-resolution DEM. This super-resolution DEM is color-added and displayed through the display processing unit.

[0367] Use Figure 43 AndFigure 44 , the effects obtained by performing the above general processes will be described.

[0368] Figure 43 It is an explanatory diagram of a red stereoscopic image using a 5m DEM generated based on Japanese Patent No. 66692984. Figure 44 It is an explanatory diagram of a super-resolution image generated by the high-speed super-resolution image stereoscopic visualization processing system according to this embodiment.

[0369] Figure 44 Since the elevation value zhi after the smoothing process by the 9×9 bilinear interpolation process and the 9×9 moving average process is used, compared with Figure 43 , it becomes a clean image without a jagged feeling.

[0370] Such an image is preferably used in combination with a general map as shown, for example, in Figure 45 . As shown in Figure 45 , since the unevenness of the entire map is clear, there is a three-dimensional feeling, and the height and subsidence of the ground (including roads) in the urban area can be known stereoscopically.

[0371] In addition, in the above-described Embodiment 1, the projective transformation is performed between the super-resolution red stereoscopic map creation process and the super-resolution tilt calculation process, but it can also be performed after the super-resolution red stereoscopic map creation process.

[0372] <Embodiment 2>

[0373] Figure 46 It is a schematic configuration diagram of Embodiment 2.

[0374] In Figure 46 , the super-resolution rasterization processing unit 135, the moving average unit 134, and the distance grid number calculation unit 148 of Figure 3 are not shown.

[0375] In Figure 46 , the memory 153 (not shown) of the super-resolution image generation unit 151, the super-resolution image generation unit 151, and the X-direction adjustment unit 152 are shown and described.

[0376] In addition, in Figure 46Among them, a smoothed contour calculation unit 156, a memory 158 for smoothed contour data, a memory 159 for Geospatial Information Authority of Japan standard map, a first image synthesis unit 160 (Geospatial Information Authority of Japan map + red), a memory 161 for the first synthesized image (Geospatial Information Authority of Japan map + red), a second image synthesis unit 162 (smoothed contour + red), a memory 164 for the second synthesized image (smoothed contour + red), a third image synthesis unit 166 (contour + Geospatial Information Authority of Japan map + red), a memory 168 for the third synthesized image (contour + Geospatial Information Authority of Japan map + red), and a display processing unit 150 are shown.

[0377] In the memory 159 for Geospatial Information Authority of Japan standard map, vector data of a 1:25,000 standard map Gki (level 16) is stored.

[0378] The smoothed contour calculation unit 156 designates the adjusted fine mesh mei in the memory 153, and searches for the adjusted fine mesh mei having the same elevation value as the elevation value after smoothing of the super-resolution fine mesh representative point dpij assigned to this adjusted fine mesh mei.

[0379] After that, for these adjusted fine meshes mei, the adjusted fine meshes mei to be connected are determined by standard deviation calculation processing or the like, and are closed.

[0380] At this time, as shown in Figure 47 Among, for example, the super-resolution fine mesh representative points (dp4,5), (dp4,6), (dp5,5), (dp5,6) at the four corners of the adjusted fine mesh mei, for example, the line connecting (dp4,5) and (dp4,6) is used as the entrance line, and the line connecting (dp5,5) and (dp5,6) is used as the exit line.

[0381] After that, the elevation values between (dp4,5) and (dp4,6) are interpolated, and the elevation values between (dp5,5) and (dp5,6) are interpolated, and a line (y = ax + b) connecting points having substantially the same elevation is generated and connected.

[0382] After that, the set of straight lines of this closed adjusted fine mesh mei is vectorized (function), and this is subjected to moving average processing (the same processing as step S60 in Figure 1 and step S280 in Figure 5 ), and is stored as smoothed contour information Ji in the memory 158 for smoothed contour data. When the smoothed contour information Ji is imaged, it is called a smoothed contour Ci.

[0383] When the adjacent adjusted fine meshes mei to be connected are in the X direction or the Y direction, the center coordinates (x, y) are linked by a straight line. In addition, when the adjacent adjusted fine meshes mei to be connected are in an inclined direction, the center of the coordinates of the two corner points on the side of the adjusted fine mesh mei in the connection direction and the center coordinates between the two points of the adjusted fine mesh mei in the connection inclined direction are linked and set as a straight line.

[0384] After that, the set of these straight lines is set as a function (it can also be set as an approximate function).

[0385] That is to say, the smoothed contour information Ji becomes a contour line connected by straight lines of the adjusted fine mesh mei without performing curvature maximization processing such as spline curve and Bézier Curve as in the prior art.

[0386] At this time, color values are assigned. That is to say, the smoothed contour information Ji is composed of the region Ei, the adjusted fine mesh mei, the size (0.5555 m), the elevation value zhi, and the connection direction (up (or down) in the X direction, up (or down) in the Y direction, or right diagonal or left diagonal), etc.

[0387] In addition, the interval between the smoothed contour lines Ci can also be 1 m, 2 m, 3 m, ….

[0388] In Figure 48 an example is shown in which the contour line (vector) of the above-mentioned smoothed contour information Ji is overlapped with a red image that is not smoothed. In addition, in Figure 49 an enlarged view of Figure 48 is shown. In addition, in Figure 50 an image of the result after smoothing the contour line using the elevation value zhi is shown. However, Figure 50 is an image after performing a moving average twice.

[0389] As shown in Figure 48 and Figure 49 as a whole, the contour lines are serrated (for example, at the place of Va). However, in Figure 50 it becomes smooth as a whole (refer to Va).

[0390] Figure 52 is an image after synthesizing such a contour line with a super-resolution red image obtained by the high-speed super-resolution image stereoscopic visualization processing system of the first embodiment. Figure 52It is an image obtained from the super-resolution DEM based on 5m DEM. In addition, the interval between contour lines is several meters (for example, 1m, 2m, or 3m).

[0391] In addition, Figure 51 It is a map obtained by synthesizing the contour lines of the 1:25,000 map and the red image generated based on 10m DEM. In addition, the interval between contour lines is 10m.

[0392] As shown in Figure 52 The smooth contour lines are displayed in detail, and the color degree of the inclination of the concavity and convexity (if the concave part is deep, it is dark red, and if the convex part is high, the color will turn white) can be known in detail and clearly.

[0393] That is, since the interval between contour lines is several meters (for example, 1m, 2m, or 3m), the contour lines of this embodiment can be used as a 1:10,000 contour map.

[0394] In addition, the first image synthesis unit 160 (Geospatial Information Authority map + red) generates the "Geospatial Information Authority map + red composite image" GFi obtained by multiplying and synthesizing the "image in memory 153 (not shown)" and the "image data of the vector data of the standard map Gki (level 16) in the memory 159 for the Geospatial Information Authority standard map", and stores this in the first composite image memory 161 (for Geospatial Information Authority map + red) (see Figure 52 ).

[0395] At this time, the first image synthesis unit 160 (Geospatial Information Authority map + red) reduces the color value of the image in memory 153 by about 50% in such a way that it is different from the color (for example, orange) when the vector of the standard map (urban map such as buildings and roads) is imaged. For example, it is set to a stable red constructed with a hue of 0°, a chroma of 50%, and a brightness of 80%.

[0396] When the RGB values are specified in the range of 0 to 255 for each color, RED is set to about "204", GREEN is set to about "102", and BLUE is set to about "102". The HEX value (hexadecimal WEB color, HTML color code) is set to #CC6666. Or, the CMYK values used in color printing are roughly set to indigo "C20%", magenta "M70%", yellow "Y50%", and black "K0%".

[0397] The second image synthesis unit 162 (smoothed contour lines + red) generates a "smoothed contour lines + red" image GaCi obtained by multiplying and synthesizing the "memory 161 for the first synthesized image (Geospatial Information Authority of Japan map + red)" and the "data obtained by imaging the smoothed contour line information CJi in the memory 158 for smoothed contour line data", and stores it in the second synthesized image memory 164 (smoothed contour lines + red).

[0398] The third image synthesis unit 166 (contour lines + Geospatial Information Authority of Japan map + red) generates a "standard map + red + smoothed contour lines" image Gami obtained by multiplying and synthesizing the "Geospatial Information Authority of Japan map + red synthesized image" Gfi in the "memory 161 for the first synthesized image (Geospatial Information Authority of Japan map + red)" and the "smoothed contour lines + red" image GaCi in the "second synthesized image memory 164 (smoothed contour lines + red)", and stores it in the third synthesized image memory 168 (see Figure 49 )

[0399] In addition, in the memory 159 for the Geospatial Information Authority of Japan standard map, vector data of a 1:25,000 scale standard map (level 16) is stored.

[0400] Even if the vector data of buildings, roads, etc. on the Geospatial Information Authority of Japan base map is read into the display memory and displayed, there will be no jagged feeling. That is, the resolution is adjusted for the complex linear road contours and building contours of the 1:25,000 scale standard map (level 16).

[0401] In addition, even if it is enlarged, there will be no jagged feeling (jagged edges). Therefore, it is possible to confirm in detail the conditions of cliffs, flat surfaces, road inclinations, etc.

[0402] Therefore, it is possible to generate a map that is slightly the same as the 1:10,000 scale map that the Geospatial Information Authority of Japan has given up making.

[0403] In addition, in the above embodiment, although it is described as an example of using the DEM of the ground terrain to perform super-resolution at high speed, it is also possible to use the DEM of the seabed terrain to perform super-resolution at high speed.

[0404] <Other Embodiments>

[0405] (Lab Colorization)

[0406] By applying Lab colorization processing to the above high-speed super-resolution image stereoscopic visualization processing system, the image becomes more vivid. In this embodiment, this system is called a high-speed super-resolution image stereoscopic visualization processing system with Lab color assignment.

[0407] For example, it is desired to avoid problems such as "the valley becomes too dark", "the water system is difficult to track", "since the track of the valley is dark, the system is difficult to track", etc.

[0408] Figure 53 It is a schematic configuration diagram of a Lab color - imparting high - speed super - resolution image three - dimensional visualization processing system of other embodiments.

[0409] In Figure 53 For components with the same reference numerals as those described above, their descriptions are omitted.

[0410] In this embodiment, in addition to the above - mentioned Figure 3 each part, it further includes a Lab color part 320 and a Lab synthesis part 340.

[0411] In addition, it is described under the assumption that "in the memory (not shown) of the super - resolution image generation part 151, the data of the above - described memory 153 (including the square - adjusted super - resolution mesh Mei) has been generated".

[0412] Each time the Lab color part 320 designates the adjusted fine mesh mei of the square - adjusted super - resolution mesh Mei as the point of interest, it generates "converting the ground openness obtained by the super - resolution image generation part 151 into the a * of Lab color, converting the underground openness into b * , and converting the inclination (also called tilt) into L * " of the super - resolution L * a * b * color image Li, and stores it in a memory (not shown).

[0413] The Lab synthesis part 340 synthesizes the super - resolution L * a * b * color image Li and the super - resolution three - dimensional visualization image Ki (super - resolution red three - dimensional visualization image) (referred to as the Lab color red super - resolution image Lki), and stores it in the memory 172.

[0414] The display processing part 150 has a display memory (not shown), reads the data corresponding to the input image type into the display memory, and displays the image (for example, the L * a * b * color red super - resolution image Lki) with the color value assigned to this data on the screen of the display part.

[0415] That is to say, the Lab color part 320, after performingFigure 54 (same as Figure 4 ), Figure 55 (same as Figure 5 ), after the processing shown in Figure 56 , perform the Lab color red super-resolution image processing shown in Figure 54 . Since Figure 4 is the same processing as Figure 55 which is the same processing as Figure 5 , the description thereof is omitted.

[0416] In step S320 of Figure 56 , the super-resolution image generation unit 151 reads each of the planar rectangular super-resolution meshes mdi of each of the planar rectangular super-resolution meshes Mdi (S330), and performs super-resolution image stereoscopic visualization processing (S400).

[0417] At this time, through the X-direction adjustment unit 152, in the memory 153 (not shown), a square-adjusted super-resolution mesh Mei is generated by adjusting the planar rectangular super-resolution mesh Mdi (for example, a rectangle, a trapezoid) to a square. (Super-resolution image stereoscopic visualization processing S400 of the super-resolution image generation unit 151)

[0418] The super-resolution image generation unit 151 obtains the inclination αi (α1, α2,...) of each of all the adjusted fine meshes mei through the above-described inclination calculation processing (refer to Figure 39 (b)).

[0419] In addition, the super-resolution image generation unit 151 obtains the ground opening degree and the underground opening degree of each of the adjusted fine meshes mei through the above-described processing, and obtains the peak-valley degree (also referred to as the floating and sinking degree) (refer to Figure 38 ).

[0420] In addition, in the super-resolution red stereoscopic visualization processing (S340) shown in Figure 56 , a gradation color value (a color in the red system) representing the combination of the peak-valley degree and the inclination (also referred to as the inclination) is assigned to the adjusted fine mesh mei. That is, image formation is performed. In the present embodiment, the super-resolution ground opening degree image is simply referred to as the ground opening degree image Dp in the same manner as above, in the present embodiment, the super-resolution underground opening degree image is simply referred to as the underground opening degree image Dq, and the inclination image is simply referred to as the inclination emphasis image Dr.

[0421] On the other hand, the Lab color unit 420 performs L * a * b *Color adjustment image generation process (S420).

[0422] This L * a * b * In the color adjustment image generation process, the image data of the adjusted fine mesh mei (fine mesh: super-resolution mesh) of the ground openness image Dp is read out, and when each such readout is performed, the data assigned to the a * channel is obtained as a * data.

[0423] In addition, the image data of the adjusted fine mesh mei (fine mesh) of the underground openness image Dq is read out, and when each such readout is performed, the data assigned to the b * channel is obtained as b data.

[0424] In addition, the image data of the tilt-emphasized image Dr is read out, and when each such readout is performed, the data assigned to the L * channel is obtained as L data.

[0425] After that, each time the a * data and the b * data and the L * data are obtained, these data are successively defined in the L * a * b * space to obtain the super-resolution L * a * b * color image data Li (refer to Figure 62 ).

[0426] After that, the Lab synthesis unit 340 synthesizes it with the super-resolution red stereoscopic visualization image Ki obtained in step S340, and stores this as the L * a * b * color red super-resolution image KLi in the memory 172 (S440).

[0427] The display processing unit 150 displays this L * a * b * color red super-resolution image KLi, etc. on the screen (S460).

[0428] In Figure 57 the process of obtaining the L * a * b * color red super-resolution image KLi is shown as an image.

[0429] In Figure 57In (a), the super-resolution L * a * b * color image data Li is shown. In Figure 57 In (b), the super-resolution stereoscopic visualization image Ki (super-resolution red stereoscopic visualization image) is shown. In Figure 57 In (c), the L after synthesizing these images * a * b color red super-resolution image KLi is shown. This L * a * b * The color red super-resolution image KLi is an image in which the transparency of L * a * b * is reduced by about 30%.

[0430] Using Figure 58 , the Lab color adjustment image processing (S420) described above is supplemented. Figure 58 is a configuration diagram of the Lab colorization unit 320. However, for the super-resolution image generation unit 151, L * a * b * is described in the synthesis unit 340 (also simply referred to as the synthesis unit 340) and the like.

[0431] As shown in Figure 58 , the Lab colorization unit 320 includes a tilt image tone correction unit 62, a ground openness image tone correction unit 64, an underground openness image tone correction unit 63, L * channelization unit 66, b * channelization unit 65, a * channelization unit 67, and L * a * b * color image formation unit 68.

[0432] Furthermore, it includes a tone correction unit 69, an XYZ color system conversion unit 71, an RGB color system conversion unit 70, a fine-tuning correction unit 72, a tilt spectrum calculation unit 52, an underground openness spectrum calculation unit 51, and a ground openness spectrum calculation unit 53, etc. For the image, valleys or depressions with high underground openness are adjusted to indigo, and peaks or tops with large ground openness are adjusted to red. Valley slopes with small ground openness also appear green.

[0433] The tilt spectrum calculation unit 52 calculates the spectrum distribution (also referred to as the tilt degree spectrum) of the super-resolution tilt enhancement image Dr in the memory 153 (not shown) of the super-resolution image generation unit 151, and stores this in the memory 55.

[0434] The inclination spectrum of the inclined emphasized image Dr, when displayed as a histogram with the inclination (0° to 90°) on the horizontal axis and the frequency (n) of pixels on the vertical axis, becomes as shown in Figure 59 in (a) of Figure 59 As shown in (a) of

[0435] The ground openness spectrum calculation unit 53 calculates the spectrum distribution (also referred to as the ground openness spectrum) of the ground openness image Dp in the memory 153 of the super-resolution image generation unit 151, and stores this in the memory 54.

[0436] The ground openness spectrum, when displayed as a ground openness histogram with the openness (0° to 180°) on the horizontal axis and the frequency (n) of pixels on the vertical axis, becomes as shown in Figure 59 in (b) of Figure 59 As shown in (b) of

[0437] The underground openness spectrum calculation unit 51 calculates the spectrum distribution (also referred to as the underground openness spectrum) of the underground openness image Dq in the memory 153, and stores this in the memory 56.

[0438] The underground openness spectrum, when displayed as an underground openness histogram with the underground openness (0° to 180°) on the horizontal axis and the frequency (n) of pixels on the vertical axis, becomes as shown in Figure 59 in (c) of Figure 59 As shown in (c) of

[0439] (Explanation of the image tone part)

[0440] The inclined image tone correction unit 62 performs tone correction in such a way that the steeper the slope, the darker it becomes. That is, the input side (horizontal axis) is set to the inclination from 0° to 50°, and the output side is set to 0 (black) to 255 (white), and a linear conversion is performed such that when the inclination αi is 50°, it is converted to "0" and when the inclination αi is 0°, it is converted to the maximum value 255 (refer to Figure 60 in (a) of

[0441] The histogram of the obtained inclination is Figure 59of (a).

[0442] The ground opening degree image tone correction unit 63 corrects the tone in such a way that the contour of the peak becomes bright. That is, with the input side (horizontal axis) being the ground opening degree of 50° to 130°, and the output side being 0 (black) to 255 (white), a linear conversion is performed such that when the ground opening degree θi is 50°, it is converted to "0" and when the ground opening degree θi is 130°, it is converted to the maximum value 255 (refer to Figure 60 of (b)).

[0443] However, when the ground opening degree θi is 90°, it is set to make it "120°". Specifically, it is performed through a look-up table. That is, as shown in Figure 60 of (b), the center system of the conversion line is set to pass through (90°, 120). The histogram of the obtained ground opening degree is shown in Figure 59 of (b).

[0444] The underground opening degree image tone correction unit 64 corrects the tone in such a way that the contour of the valley becomes dark. That is, with the input side (horizontal axis) being the underground opening degree of 50° to 130°, and the output side being 0 (black) to 255 (white), a linear conversion is performed such that when the underground opening degree αi is 50°, it is converted to "255" and when the underground opening degree αi is 130°, it is converted to "0" (refer to Figure 60 of (c)). However, when the underground opening degree αi is 90°, the output side is set to become "120". Specifically, it is performed through a look-up table. The histogram of the obtained underground opening degree is shown in Figure 59 of (c).

[0445] That is, if the relationship between the ground opening degree and the underground opening degree is shown as a scatter diagram through the tone conversion unit, it becomes as shown in Figure 39 below. Figure 61 The ground opening degree (50° to 130°) is depicted on the horizontal axis, and the underground opening degree (50° to 130°) is depicted on the vertical axis. This scatter diagram is centered at (90°, 90°). The closer the scatter diagram is to a straight line, the more blue there is, if it is separated, the more yellow there is, and if it is further away, the more red there is.

[0446] In addition, the color of the plotted points shows the color corresponding to the inclination amount of the same point of interest. As shown in Figure 61As shown, it can be known that there is an inverse relationship between the ground opening degree and the underground opening degree. This relationship becomes stronger as the distance becomes shorter. At the peak, the ground opening degree is large and the underground opening degree is small. At the valley, the ground opening degree is small and the underground opening degree is large.

[0447] According to the color of the plotted points, it can be known that there is a weak positive proportional relationship between the sum value of the ground opening degree and the underground opening degree and the inclination.

[0448] (Channeling section)

[0449] L * Each time the inclination (0° → 50°) is converted to a color value (255 → 0) by the inclination image tone correction section 62, the channeling section 66 assigns it to L * At the channel (refer to Figure 60 ) of (a).

[0450] a * Each time the ground opening degree θi (50° → 130°) is converted to a color value (0 → 255) by the ground opening degree image tone correction section 63, the channeling section 67 assigns it to a * At the channel.

[0451] b * Each time the underground opening degree φi (50° → 130°) is converted to a color value (255 → 0), the channeling section 65 assigns it to b * At the channel.

[0452] L * a * b * The color image creation section 68 defines the data of L * the L of the channeling section 66 * and a * the a of the channeling section 67 * and b * the b of the channeling section 65 * in the L * a * b * space, and obtains L * a * b * color image Li (Lai, Lbi) (refer to Figure 62 ).

[0453] (Others)

[0454] Since the tone correction section 69 is for L * a * b *The color space of the color image Li is wider than that of the RGB space. Therefore, after roughly adjusting its color through level correction, a tone curve is used to make fine adjustments.

[0455] For example, change the inclination from 0° to 50° to 0° to 30° or 0° to 70°, and redistribute the color values. In addition, change the ground openness (50° to 130°) and the underground openness (50° to 130°) to 60° to 120° or 70° to 110°, and redistribute the color values.

[0456] The XYZ color system conversion unit 71 converts the Lab-adjusted image into the XYZ color system (stored in the color space defined by the XYZ color system) (the Lab image in the XYZ color system).

[0457] The RGB color system conversion unit 71 converts the Lab image in the XYZ color system into the RGB color system (stored in the RGB space) (the Lab image in the RGB layer). The Lab image in this RGB layer is stored in the memory 42.

[0458] The Lab composite unit 340 (image composite processing) composites (multiplies and composites) the Lab color image in the RGB layer in the memory 42 with the super-resolution stereoscopic visualization image Ki (super-resolution red stereoscopic visualization image), and stores it as the Lab color red super-resolution image Lki in the memory 172.

[0459] The fine-tuning correction unit 72 adjusts the contrast (transparency), etc. of the Lab color red super-resolution image Lki (caused by the operator's input).

[0460] That is, by overlapping and compositing with these images, the expression of the valleys that have become too dark is adjusted and improved to be approximately indigo. Therefore, there will be no situation where the valleys become too dark and are difficult to observe.

[0461] <Embodiment 3>

[0462] This Embodiment 3 is a method for emphasizing the water system.

[0463] Figure 65 It is a schematic configuration diagram of Embodiment 3. For the same component symbols as above, their descriptions are omitted. As shown in Figure 65 , it is equipped with a water system adjustment unit 180. The water system adjustment unit 180 makes adjustments in such a way that the bright side of the histogram of the underground openness is skipped and only the dark side becomes an image. Through this, the parts with a high underground openness (valley parts or parts that are relatively lower than the surroundings) are extracted.

[0464] After that, in the same way as Embodiment 2, the super-resolution L* a * b * The color image Li overlaps with the super-resolution three-dimensional visualization image Ki (super-resolution red three-dimensional visualization image).

[0465] In addition, different from contour maps and the like, the red three-dimensional map does not have the concept of height and only represents unevenness. Therefore, when the elevation difference within the object range is large, the overall sense of undulation may become insufficient. When representing a large terrain in the red three-dimensional map, it can be achieved by increasing the consideration range of the openness according to the scale of the represented terrain (for example, if you want to see the terrain undulation within a range of about 1 km, set the openness range to 1000 m).

[0466] However, in practice, in the calculation of the openness, it is restricted by the micro-topography existing around the attention place, and it is not very likely to be calculated all the way to 1 km.

[0467] For example, for a 1m DEM, if the openness range is set like 1 km, at the valley or peak part, the openness value will saturate, the valley will become too dark, and the peak will become too bright.

[0468] To solve this problem, the resolution of the DEM to be calculated is reduced (the resolution of the terrain is reduced) and the calculation is performed.

[0469] By this, it becomes possible to perform a calculation considering the earth system (refer to Figure 66 ).

[0470] At 1m DEM and 4m DEM, in the case of 4m DEM, the overall sense of undulation is strong.

[0471] In addition, the method of the above embodiment can be applied to the terrain of Venus or Mars. Furthermore, it can also be applied to the visualization of unevenness measured by an electron microscope. In addition, if it is applied to a game machine, a three-dimensional effect can be obtained even without wearing glasses.

[0472] In the above embodiment, although the sinking degree (peak-valley degree) obtained from the ground openness and the underground openness is used to generate the super-resolution image, for example, it can also be overlapped and displayed on an image obtained by the sky visibility rate, the terrain protection coefficient, the plane curvature, the high-pass filter, and the Mexican hat function.

[0473] Or, it can also be an image in which the sky visibility rate, the terrain protection coefficient, the plane curvature, the high-pass filter, the Mexican hat function, etc. are inverted, and this is set as the underground openness image.

[0474] In addition, the DEM of the base map can also be ALB (Airborne lidar Bathymetry) (point cloud density: 1 point / m 2 ).

[0475] -Symbol Explanation-

[0476] 110: Database for base map

[0477] 112: Region definition unit

[0478] 115: Odd-number division unit for 5m DEM

[0479] 132: Grid coloring processing unit

[0480] 134: Moving average unit

[0481] 135: Super-resolution rasterization processing unit

[0482] 137: TIN bilinear interpolation unit

[0483] 145: Plane rectangular coordinate conversion unit

[0484] 148: Calculation unit for considering distance grid number

[0485] 151: Super-resolution image generation unit

[0486] 152: X-direction adjustment unit.

Claims

1. A high-speed super-resolution image three-dimensional visualization processing system, comprising: (A) For each of a plurality of square mesh groups in a specific area of a numerical elevation model, obtaining a unit of a super-resolution square mesh in which the square mesh is defined by a fine square super-resolution mesh group; (B) For each of the super-resolution square meshes, performing interpolation processing, and assigning the elevation value after interpolation to each of the super-resolution fine meshes of the super-resolution square mesh; (C) For each of the super-resolved square meshes, a unit that performs a moving average process a specific number of times on each super-resolved fine mesh and updates the interpolated elevation value to the smoothed elevation value; and (D) Generating a unit of a planar rectangular super-resolution mesh in which the super-resolution square mesh after the unit of (C) is defined by a planar rectangular coordinate; and (E) Generating a unit of a square super-resolution three-dimensional visual image based on the planar rectangular super-resolution fine meshes of the planar rectangular super-resolution mesh.

2. The high-speed super-resolution image three-dimensional visualization processing system according to claim 1, wherein the unit of (E) comprises: (E1) Generating a unit of a square-adjusted super-resolution mesh that adjusts the planar rectangular super-resolution mesh to a square; (E2) Designating the square-adjusted fine meshes of the square-adjusted super-resolution mesh as points of interest in sequence, and obtaining the inclination with the adjacent square-adjusted fine meshes based on the elevation value after smoothing processing, and assigning it to the square-adjusted fine meshes of the points of interest; (E3) Designating the square-adjusted fine meshes as points of interest in sequence, and for each of the points of interest, obtaining the peak-valley degree based on the ground openness and the underground openness between the point of interest and the adjacent square-adjusted fine meshes, and assigning the gradation color value representing the combination of the peak-valley degree and the inclination to the square-adjusted fine meshes of the points of interest; and (E4) After the unit of (E3), defining the square-adjusted fine meshes and their gradation color values in a display memory and displaying them as the super-resolution three-dimensional visual image.

3. The high-speed super-resolution image three-dimensional visualization processing system according to claim 1, wherein it comprises: A unit for obtaining a ground openness image (Dp) in which the larger the value of the ground openness, the brighter the color is assigned, an underground openness image (Dq) in which the larger the value of the underground openness, the darker the color is assigned, and an inclination emphasis image (Dr) in which the larger the value of the inclination, the more the color with red emphasized is assigned; A unit for obtaining a first composite image (Ki: super-resolution red image) obtained by overlapping the ground openness image (Dp), the underground openness image (Dq), and the inclination emphasis image (Dr); Read out the image data of the above-ground openness image (Dp) and obtain, each time such reading is performed, a unit to which a data assigned to channel a is allocated. * The unit of a data for channel a; Read out the image data of the underground openness image (Dq) and obtain, each time such reading is performed, a unit of b data assigned to the b * channel; Read out the image data of the tilted emphasis image (Dr) and obtain a unit of L data assigned to the L * channel each time such reading is performed; By defining these data successively in L each time the a data, b data, and L data are obtained, * a * b * a unit that obtains the Lab image data (Li) of the ground openness image (Dp), the underground openness image (Dq), and the tilt emphasis image (Dr) in space; and A unit for generating a second composite image (lab color super-resolution red image KLi) obtained by synthesizing the Lab image (Li) and the first composite image (Ki: super-resolution red image).

4. The high-speed super-resolution image stereoscopic visualization processing system according to claim 1, wherein the unit of (E) comprises: (E1) a unit that generates a square-adjusted super-resolution mesh that adjusts the planar rectangular super-resolution mesh to a square; (E2) a unit that sequentially designates the square-adjusted fine meshes of the square-adjusted super-resolution mesh as points of interest, and obtains the inclination thereof with the adjacent square-adjusted fine meshes based on the elevation value after the smoothing process, and assigns it to the square-adjusted fine meshes of the points of interest; (E3) a unit that sequentially designates the square-adjusted fine meshes as points of interest, and for each of these points of interest, respectively obtains the peak-valley degree based on the ground openness and the underground openness between it and the adjacent square-adjusted fine meshes, and assigns the gradation color value representing the combination of the peak-valley degree and the inclination to the square-adjusted fine meshes of the points of interest; and (E4) a unit that, after the unit of (E3), defines the square-adjusted fine meshes and their gradation color values in the display memory and displays them as the super-resolution stereoscopic visual image.

5. The high-speed super-resolution image stereoscopic visualization processing system according to claim 1, wherein the unit of (A) comprises: (A1) a unit that reads the square mesh group in a specific area of the numerical elevation model stored in the numerical elevation model memory into the first memory; and (A2) a unit that, for each square mesh in the first memory, equally divides the side in the latitude direction and the side in the longitude direction of this square mesh with an odd number of division points (the odd number does not include 1), and generates the super-resolution square mesh having the super-resolution fine mesh group.

6. The high-speed super-resolution image stereoscopic visualization processing system according to claim 1, wherein the unit of (C) performs: a unit that reads the elevation value after the smoothing process of the area into the display memory and displays it as the super-resolution stereoscopic visual image on the screen, and performs the super-resolution smoothing process again with the input of the super-resolution smoothing instruction.

7. The high-speed super-resolution image stereoscopic visualization processing system according to claim 1, wherein the unit of (C) has: (C1) a unit that sequentially designates the super-resolution square meshes, and for each of the designated super-resolution square meshes, sequentially designates the super-resolution fine meshes; (C2) a unit that applies a moving average mesh divided by the number of division points a specific number of times to the super-resolution fine meshes and generates the elevation value after the smoothing process; and (C3) a unit that updates the elevation value after the interpolation of the designated super-resolution fine meshes to the elevation value after the super-resolution smoothing process of the unit of (C2).

8. The high-speed super-resolution image stereoscopic visualization processing system according to claim 1, wherein Comprising: (F) Among the planar rectangular super-resolution fine meshes or the square-adjusted fine meshes of the said step, designating a planar rectangular super-resolution fine mesh as the starting point, extracting the straight lines passing through the closed mesh groups having the same elevation value after smoothing processing as the designated mesh, and vectorizing these straight lines to generate contour vectors as units; and (G) A unit for setting the said contour vectors as images and writing them into the display memory for display on the screen.

9. The high-speed super-resolution image three-dimensional visualization processing system according to claim 1, wherein Comprising: A standard map memory storing 1:25,000 standard map information in which roads, buildings, rivers, and marshes are defined by vector information, And, comprising: (H) A unit for visualizing the said standard map information and displaying this image together with the image of the said super-resolution image or the contour vectors or both.

10. The high-speed super-resolution image three-dimensional visualization processing system according to claim 1, wherein The tone display of the said inclination is set to a color of the red system.

11. The high-speed super-resolution image three-dimensional visualization processing system according to claim 1, wherein There is provided a map memory unit for storing vector data of roads, buildings, rivers, marshes, trees, or some combination or all of these as a standard map; A unit for reducing the tone of the said inclination by 30% to 60%; And A unit for visualizing the said vector data and further overlapping and displaying it on the image on which the overlapping display has been performed.

12. A high-speed super-resolution image three-dimensional visualization processing program for causing a computer to execute the functions of the following units: (A) For each of the square mesh groups in a specific area of the numerical elevation model, respectively generating, in a memory unit, super-resolution square meshes in which the square meshes are defined by a fine square super-resolution fine mesh group; (B) For each of the said super-resolution square meshes, performing interpolation processing and allocating the elevation values after interpolation to each of the super-resolution fine meshes of the super-resolution square mesh; (C) For each of the said super-resolution square meshes, applying a specific number of moving average processes to each of the super-resolution fine meshes and updating the elevation values after interpolation to smoothed elevation values; (D) Generating, in a memory unit, planar rectangular super-resolution meshes in which the super-resolution square meshes after the unit (C) are defined by planar rectangular coordinates; and (E) Generating, in a memory unit, a super-resolution three-dimensional visual image based on the planar rectangular super-resolution fine meshes of the planar rectangular super-resolution meshes.

13. The high-speed super-resolution image three-dimensional visualization processing program according to claim 12, wherein The unit of (E) comprises: a unit that generates a square-adjusted super-resolution mesh by adjusting the planar rectangular super-resolution mesh to a square; a unit that sequentially designates the square-adjusted fine meshes of the square-adjusted super-resolution mesh as points of interest, and obtains the inclination between each of them and the adjacent square-adjusted fine meshes based on the elevation values after the smoothing process, and assigns it to the square-adjusted fine meshes of the points of interest; a unit that sequentially designates the square-adjusted fine meshes as points of interest, and for each of these points of interest, separately obtains the peak-valley degree between it and the adjacent square-adjusted fine meshes, and assigns the gradation color value representing the combination of this peak-valley degree and the inclination to the square-adjusted fine meshes of the points of interest; and a unit that, after the unit of (E3), defines the square-adjusted fine meshes and their gradation color values in the display memory and displays them as the super-resolution stereoscopic vision image.

14. The high-speed super-resolution image stereoscopic visualization processing program according to claim 12, wherein the unit of (A) includes: a unit that reads a square mesh group in a specific area of the numerical elevation model stored in the memory unit for the numerical elevation model into the first memory; and a unit that, for each of the square meshes in the first memory, equally divides the sides in the latitude direction and the sides in the longitude direction of each square mesh with an odd number of division points (the odd number does not include 1) to generate the super-resolution square mesh having the super-resolution fine mesh group.

15. The high-speed super-resolution image stereoscopic visualization processing program according to claim 12, wherein the unit of (C) performs: a unit that reads the elevation values after the smoothing process of the area into the display memory and displays them as the super-resolution image on the screen, and performs the super-resolution smoothing process again with the input of the super-resolution smoothing process instruction.

16. The high-speed super-resolution image stereoscopic visualization processing program according to claim 12, wherein the unit of (C) performs: a unit that sequentially designates the super-resolution square meshes, and for each of the designated super-resolution square meshes, sequentially designates the super-resolution fine meshes; a unit that applies a moving average mesh divided by the number of division points a specific number of times to the super-resolution fine meshes and generates the elevation values after the smoothing process; and a unit that updates the elevation value after the interpolation of the designated super-resolution fine meshes to the elevation values after the super-resolution smoothing process of the unit of (C2).

17. The high-speed super-resolution image three-dimensional visualization processing program according to claim 12, wherein, causes a computer to execute the functions of the following units: (F) Among the planar rectangular super-resolution fine meshes or the square-adjusted fine meshes of the said step, a planar rectangular super-resolution fine mesh designated as the starting point is specified, and a straight line passing through a closed mesh group having the same elevation value after smoothing processing as the specified mesh is extracted, and these straight lines are vectorized, and a unit generated as a contour vector; and (G) A unit that sets the said contour vector as an image and writes it into the display memory and displays it on the screen.

18. The high-speed super-resolution image three-dimensional visualization processing program according to claim 12, wherein, Causes a computer to execute the functions of the following units: A unit that stores in a standard map memory the standard map information at a scale of 1:25,000 that defines roads, buildings, rivers, and marshes through vector information; and (H) A unit that visualizes the said standard map information and displays this image together with the image of the said super-resolution image or the contour vector or both.

19. The high-speed super-resolution image three-dimensional visualization processing program according to claim 12, wherein The hue display of the said inclination is set to a color in the red system.

20. The high-speed super-resolution image three-dimensional visualization processing program according to claim 12, wherein, Causes a computer to execute the functions of the following units: A unit that stores in a map memory unit vector data of roads, buildings, rivers, marshes, or trees, or some combination or all of these as a standard map; A unit that reduces the hue of the said inclination by 30% to 60%; and A unit that visualizes the said vector data and further overlays and displays it on the image on which the overlay display has been performed.

Images