JP6196910B2 - Simulation method and simulator - Google Patents

Description

  The present invention relates to an installation such as a traffic light that is planned to be installed on a railroad track or an obstacle that is expected to be accidentally present on the railroad track, etc. The present invention relates to a simulation method and a simulator used for confirming the visibility of an installation or an obstacle when looking forward in the traveling direction (before installation of the installation or before the occurrence of an obstacle).

  For example, when planning the installation of installations such as traffic lights on railroad tracks (hereinafter simply referred to as tracks), is the installation location a place that is easily visible (high visibility) from the driver on the car? It is necessary to confirm in advance (before construction). In addition, when there are obstacles such as falling rocks on the track, or when there are fallers, fallen objects, or other obstacles on the station premises, the driver should be able to avoid the collision at any point in order to avoid a collision. It is necessary to learn in advance (before actual driving) whether to start braking. In the event of falling rocks, etc. on the motorway, it is desirable to install a warning sign etc. in front of the vehicle approaching the location.

  For example, special signal light emitters (hereinafter referred to as “special occurrences”), which are installed on railroad tracks, are installed in places requiring warning due to railroad crossing obstacles, falling rocks, avalanches, strong winds, etc. When there is, the red light is turned on to stop the vehicle (or train, the same applies hereinafter) traveling on the track.

  Therefore, when setting up a special event, it must be confirmed before construction, since it will be installed outside the place that needs to be warned, and in a place where it can be seen when the vehicle approaches a predetermined distance. Installations that require line-of-sight confirmation at the time of new installation include, for example, relay traffic lights that relay the signal display of traffic lights whose visibility is obstructed by mountains, etc., and other facilities and equipment.

  Conventionally, when establishing a special event, a simulated special event is set up on the site, and the signal personnel and driving personnel jointly check the outlook based on the presence of the site, verify the validity of the installation kilometer, and then install the special event We are constructing. Therefore, a great deal of labor and time is required for the confirmation and construction of the line of sight.

  Although it is an urgent task to prevent accidents at railroad crossings, it is difficult to meet the demand with the above-mentioned conventional method, which requires a great deal of time and labor for confirming and constructing the line of sight. It is.

  In addition, if obstacles such as falling rocks, falling pedestrians, falling objects, etc. occur on the track, how far the emergency braking starts from that location is effective for preventing accidents. It is not preferable to check at the site because there is a danger (there is no particular prior art document).

  The present invention has been made in view of the above circumstances, and when a new installation of an installation such as a traffic light is planned on the track, or when an obstacle is expected to be generated on the track, Developed image processing technology that can display CG or illustrations, images, etc. of installed objects or obstacles in the correct position with respect to the actual video ahead in the direction of travel, and using this will improve construction speed or increase accident prevention rate It is an object of the present invention to provide a simulation method and a simulator capable of achieving the above.

  The simulation method according to the present invention that achieves the above object is as follows. (A) A front-captured image obtained by capturing a forward direction (front of the track) of a vehicle traveling on a track with a camera provided on the vehicle at predetermined short time intervals. Using the data and the previously created facility / equipment-kilometre correspondence data, the forward-captured image and the kilometre are matched to create kilometer data representing the kilometer corresponding to each frame image, (b ) On the other hand, a rail shape extraction process is performed to extract rail shape from front shot image data and generate rail shape data, and (c) facility / equipment-kilometre correspondence data, kilometer data, and rail shape data. Based on the desired installation position data, a desired installation position derivation process is performed to derive a desired installation position derivation process for deriving a desired installation position. (D) Based on the desired installation position data, In addition to reading forward-captured image data of the position to be installed, it reads out pre-stored installation object image data, and displays the image of the installation object on the front-captured image that is a real image of the position where the installation object is desired. An overlay display process is performed.

  The image of the installation object includes an actual image, an image created by CG, an illustration, and the like.

The simulator according to the present invention that achieves the above object is (a1) a camera that is provided on a vehicle and photographs the front of a track at a predetermined short time interval when used for prior confirmation of the visibility of an installation; (b1) A storage unit for storing data, (c1) an image processing device for processing forward-captured image data input from the camera, (d1) an input unit for inputting a control command to the image processing device, (e1) by the image processing device A monitor for displaying the processed image,
The storage unit (b11) stores first captured image data input from the camera and image data after image processing by the image processing device, and (b12) facility / equipment-kilometre correspondence data. A second storage unit for storing, (b13) a third storage unit for storing rail shape data generated by rail shape extraction processing by the image processing device, and (b14) a fourth storage unit for storing image data of the installation object Including
The image processing apparatus includes (c11) a storage / reading control unit that controls storage of data in the storage unit and reading of data from the storage unit, and (c12) each frame from the starting point on the line image stored in the first storage unit. (C13) a display ratio adjusting unit for adjusting the display ratio of the image of the installation; (c13) calculating a kilometer distance from the image shooting point (which is also a viewing point) and extracting the rail shape from the front shot image data; c14) an image composition unit that overlay-displays an image in which the display ratio of the installation object is adjusted on the image of the visual recognition point of the track displayed on the monitor,
The input unit includes (d11) a point designating unit for designating a point (about kilometer) on the track image to be read with respect to the second storage unit, and (d12) an installation stored in the fourth storage unit. It has an installation specification part for specifying,
The image processing device reads the image of the viewing point designated by the operation of the point designation unit from the first storage unit and displays it on the monitor, and the rail shape data obtained by the rail shape extraction process from the image of the point and the facility / Based on the equipment-kilometre data and the kilometer data, the location where the installation is desired is derived, and the image data of the installation designated by the operation of the installation designation unit is read out and displayed on the monitor. The image of the installation object is displayed on the image (actual video) of the viewing point being displayed at a ratio adjusted according to the kilometer of the installation object desired point by the display ratio adjustment unit.

The simulator according to the present invention that achieves the above object is (a1) a camera that is provided on the vehicle and images the front of the track at predetermined short time intervals when used for prior confirmation of obstacle visibility, (b1) A storage unit for storing data, (c1) an image processing device for processing forward-captured image data input from the camera, (d1) an input unit for inputting a control command to the image processing device, (e1) by the image processing device A monitor for displaying the processed image,
The storage unit (b11) stores first captured image data input from the camera and image data after image processing by the image processing device, and (b12) facility / equipment-kilometre correspondence data. A second storage unit for storing; (b13) a third storage unit for storing rail shape data generated by rail shape extraction processing by the image processing device; and (b14) a fourth storage unit for storing image data of obstacles. Including
The image processing apparatus includes (c11) a storage / reading control unit that controls storage of data in the storage unit and reading of data from the storage unit, and (c12) each frame from the starting point on the line image stored in the first storage unit. (C13) a display ratio adjusting unit for adjusting the display ratio of the obstacle image; (c13) calculating a kilometer distance to the shooting point of the image (which is also a viewing point) and extracting the rail shape from the forward shot image data; c14) an image composition unit that overlay-displays an image in which the display ratio of the obstacle is adjusted on the image of the visual recognition point of the track displayed on the monitor,
The input unit includes (d11) a point designating unit for designating a point (about kilometer) on the track image to be read with respect to the second storage unit, and (d12) an obstacle stored in the fourth storage unit. With an obstacle designating part to designate,
The image processing device reads the image of the viewing point designated by the operation of the point designation unit from the first storage unit and displays it on the monitor, and the rail shape data obtained by the rail shape extraction process from the image of the point and the facility / Based on the equipment-km data and the kilometer data, the position where the obstacle is desired is derived, and the image data of the obstacle specified by the operation of the obstacle designating unit is read and displayed on the monitor. The image of the obstacle is displayed on the image (actual video) of the viewing point being displayed at a ratio adjusted by the display ratio adjustment unit according to the kilometer of the obstacle installation desired point.

ADVANTAGE OF THE INVENTION According to this invention, when using for the visibility prior confirmation of an installation object, the image of an installation object can be displayed on the correct position designated with respect to the real image ahead of the track image | photographed with the camera on the vehicle. Therefore, since the visibility (visibility) of the prospect at the installation scheduled point of the installation object can be confirmed without going to the site, the construction speed at the site can be remarkably improved as compared with the prior art.
Moreover, when using for the prior visual confirmation of an obstruction, the image of an obstruction can be displayed on the designated generation | occurrence | production prediction point with respect to the real image ahead of the track image | photographed with the camera on the vehicle. Therefore, since the visibility at the predicted occurrence point of the obstacle can be confirmed without going to the site, the accident prevention rate can be improved.

It is a block diagram which shows an example of a structure of the installation object visibility prior confirmation apparatus which concerns on embodiment of this invention. It is a conceptual diagram which shows the outline | summary of the function of the simulator which concerns on this invention, and related input data, intermediate data, and output data. It is a figure explaining normalization of a front photography picture. It is a conceptual diagram of a normalization process. It is a figure which shows the coordinate system and symbol used for normalization. It is a figure which further shows normalization of a screen. It is a figure which shows the relationship between the trapezoid of the frame image, and the rectangle by which the trapezoid was normalized. It is a figure which shows the relationship between the frame image before and after normalized and pixel movement amount. It is a figure which shows the outline | summary of relative speed measurement. It is a flowchart which shows the flow of a process of relative speed measurement. It is a flowchart which shows the flow of a process of distance calculation. It is a graph which shows an example of the result of relative distance measurement. It is a figure which shows the virtual example of the kilometer data of the facilities and equipment corresponding to a relative distance. It is a figure which shows the outline | summary of frame-kilometer graph preparation. It is a figure which shows an example of the produced frame-kilometer graph. It is a figure which shows the flow of rail shape extraction. It is a figure which shows the result of the rail extraction by the existing system. It is a figure which shows the relationship between the point cloud data on a screen rail, and the point cloud data on a real space rail. It is a figure which shows the real space coordinate system of a rail. It is a figure which shows the real space coordinate system of the rail at the time of the distance d separation. It is a figure explaining conversion of installation object position data. It is a figure which shows the alignment correction point group data on a real space rail. It is a flowchart explaining the flow of operation | movement of the visibility confirmation which concerns on embodiment of this invention. It is a figure which shows an example of the track | line image of the visual recognition point P1 of 800 m outside from an installation desired point. It is a figure which shows the example by which the image of an installation object is displayed by overlay on the front picked-up image.

  In the following, the basic idea of the simulation method of the present invention will be described with reference to the drawings, with respect to an embodiment in the case of application to the installation object visibility prior confirmation method.

  In general, the object visibility prior confirmation method according to the present invention is based on the image correlation of the image data in front of the track (front-captured image data) captured by the camera on the vehicle, and the actual video frame and kilometer. After matching (matching) and extracting the rail from the forward shot image data, the three-dimensional position of the installation object in the screen is estimated, and the height, size, and distance of the installation object (distance from the rail center) ) Is determined and displayed on the monitor.

  A simulator VCA used in the above-described method for checking the visibility of an installed object according to the present invention is provided on a vehicle as shown in FIG. A camera 1 to be created, an image processing device 2 that processes forward-captured image data input from the camera 1, a storage unit 3 that stores various data, and an input unit 4 that inputs control commands to the image processing device 2 And a monitor 5 for displaying an image read out by the image processing apparatus 2 or a processed image.

  The image processing apparatus 2 stores the forward photographed image data input from the camera 1 in a predetermined area of the storage unit 3 and is designated by the forward photographed image data at the point designated by the input unit 4 and the input unit 4. An image data storage / reading control unit 21 that reads out the image data of the installed object and outputs it to the monitor 5; calculates the distance from the starting point on the stored track image to the shooting point on each image; The operation unit 22 that extracts the rail shape from the data, the display ratio adjustment unit 23 that adjusts the display ratio of the image of the installation object based on the obtained rail shape data, and the installation object whose display ratio is adjusted An image compositing unit 24 that overlay-displays the image on the front-captured image (actual image) of the designated point.

  The storage unit 3 stores a program storage unit 31 that stores a program for realizing the object, and forward-captured image data that is input from the camera 1 and image data that is subjected to image processing by the image processing device 2 A storage unit (first storage unit) 32; a location information storage unit (second storage unit) 33 for storing facility / equipment-kilometre correspondence data obtained by image processing of the image data by the image processing apparatus 2; A rail shape data storage unit (third storage unit) 34 for storing rail shape data obtained by rail shape processing by the image processing apparatus 2 and an installation object image data storage unit (fourth storage unit) for storing image data of installation objects 35).

  The input unit 4 includes an image capture switch 41 for instructing capture of forward-captured image data from the camera 1 and a visibility check for copying the captured forward-captured image to the monitor 5 and starting a visibility confirmation operation. A switch 42, a point designating unit 43 for designating a point on the forward-captured image to be read out from the position information storage unit 33, and an installation for designating an installation stored in the installation image data storage 35 And a designation unit 44.

  Below, simulator VCA is explained in full detail, and the installation thing visibility prior confirmation method is demonstrated collectively.

  In the present embodiment, a camera 1 that captures the front in the traveling direction of a vehicle traveling on a track is mounted. The camera 1 is a digital video camera that can capture images at regular time intervals, for example, at intervals of 1/30 second. In the case of a digital video camera, still images taken at regular time intervals can be output as digital image data, so that the output signal can be input to the image processing apparatus 2 as it is (without requiring AD conversion). Because it can. When the camera 1 is not a digital video camera, an output signal from the camera is AD converted by an AD converter (not shown) and then input to the image processing apparatus 2.

  The simulator VCA of FIG. 1 shows an example of a case where it is configured by a computer. Normally, the camera 1 is permanently installed in the vehicle or attached to the vehicle when necessary, but the components other than the camera 1, that is, the image processing device 2, the storage unit 3, the input unit 4, and the monitor 5 are always coupled. Then, the camera 1 that has been placed at a maintenance office or the like and has finished photographing is removed from the vehicle and connected to the image processing apparatus 2 with a cable or the like. However, the camera 1 attached to the vehicle may send forward image data (sometimes referred to simply as image data) of the camera to the image processing device 2 via a known telecommunication line.

[Summary of simulator functions]
FIG. 2 is a conceptual diagram showing an outline of functions of the simulator VCA according to the present invention and related input data, intermediate data, and output data.

Since the simulator VCA processes necessary input data and outputs target data as a result, the simulator VCA has the following four data processing functions. Each will be described in detail later.
(1) Matching the front shot image and kilometer
(2) Rail shape extraction
(3) Derivation of installed object position in the image
(4) Overlay display of front shot image and installation object image

  First, input data required when using the visibility checking method according to the present invention, intermediate data obtained in the middle of data processing, and output data obtained as a result of data processing will be described as preliminary knowledge.

[Input data]
In order to use the visibility checking method according to the present invention, the following data must be stored in the simulator VCA.
(A) Front shot image data (b) Facility / equipment-kilometre correspondence data (c) Installation and obstacle image data

  The forward photographed image data (D1 in FIG. 2) is video data obtained by photographing the front of the track at regular frame times by a camera provided in the head car of the train. The forward shot image data is stored and used in the image data storage unit 32 of the simulator VCA by electrically connecting a camera that has shot the front of the track to the simulator VCA or by an electric communication line.

  Facility / Equipment-Kilometre correspondence data (D2 in FIG. 2) corresponds to each facility or equipment on the track where the installation of the installation is planned and the kilometer from the starting point of the track to each facility or equipment. It is the data which was made to make. This facility / equipment-kilometre correspondence data is also stored in the position information storage unit 33 of the simulator VCA and used.

  The installed object and obstacle image data (D3 in FIG. 2) is image data of an installed object, for example, a facility or equipment installed on or near a track such as a special signal light emitter. The installed object and obstacle image data are also stored in the installed object image data storage unit 35 of the simulator VCA and used. The image of the installation object and the obstacle may be any of an actual image, an image created by CG, an illustration, and the like.

[Intermediate processing data]
In the simulator VCA, as a first data processing function, a matching process between a forward-captured image and about a kilometer, which will be described later, is performed. As a result of the processing, each image frame (photographing point) of the forward-captured image data (D1 in FIG. 2) and the kilometer data that graphs the kilometer from the starting point of the route in the image frame (photographing point) 2 of D4) is obtained.
In the preferred embodiment of the simulator VCA, a rail shape extraction process described later is performed as the second data processing function. As a result of the processing, rail shape data (D5 in FIG. 2) is obtained.
Further, in the simulator VCA, the installation position in the image is derived as a third data processing function. As a result of the processing, installed object and obstacle position data (D6 in FIG. 2) is obtained.

[output data]
Then, in the simulator VCA, the installed object and obstacle image data (D3 in FIG. 2) is read as the fourth data processing function, and the installed object and obstacle position data (FIG. 2) obtained by the third data processing function. Based on D6), a predetermined installation object image is displayed in an overlay at a predetermined position on the displayed forward photographed image. The monitor 5 on which the front shot image and the installation object image are displayed is based on the facility / equipment-kilometre correspondence data (D2 in FIG. 2). Distance, height and train position are displayed together.

[Details of simulator functions]
(1) Matching of the forward-captured image and the kilometre To perform this process, as a preparation process, the relative speed at each shooting point of the forward-captured image is measured. Then, the image frame-relative distance graph is scaled to create an image frame-relative distance graph.
Detailed data processing is divided into the following five.
1) Normalization of the forward image
2) Image frame-relative speed graph creation
3) Image frame-relative distance graph creation
4) Matching of facility / equipment position and relative distance
5) Image frame-kilometer graph creation The following is a description in order.

1) Normalization of the forward image As shown in FIG. 3A, the apparent distances Δd and Δd on the forward captured image obtained by the camera 1 are different from the actual distances. This makes it impossible to measure accurately. Therefore, in order to measure the relative speed, it is necessary to normalize (correct) so that the actual distance of any apparent distance on the screen is the same everywhere. The normalization of the front shot image is, as conceptually shown in FIG. 4, a video obtained by shooting a perspective-transformed image of the video of the camera 1 installed obliquely downward on the vehicle from the camera 1 ′ directly above. It is to convert into the corresponding image as shown in FIG.

  The normalization process for the forward captured image will be described. FIG. 5 shows a coordinate system and symbols used for normalization processing. H is the height of the camera 1, e is the direction of the camera 1, θ is the depression angle of the camera 1, f is the focal length of the camera 1, x is the horizontal direction of the screen, and y is the vertical direction of the screen.

となる。仮想画面の中心、つまりx-y 座標の原点をX-Y-Z 座標系で表すと、

である。仮想画面上のある点(Xn,Yn)はX-Y-Z 座標系で

と表すことができる。カメラ焦点の原点から( Xn ,Yn ) に向かうベクトルは、

なので、原点からこのベクトルの延長線上に( Xn ,Yn ) の実際の位置( Xn,Yn,Zn ) があるので未知数αを用いて、

と表現できる。Ｙｎ＝−Ｈなので

である。これより未知数αは

として与えられる。これを用いて

が得られる。 First, the relationship between the xy coordinates on the screen and the XYZ coordinate system is organized. When the direction e of the camera 1 is represented by a unit vector,

It becomes. When the center of the virtual screen, that is, the origin of the xy coordinates is expressed in the XYZ coordinate system,

It is. A point (Xn, Yn) on the virtual screen is in the XYZ coordinate system.

It can be expressed as. The vector from the camera focus origin to (Xn, Yn) is

So, since the actual position (Xn, Yn, Zn) of (Xn, Yn) is on the extension line of this vector from the origin, using the unknown α,

Can be expressed. Since Yn = -H

It is. From this, the unknown α

As given. With this

Is obtained.

とおく。これよりレール方向の点Ｚｂ, Ｚｔ（Ｒｂ,Ｒｔ）間の距離（Ｚｔ-Ｚｂ）は、

となる。また、左右レールの距離は１０６７ｍｍなので

である。これより、

が得られる。これを用いると、焦点距離ｆが

として求められる。これらを代入することでＺｔ−Ｚｂが求められる。正規化された４点のｘ方向の長さとｙ方向の長さの比は、

である。よって、画面上の４点を、

へ移すことで、図６（ａ）の画像は、図６（ｂ）に示すように、映像の正規化がされる。
前方撮影画像の正規化処理は、要約すると、図７に示すように、カメラ１で撮影した台形の画像ＦＡ１を、台形の底辺Ｗを共通にする矩形の画像ＦＡ２に変換することである。 Here, as shown in FIG. 6A, two sets of points on the left and right rails that are at the same distance in the Z direction on the forward photographed image are set, and are set as Lt, Rt; Lb, Rb. The position of the corresponding point in the XYZ coordinate system and the position on the screen

far. From this, the distance (Zt−Zb) between the points Zb, Zt (Rb, Rt) in the rail direction is

It becomes. Also, the distance between the left and right rails is 1067 mm

It is. Than this,

Is obtained. Using this, the focal length f is

As required. By substituting these, Zt-Zb is obtained. The ratio of the normalized length of the four points in the x direction to the length in the y direction is

It is. Therefore, 4 points on the screen

6A, the image of FIG. 6A is normalized as shown in FIG. 6B.
In summary, as shown in FIG. 7, the normalization processing of the forward captured image is to convert the trapezoidal image FA1 captured by the camera 1 into a rectangular image FA2 having a common base W of the trapezoid.

2) Image frame-relative velocity graph creation In this process, relative to the image normalized as described above, relative to the amount of pixel movement on the screen at two timings t and t + 1, as shown in FIG. The typical speed. In that case, in consideration of measurement errors and the like, processing is performed to remove outliers and maintain continuity of speed change.

  The outline of the relative speed measurement will be described with reference to FIG. 9. One frame screen is divided into a plurality of equal areas, and each area is set as a movement amount measurement range (FIG. 9A). Next, the movement destination of each area is searched from the entire screen, and the movement amount is obtained (FIG. 9B). If there is an outlier in the obtained movement amount, the area is excluded (FIG. 9 (c)). Then, the average movement amount of the area not excluded is obtained (FIG. 9 (d)). Details of the relative velocity measurement are as described in the flowchart illustrated in FIG.

3) Image frame-relative distance graph creation In this process, a relative distance graph is created by simply accumulating relative speeds.
FIG. 11 shows a processing flow of distance calculation. FIG. 12 shows the result of executing the relative distance calculation program. The unit of relative distance is a pixel.

4) Matching of facility / equipment position and relative distance In the matching process of facility / equipment position and relative distance, the relative distance obtained by the above relative distance calculation is associated with the kilometer of the corresponding facility / equipment. Correspondence data about the kilometer of the facility / equipment corresponding to the relative distance can also be created manually from the facility / equipment kilometer correspondence data and the image frame-relative distance graph.
FIG. 13 is a table showing an example (virtual data) of correspondence data of kilometres of facilities / equipment corresponding to relative distances.

参照データの得られているＦ０：Ｆ１，Ｆ１：Ｆ２ごとに相対距離と参照データから得られる実際の距離の比率によって距離を変換する。 5) Image frame-kilometer graph creation In this image frame-kilometer graph creation process, the image frame-relative distance graph is enlarged / reduced for each section with facilities and facilities kilometer, and the image frame-kilometer graph is created. Create
That is, as shown in the following conversion formula, the relative distance di of the image frame-relative distance graph of FIG. 14A is obtained using the reference data {(if, Af)}, and the image frame-km Di is obtained. The graph of the image frame-km shown in FIG. 14B is created.

The distance is converted by the ratio of the relative distance and the actual distance obtained from the reference data for each of F0: F1 and F1: F2 where the reference data is obtained.

  FIG. 15 is an example of an image frame-kilometer graph obtained as a result of the execution of the image frame-kilometer graph creation program.

(2) Rail shape extraction For example, when an idiosyncratic image is displayed as an overlay on a forward shot image, it cannot be correctly performed unless the distance from the idiomatic rail (distance from the center of the rail) and the height are obtained. The rail shape extraction program extracts a rail shape from a front shot image, and uses the extracted rail shape data to display a spontaneous image at the correct position on an image having only two-dimensional information. It is what makes it possible.

  The rail shape extraction program uses a rail shape extraction program based on an existing system, but has been improved to obtain an accurate rail shape by paying attention to rail curvature conditions and rail width conditions.

FIG. 16 shows the flow of rail shape extraction. That is, in the rail shape extraction, the following processing is performed.
1) Rail shape extraction by existing system
2) Create point cloud data on the screen rail
3) Creating point cloud data on real space rails
4) Creation of distance data to point cloud on real space rail
5) Creating alignment point cloud data on real space rails
6) Creation of point cloud data and rail angle data on the center line of real space rails

と表す。 1) Rail shape extraction by the existing system The left and right rail shapes on the screen extracted by the existing system are used as a function of the horizontal direction with respect to the screen height.

It expresses.

  FIG. 17 shows the result of rail shape extraction by the existing system.

A function that maps the position on the screen to the actual rail position is F, and its inverse function is F- ¹ . In this function, the camera position with respect to the rail (height from the rail, position w in the rail orthogonal direction with respect to the rail center), camera depression angle θ (pitching), other rolls, and yawing angle are set to zero. The rail shape is obtained from the camera focal length f and the camera CCD size. Provided that the rail is horizontal. The inclination of the rail is slight and it is difficult to accurately measure this. On the other hand, even if it is assumed to be horizontal, it is considered that this does not have a significant effect.

2) Create point cloud data on the screen rail
3) Creation of point cloud data on real space rails As shown in FIG. 18 (a), numbers are assigned from the bottom edge of the screen at equal intervals over the left and right rails on the screen at equal intervals, and the point cloud {r _L (i )}, {r _R (i)}. As shown in FIG. 18B, this point is converted into a point group {R _L (i)}, {R _R (i)} on the real space by F.

によって求められる。これにより、Ｒ_L(０)とＲ_R(０)からｄ離れたレール位置が左右レールに設定した点群の何番目と何番目の間にあるかを知ることができる。また、図１９,２０に示すように、距離ｄ離れたレール位置の近似値を

と表現することができる。 4) Creation of distance data to the point cloud on the real space rail Using the positions of {R _L (i)} and {R _R (i)} in Fig. 18 (b) The distance to each point is obtained by accumulating the distances between them. This means that when the distance from R _L (i + l) to R _L (i) is | R _L (i + l): R _L (i) |

Sought by. As a result, it is possible to know the position of the point group set on the left and right rails and the position between the rail positions d apart from R _L (0) and R _R (0). Also, as shown in FIGS. 19 and 20, the approximate value of the rail position separated by the distance d is

It can be expressed as

である。j番目のセンタ位置は距離jlだけ離れた左右のレール位置から

と求めることができる。θ(ｊ)をレールの絶対角度、Δθ(ｊ)を対象地点でのレール曲率とする。初期値として、たとえば
ｃｏｓθ（０）＝１
ｃｏｓΔθ（０）＝1
と曲率を設定し、j＝1、2と順々にθ(j)、Δθ(j)を求めていく。j番目とj＋１番目のレールセンタ位置について、Ｒｃ(j)とＲｃ(j＋１)間の距離はlであるので、Ｒｃ(j)からθの向きにlだけ伸ばしたベクトルをＶ、Ｒｃ(j)からＲｃ(j＋１)までのベクトルをＶ’とすると、

によってＲｃ(j)でのレール曲率が求められる。θの正負はＶとＶ’のベクトルの終点の位置関係から求められる。
ただし、この値は誤差を含んでいるため、曲率の最大値をΔθmaxとするとき、

として曲率の設定上限を超えることがないようにする。また、曲率変化の連続性を考慮して

と曲率が徐々に更新されるようにする。これによりＲｃ(j＋１)でのレールの向きは

で求められ、Ｒｃ(j＋１)でのレールセンタ位置は

と与えられる。これにより、レールの最大曲率条件、曲率の変化率を考慮したレールセンタ位置を求めることができる。 5) Creating alignment point cloud data on real space rails
6) Creation of point group data and rail angle data on the center line of the real space rail Find the position of the rail center and the rail angle in the real space. Set points on the rail center so that the unit distance is equal to 1 in real space, and the number is j. As for the center of the rail, the position of the lower end of the screen is

It is. The j-th center position is from the left and right rail positions separated by a distance jl

It can be asked. Let θ (j) be the absolute angle of the rail and Δθ (j) be the rail curvature at the target point. As an initial value, for example
cos θ (0) = 1
cosΔθ (0) = 1
The curvature is set, and θ (j) and Δθ (j) are obtained in order of j = 1 and 2. Since the distance between Rc (j) and Rc (j + 1) is l for the j-th and j + 1-th rail center positions, V, Rc (j) are vectors obtained by extending l from Rc (j) in the direction of θ. And the vector from Rc (j + 1) to V ′,

Is used to obtain the rail curvature at Rc (j). The sign of θ is obtained from the positional relationship between the end points of the vectors V and V ′.
However, since this value includes an error, when the maximum value of curvature is Δθmax,

To avoid exceeding the upper limit of curvature. Also consider the continuity of curvature change

And let the curvature be gradually updated. As a result, the direction of the rail at Rc (j + 1) is

The rail center position at Rc (j + 1) is

And given. Thereby, the rail center position in consideration of the maximum curvature condition of the rail and the rate of change of the curvature can be obtained.

と表現できる。レールの向きも

で求めることができる。 (3) Derivation of installed object position in the image
In order to perform this process, it is necessary to convert the position of the installation object, for example, the spontaneous position from the absolute coordinate system to the forward image coordinate system. Therefore, the installation object position is mapped from the rail position and camera conditions obtained by the rail extraction process. The rail center position away from Rc (0) by the distance d is

Can be expressed. Rail orientation

Can be obtained.

As shown in FIG. 21, when the distance d from the Rc (0) of the installation object (spontaneous) to be installed is obtained, the direction of the rail at that position is obtained from the above formula, The installation position in the real space with the camera position as a reference is obtained from the value. By moving this position to a position on the screen by F- ¹ , a position on the screen (spontaneous position p) is obtained.

の点群データが得られる。
この点群をＦ^-１で画面上に写すと、レール幅１０６７ｍｍの制約条件を満たすレール曲線が得られる。 [Correction of rail shape on screen]
As shown in FIG. 22, the corresponding left and right rail positions are positions moved by 1067/2 [mm] from the rail center position R′c (j) to the left and right directions orthogonal to the direction θ.

Point cloud data is obtained.
When this point cloud is copied onto the screen with F ⁻¹ , a rail curve that satisfies the constraint condition of the rail width of 1067 mm is obtained.

(4) Overlay Display of Front Shooting Image and Installation Object Image An image of the specified installation object is displayed by chroma key (image composition) at the position on the screen obtained as described above.

  Subsequently, the installation visibility preliminary confirmation work performed using the simulator VCA will be described consistently.

  In the simulator VCA in which the forward image data is stored, for example, when a special occurrence is newly established outside a specific railroad crossing on the target route, an operation for confirming the visibility at a desired installation point in advance is shown in FIG. explain.

  When the visibility confirmation switch 42 of the input unit 4 of the simulator VCA is turned on, the image processing apparatus 2 reads out the frame image of the starting point (position information P0) of the route from the image data storage unit 32 and displays it on the monitor 5 (p41). . By looking at this image display, it is possible to confirm whether or not the route is a special installation target route.

  Subsequently, it is determined whether any point (position information P1 and P2) is designated by the point designation unit 43 (p42), and if designated, the line of the designated point (position information P1 and P2) Pi The image is displayed on the monitor 5 (p43). Normally, when a code of a railroad crossing to be specially installed is first input to the point designating unit 43, it is converted into position information of the railroad crossing corresponding to the code, so that an image of the railroad crossing is displayed on the monitor 5. Display and check the scene of the railroad crossing. Subsequently, when the point designation unit 43 is operated to input a viewing point P1 that is 600 to 800 m away from the level crossing as an example of the point where the special setting is desired, the image of the currently displayed level crossing is displayed. It changes to the track image of the visual recognition point P1 of 600 to 800 m outside the point. Without displaying the image of the railroad crossing, the track image of the visual recognition point P1 that is 600 to 800 m outside the point where the special occurrence is desired may be displayed directly. FIG. 24 is an example of a track image of a viewing point P1 800 m outward from the desired installation point.

  Next, it is determined whether or not the installation installation desired point P2 is specified in the point specifying unit 43 (p44). When specified, the line image data of the specified installation installation desired point P2 is stored in the image data storage unit 32. Is displayed on the monitor 5 (p45). Subsequently, the image processing apparatus 2 determines whether or not an installation object is designated from the installation object designating unit 44 of the input unit 4 (p46), and when the installation object is designated, the installation object image data storage unit is designated. The image data of the installation is read from 35, and the height, shape, installation position (separation) of the installation is determined by the display ratio adjusting unit 23 according to the distance from the viewing point P1 to the installation installation desired point P2. The ratio of the display ratio is calculated (p47), the image of the installation object is reduced based on the ratio of the calculation result, and the image of the reduced installation object is installed by the image composition unit 24 as shown in FIG. An overlay display is performed on the forward photographed image of the visual recognition point P1 outside the desired point (p48). Therefore, on the monitor 5, the image of the designated installation object is synthesized and displayed on the designated installation desired point P2 inward of the visual recognition point P1 and the image captured in front of the visual recognition point P1 (p49).

  FIG. 25A shows a state in which the image of the installation object designated at the desired installation point P2 specified inward of the visual recognition point P1 is synthesized and displayed on the monitor 5 in front of the visual recognition point P1. Show. From the display state of the monitor 5, it is possible to determine whether or not the visibility of the installation object at the site is good from the visibility of the image of the installation object (p 410). When the viewing point is gradually brought closer to the inside, as shown in FIG. 25 (b), the idiosyncratic image of the monitor is enlarged and displayed according to the approach distance. It can be confirmed whether or not it is suitable. If it matches, the installation planned position is stored (p411). If the idiosyncratic image cannot be confirmed sufficiently, re-enter the desired installation point (p412), confirm the visibility again (p413), and if the visibility is good, set the desired installation point as the planned installation position. The image at that time may be saved (p414). Instead of storing the image, the planned installation position may be recorded on the paper surface in kilometres.

  By preserving the image of the installation position or recording the paper, the work for confirming the visibility of the installation object is completed.

[Other embodiments]
When displaying the image of the viewing point together with the position information of the point, there is an advantage that it is easy to recognize or record the visible range, but the position information of the range with good visibility is automatically saved. If it can be confirmed later, it is not always necessary to display position information on the image of the viewing point.

  As described above, in the above-described embodiment, it is possible to achieve the objectives of prior confirmation of the visibility of the installed object and improvement of the construction speed.

  In the above-described embodiment, the present invention is applied to the prior confirmation of the visibility of the installed object. However, the present invention can be applied to other applications such as a car stopped at a railroad crossing, a falling rock falling on a railroad track, and a person falling on a campus railroad track. Or, it is possible to provide a driving simulator for vehicle driver training for checking the visibility of obstacles in advance, that is, assuming an abnormal time such as the occurrence of an obstacle such as a falling object. .

  In this case, the installation object designating unit 44 of the input unit 4 in FIG. 1 is an obstacle designating unit, and the installation object image data storage unit 35 of the storage unit 3 is an obstacle image data storage unit. Obstacles at the designated point where obstacles are expected to be displayed by displaying the CG of obstacles (cars, falling rocks, falling homes, etc.) superimposed on the actual video of the front shot image by the same operation as in the case of prior confirmation. Can be confirmed.

  In the above configuration, when the obstacle designating unit is not operated, it can be used as a normal driving simulator without any abnormality.

Claims (4)

Using the camera provided on the vehicle, the front shot image data obtained by photographing the front of the track in the vehicle traveling direction at predetermined short time intervals and the pre-created facility / equipment-km data are used. To create a kilometer data representing the kilometer corresponding to each frame of the image,
On the other hand, a rail shape extraction process for generating rail shape data by extracting a rail shape from the front shot image data is performed,
Based on the facility / equipment-kilometre correspondence data, the kilometrage data, and the rail shape data, a desired installation position derivation process is performed to derive a desired installation position derivation process for deriving a desired installation position. make,
Based on the desired installation position data, the front-captured image data of the position where the installation object is desired is read out, and the installation image data stored in advance is read out, and the installation object position is desired. Performing an installation display process for displaying an image of the installation on the front photographed image that is a real image of
A method for confirming the visibility of an installation in advance.   The visibility confirmation method for an installed object according to claim 1, wherein the image of the installed object includes an actual image, an image created by CG, an illustration, and the like. Method. A camera that is provided on a vehicle and captures the front in the traveling direction of the vehicle traveling on the track at a predetermined short time interval, a storage unit that stores various data, an image processing device that processes forward captured image data input from the camera, An input unit for inputting a control command to the image processing apparatus, and a monitor for displaying an image after being processed by the image processing apparatus;
The storage unit stores a first captured image data input from the camera and image data after image processing by the image processing apparatus, and a first storage unit storing facility / equipment-kilometre correspondence data. 2 storage units, a third storage unit that stores rail shape data generated by rail shape extraction processing by the image processing device, and a fourth storage unit that stores image data of the installed object,
The image processing apparatus includes: a storage / reading control unit that controls storage of data in the storage unit and reading of data from the storage unit; a starting point on the line image stored in the first storage unit; Calculate the kilometer to the shooting point (which is also the viewing point), extract the rail shape from the front shot image data, the display ratio adjustment unit to adjust the display ratio of the image of the installation, the visual recognition of the track An image composition unit that synthesizes an image in which the display ratio of the installation object is adjusted to the image of the point,
The input unit is for designating a point designating unit for designating a point (about kilometer) on the track image to be read out to the second storage unit, and an installation stored in the fourth storage unit. The installation designation part of
The image processing device reads an image of a viewing point designated by an operation of the point designation unit from the first storage unit and displays it on a monitor, and rail shape data obtained by rail shape extraction processing from the image of the point And facility / equipment-Deriving the desired location of the installation based on the kilometer data and the kilometer data, and reading out the image data of the installation designated by the operation of the installation designation unit, and monitoring The image of the installation object is displayed on the image of the viewing point displayed on the screen at a ratio adjusted by the display ratio adjusting unit according to the kilometer of the installation object desired point. A simulator for using the method for confirming the visibility of installed objects according to 2. A camera that is provided on the vehicle and captures the front of the track at predetermined short time intervals, a storage unit that stores various data, an image processing device that processes forward captured image data input from the camera, and controls the image processing device An input unit for inputting a command, and a monitor for displaying an image after being processed by the image processing device;
The storage unit stores a first captured image data input from the camera and image data after image processing by the image processing apparatus, and a first storage unit storing facility / equipment-kilometre correspondence data. 2 storage units, a third storage unit that stores rail shape data generated by rail shape extraction processing by the image processing device, and a fourth storage unit that stores image data of obstacles,
The image processing apparatus includes: a storage / reading control unit that controls storage of data in the storage unit and reading of data from the storage unit; a starting point on the line image stored in the first storage unit; Calculate the kilometer to the shooting point (which is also the viewing point), extract the rail shape from the front shot image data, display ratio adjustment unit to adjust the display ratio of the obstacle image, and display on the monitor An image composition unit that overlays an image in which the display ratio of the obstacle is adjusted on the image of the viewing point of the track,
The input unit designates a point designation unit for designating a point (about kilometer) on the track image to be read with respect to the second storage unit, and an obstacle stored in the fourth storage unit With an obstacle designation part,
The image processing device reads an image of a viewing point designated by an operation of the point designation unit from the first storage unit and displays it on the monitor, and a rail shape obtained by rail shape extraction processing from the image of the point Based on the data and facilities / equipment-kilometre correspondence data and kilometer data, the position where the obstacle is desired is derived, and the image data of the obstacle designated by the operation of the obstacle designating unit is read out. The obstacle image is displayed on the image (actual video) of the viewing point displayed on the monitor at a ratio adjusted by the display ratio adjusting unit according to the kilometer of the estimated obstacle occurrence point. Simulator characterized by.

Images