JP2007122247A - Automatic landmark information production method and system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an automatic landmark information production method and system, capable of automatically generating a visual landmark by a stereoscopic image, and disposing it in a prescribed position on a map. <P>SOLUTION: This automatic landmark information production method automatically forming landmark information has: a shape setting step for generating the landmark of the stereoscopic image by use of pick-up images obtained by imaging an object from a plurality of directions, and specifying appearance shape of the object from the plurality of pick-up images by use of the plurality of pick-up images of the object each obtained by imaging the object from a specific place to a specific direction when disposing the generated landmark in the prescribed position on the map; a landmark generation step for generating the landmark from the appearance shape found in the shape setting step; a position setting step for detecting a position of the object from the plurality of pick-up images; and a landmark disposal step for disposing the generated landmark in the detected position of the object on the map. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an automatic landmark information creating apparatus and system for automatically generating landmarks (marks on a map) composed of 3D images such as 3D and automatically arranging them at predetermined positions.

Conventionally, a technique for arranging and displaying landmarks on a map for map guidance has been proposed (see, for example, Patent Document 1).
JP-A-10-333562

However, it takes enormous effort and time to create a variety of landmark figures. For example, it is desirable that the shape of the landmark accurately reflects the shape of the landmark building, etc., but in order to create such a landmark, take a picture of the building, It needs to be analyzed, and data collection alone is very time consuming, labor intensive, and expensive.
In addition, when a landmark is expressed as a 3D stereoscopic image, it is necessary to take pictures from various angles, and the burden of creating the landmark increases. Furthermore, it can be said that creating landmarks such as famous buildings is relatively easy, but for example, creating detailed maps with specific details in a specific local area (such as a very small area in a specific region). In some cases, small buildings along a small path (mountain-covered mountain roads, loopholes, narrow roads between buildings that are hidden behind high-rise buildings, etc.) cannot be seen in aerial photographs. If you are trying to create a visual landmark using 3D images, it is necessary for the full-time person to search all over the area and take pictures of the landmark buildings. It becomes.

Even after the map was created, the situation was constantly changing due to changes in topography, new construction and removal of buildings, especially in old buildings where houses under construction were completed when viewed microscopically in a small area. Minor changes in the situation occur on a daily basis, such as a demolished building, a new building on the site, and a bus stop moving.
Therefore, if such a change in the situation cannot be followed, depending on the use of the map, there is a concern that the information on the map becomes obsolete. In other words, the landmarks displayed on the map also need to be updated following changes in the situation, and enormous time, effort, and cost are required for the update.

  The present invention has been made in view of such a situation, and an object of the present invention is to generate automatic landmark information that can automatically generate visual landmarks based on stereoscopic images and automatically place them at predetermined positions on a map. It is to provide a method and system.

The above object of the present invention is achieved by the following configuration.
(1) An automatic landmark information generation method for generating a landmark consisting of a three-dimensional image using captured images obtained by capturing an object from a plurality of directions, and arranging the generated landmark at a predetermined position on a map. Using a plurality of captured images of the object imaged from a specific location in a specific direction, and determining the external shape of the object from the plurality of captured images, and obtaining the shape by the shape setting step A landmark generation step for generating the landmark from the outer appearance shape, a position setting step for detecting the position of the object from the plurality of captured images, and the generation at the position of the detected object on the map. And a landmark arrangement step for arranging the landmarks. An automatic landmark information generation method comprising:

According to this automatic landmark information creation method, a plurality of captured images of an object captured in a specific direction from a specific location are used, and an external shape of the target is specified from the captured images. A landmark of a stereoscopic image is generated from the external appearance shape, the position of the object is detected from a plurality of captured images, and the generated landmark is arranged at the detected position on the map, thereby It can be automatically generated and automatically placed at a predetermined position on the map. As a result, it is possible to automatically create a newest map including visual landmarks based on 3D images without much effort and cost.
For example, a captured image is narrowed down under a predetermined condition, and a building or the like (a building having a size larger than a predetermined size, for example, a three-story building) that is a common landmark among a plurality of images after the narrowing down House, hotel, shrine, etc.), analyze the image (including other additional information) for the building, etc., detect the position and shape of the building, etc., and based on the shape detection, A 3D image can be formed by computer graphics (CG), and the 3D image can be arranged as a landmark on a map surface.

(2) The automatic landmark information creation method according to (1), wherein the shape setting step represents an imaging position on the map and an imaging direction with respect to each of a plurality of captured images obtained by capturing the object. An automatic landmark information generating method, wherein a three-dimensional shape of the object captured in common in the captured image is detected using a roll angle and a pitch angle.

  According to this automatic landmark information creation method, by using the imaging position on the map, the azimuth angle indicating the imaging direction, the roll angle, and the pitch angle for each of the plurality of captured images obtained by capturing the target object, It is possible to geometrically detect the three-dimensional shape of the object imaged in common.

(3) The automatic landmark information creation method according to (1) or (2), wherein the position setting step includes an imaging position and an imaging direction on the map for each of a plurality of captured images obtained by imaging the object. An automatic landmark information generating method, wherein a position of the object imaged in common in the captured image is detected using an azimuth angle, a roll angle, and a pitch angle representing

  According to this automatic landmark information creation method, by using the imaging position on the map, the azimuth angle indicating the imaging direction, the roll angle, and the pitch angle for each of the plurality of captured images obtained by capturing the target object, The position of the object being imaged in common can be detected geometrically.

(4) The automatic landmark information creation method according to any one of (1) to (3), wherein the captured image is obtained by each of the imaging units attached to each of the plurality of moving bodies. An automatic landmark information creation method characterized by

  According to this automatic landmark information creation method, each imaging unit attached to each of a plurality of moving bodies provides a captured image, so that there is an information source corresponding to the number of moving bodies. Information such as a plurality of captured images at the place can be efficiently obtained in a short time.

(5) The automatic landmark information creation method according to any one of (1) to (4), wherein imaging date / time information indicating an imaging time point is added to the captured image, and among the generated landmarks, Extract a specific landmark created between the creation time of the existing map prepared in advance and a predetermined time after the creation time, and compare the existing landmark existing in the existing map with the specific landmark When there is no existing landmark corresponding to the specific landmark, the specific landmark that does not exist is newly arranged on the existing map, and when the specific landmark corresponding to the existing landmark does not exist The automatic landmark information creating method, wherein the existing landmark that does not exist is deleted from the existing map.

  According to this automatic landmark information creation method, the newly created landmark and the landmark to be deleted are determined by comparing the generated specific landmark with the existing landmark on the existing map, and the map is updated. By doing so, new information can be included in the map.

(6) An automatic landmark information creation system that arranges the landmarks at predetermined positions based on the automatic landmark information creation method according to any one of (1) to (4), wherein A position detection unit that detects a position of the object, a shape detection unit that detects a shape of the object from the captured image, a landmark generation unit that generates a landmark from the detected shape of the object, and An automatic landmark information creation system, comprising: a landmark placement unit that places a landmark obtained by a landmark generation unit at a position detected by the position detection unit.

  According to this automatic landmark information creation system, the position detection unit detects the position of the object from the captured image, the shape detection unit detects the shape of the object from the captured image, and the landmark generation unit detects the detected object. A landmark is generated from the shape of the object, and the landmark obtained by the landmark generator is placed at the position detected by the position detector by the landmark placement unit, so that a visual landmark based on a stereoscopic image is automatically generated. And can be arranged at a predetermined position on the map.

(7) The automatic landmark information creation system according to (6), further including a map correction unit that updates the existing map based on the automatic landmark information creation method according to (5). Landmark information creation system.

  According to this automatic landmark information creation system, the newly created landmark and the landmark to be deleted are determined by comparing the generated specific landmark with the existing landmark on the existing map, and the map is updated. By doing so, new information can be included in the map.

  According to the automatic landmark information creation method and system according to the present invention, a visual landmark based on a stereoscopic image is automatically generated and arranged at a predetermined position on a map, and the latest map including the landmark is reduced in labor. It can be created automatically without difficulty and at a reasonable cost.

Next, preferred embodiments of an automatic landmark information creating method and system according to the present invention will be described in detail with reference to the drawings.
The automatic landmark information creation method of the present invention generates a landmark composed of a three-dimensional image using captured images obtained by capturing an object from a plurality of directions, and places the generated landmark at a predetermined position on a map. In addition, a plurality of captured images of the target imaged from a specific location in a specific direction are used, and a shape setting step for specifying the appearance shape of the target object from the plurality of captured images and a shape setting step A landmark generation step for generating the landmark from the outer appearance shape, a position setting step for detecting the position of the object from a plurality of captured images, and the generated landmark on the map. And a landmark placement step for placing the landmark at a position.

(First embodiment)
Hereinafter, the details of the above steps and the system configuration example will be described.
FIG. 1 is a schematic configuration diagram showing an example of an automatic landmark information creation system of the present invention.
As shown in the figure, movement information collection devices 16a and 16b are attached to a transport vehicle 14a such as a courier as a moving body and a vehicle 14b such as a taxi, respectively. The movement information collection devices 16a and 16b include An imaging unit, a triaxial geomagnetic sensor, a GPS receiver, a wireless communication device, and the like are mounted.

  The transport vehicle 14a, the taxi vehicle 14b, or the public bus is a moving body that may repeatedly travel on the road 12 in the common area almost every day, and the movement information collecting device is provided on the moving bodies 14a and 14b. By installing 16a and 16b, an image around the road 12 can be taken on a daily basis without being particularly conscious.

  In the present invention, the moving body may be a motorcycle or a bicycle in addition to the vehicle as described above, and other items such as a helmet that covers a person or a person's head, a pet animal (for example, a pet animal) Can be handled as a moving object. The number of moving bodies is not limited to two as in the illustrated example, and it is preferable to use a large number of moving bodies simultaneously. The presence of a plurality of mobile information collection devices increases the amount of information collected and enables efficient collection of necessary information in a short period of time.

  Here, the area where the image is captured, the imaging target, the imaging timing and the like can be set in advance. For example, when traveling on a road 12, a condition that an image of the vicinity of the building 10 along the road 12 must be captured is set in the movement information collection devices 16a and 16b. When it is detected that the movement information collection devices 16a and 16b) are approaching the building 10, the imaging is automatically performed. The captured image at this time may be a moving image or a still image. This makes it possible to capture a desired image without making the driver aware of the driver of the transport vehicle 14a or the taxi vehicle 14b and without applying any burden.

  The movement information collection devices 16a and 16b attached to the transport vehicle 14a and the taxi vehicle 14b add the attitude angle (direction on the horizontal plane) of each imaging unit detected by the triaxial geomagnetic sensor to the image information captured by the imaging unit. Azimuth angle, roll angle representing the tilt angle in the vertical plane, pitch angle that is the tilt angle in the vertical direction from the horizontal plane) information, and current position information of the movement information collection device measured by the GPS receiver. Is transmitted wirelessly as cartographic information.

  The transmitted image information (including posture angle information and current position information) is sent to a map creation device (map server) 20 via a relay station 18 and a communication network (such as the Internet) 19.

  The map server 20 includes a communication unit 21, an image storage unit 22, a condition setting unit 23, a position detection unit 24, a shape detection / 3D image creation unit (shape detection unit, landmark generation unit) 26, a landmark An information database 27, a landmark arrangement unit 28, an output unit 29, a map database 30, and a map correction unit 31 are included.

  The image storage unit 22 stores the image information transmitted from the movement information collection devices 16a and 16b. The landmark information database 27 stores information on the position and shape of the 3D image landmark. The map database 30 stores the created map (or updated map).

  The condition setting unit 23 is used to set conditions for narrowing down the landmark creation area and the type, shape, etc. of the subject for which the landmark is to be created. That is, innumerable subjects are recorded in the image stored in the image storage unit 22, and it is not realistic to represent all of these subjects as objects and to represent the landmarks of the 3D image. Therefore, it is possible to narrow down the objects that are landmark candidates by setting the area where the landmark is to be created and the type, size, shape, etc. of the target subject via the condition setting unit 23. It is what. This point will be specifically described later.

  The object position detection unit 24 analyzes the map creation information (including captured images, posture angle information and current position information) received from the movement information collection devices 16a and 16b, received by the communication unit 21, The position of the target object as a mark candidate is detected. This position detection operation will be described later with reference to FIGS.

  The shape detection / 3D image creation unit 26 becomes a landmark candidate from the images captured by the movement information collection devices 16a and 16b and the azimuth angle, roll angle, and pitch angle of the imaging units of the movement information collection devices 16a and 16b. The shape of the subject (object) is detected, and a 3D image is created by computer graphics (CG) technology. The created 3D image is accumulated in the landmark information database 27 and sent to the landmark arrangement unit 28.

  The landmark arrangement unit 28 arranges the landmark that has been made into a three-dimensional image by the CG technique at the position detected by the position detection unit 24 on the map surface, and automatically creates the latest map.

  The created map is stored in the map database 30 and is displayed on a display (not shown) or printed by a printer (not shown) via the output unit 29. In FIG. 1, reference numeral 100 indicates a printed map, and LM indicates a landmark of the building 10 by a 3D image.

  The map correcting unit 31 corrects the map. The operation of the map correction unit 31 will be described in the second embodiment.

  As described above, in the automatic landmark information creation system of the present embodiment, the movement information collection devices 16a and 16b are mounted on the moving bodies 14a and 14b that may travel repeatedly in a common area, and the daily routine of a predetermined area is performed. By capturing a landscape or the like almost regularly, it is possible to easily collect a large number of images in a predetermined area (attention area) without imposing a burden on a specific person.

  Then, the map server 20 analyzes the image, extracts a subject as a landmark candidate, detects a position and shape, and creates a 3D image of the shape by CG. By automatically placing on the map surface, a new map (especially a detailed and detailed map from the viewpoint of the local consumer) can be automatically created without difficulty and with low cost. it can.

FIG. 2 is a block diagram showing an internal configuration of the movement information collection device of FIG.
As shown in the figure, the movement information collection device 16 (16a, 16b) mounted on the moving body 14 (14a, 14b) includes a wireless antenna AN, a transmission / reception unit 40, a data processing unit 42, an imaging unit (CCD). Camera) 44, image processing unit 46, control unit 48, GPS receiver 50, timing unit 51, triaxial geomagnetic sensor (triaxial gyro) 52, and attitude angle calculation unit 54. Note that the time measuring unit 51 is provided for the purpose of detecting the time point of imaging.

  FIG. 3 is a block diagram showing an internal configuration of the map server of FIG. In FIG. 3, the same reference numerals are given to the portions common to the above-mentioned drawings.

  As illustrated, the communication unit 21 includes a communication front end 60 and a data processing unit 61. The control unit 63 comprehensively controls the operation of the map server 20.

  In addition, an image storage unit 22, a landmark information database 27, an output unit 29 (including a display unit 37 and a printing unit 39), a map database 30, and a ROM 35 (a landmark creation / arrangement program 38) are mounted. ) And an operation key 62 are provided. The operation key 62 functions as a means for the operator to input a desired condition to the condition setting unit 23 in FIG.

  Next, a method for detecting the position and shape of an object that is a landmark candidate will be described in detail with reference to FIGS.

  FIG. 4A is a diagram for explaining a method of specifying the position of an object from the positions of two cameras and the attitude angles (azimuth angle, roll angle, pitch angle) of each camera, and FIG. These are the figures for demonstrating the method to identify the shape of a target object from the captured image by each camera, and the attitude angle (azimuth angle, roll angle, pitch angle) of each camera.

  As described above, the map creation information sent from the movement information collection devices 16a and 16b is the captured image and the position and posture angle information of the camera itself, and the position and shape of the object are unknown. . For this reason, it is necessary to recognize the object by arithmetic processing or image processing.

First, the position of the object can be calculated as follows.
In FIG. 4A, the imaging units of the two moving bodies correspond to the movement information collection devices 16a and 16b in FIG. 1, but here, a simple model will be described. Assume that the two imaging units are A and B, respectively, and the building that is the object is C.

  In FIG. 4A, after extracting an object as a landmark candidate, first, a line segment AB is obtained from the position information of each imaging unit A, B measured by the GPS receiver (reference numeral 50 in FIG. 2). (The distance between the cameras) X and the direction of the line segment AB are calculated. Next, a line connecting each imaging unit A, B and the object C based on the orientation, roll angle, and pitch angle of each imaging unit A, B measured by a triaxial geomagnetic sensor (reference numeral 52 in FIG. 2). Minute angles Aθ and Bθ are calculated. As a result, the vertex of the triangle (one point on the object C) whose base is X is uniquely determined, whereby the absolute position of the object C is obtained.

  The above procedure for position detection is summarized as shown in FIG. FIG. 5 is a flowchart showing a processing procedure for detecting the position of an object to be a landmark.

  That is, first, a common object is extracted from the feature amounts of the captured images from the imaging unit A and the imaging unit B (step 201: S201), and then the imaging unit is obtained from the positional information of the imaging unit A and the imaging unit B. The distance X between A and the imaging unit B and the direction of the line segment AB are obtained (S202).

  Next, the orientation AC between the object C and the imaging unit A is determined from the azimuth angle (direction), roll angle, and pitch angle of the imaging unit A (S203). Similarly, the azimuth angle (direction) of the imaging unit B is determined. The orientation BC between the object C and the imaging unit B is determined from the roll angle and the pitch angle (S204).

  Next, an angle BAC (Aθ) is determined from the vector AB and the direction AC (S205), and similarly, an angle ABC (Bθ) is determined from the vector AB and the direction BC (S206).

  Next, the position of the object C is specified from the distance X, the direction AB, the angle BAC (Aθ), and the angle ABC (Bθ) (S207). Then, it is detected whether or not there is another set of images in which the same object C is projected (S208). If there is a detection, the process returns to S201 and the above processing is executed to detect the position. If the accuracy is improved and no detection is made, the position detection process is terminated.

  Next, a procedure for detecting the shape of the object C will be described. The shape of the object C can be obtained as follows.

  In FIG.4 (b), P and Q have each shown the image imaged with the imaging parts A and B of two movement information collection apparatuses. As shown in the figure, the common building C (reference numeral 10 in FIG. 1) is shown in both images (P, Q), but the shapes that can be seen because the posture of each imaging unit with respect to the object is different. Is different. Therefore, the appearance of the object C is corrected by correcting the captured image of each imaging unit with reference to the azimuth, roll angle, and pitch angle of each imaging unit A and B, and then performing image processing such as contour extraction. The shape can be detected. Then, the detected shape of the object C is expressed as a 3D image composed of polygons using computer graphics (CG), and the 3D image is stored.

  FIG. 6 summarizes the procedure for detecting the shape of the landmark (object) C and 3D imaging. FIG. 6 is a flowchart showing a processing procedure for shape detection and 3D imaging of a landmark (object).

  First, a common object is extracted from the feature quantities of a plurality of images (S301), and the position, direction, roll angle, and pitch angle of the extracted object image are acquired (S302), and the shape of the object is analyzed. (S303).

  If there is another image in which the same object is imaged, the same processing is repeated using the image to improve the accuracy of shape detection (S304).

  Then, the detected shape of the object C is represented by polygons using computer graphics (CG) to create a 3D image (S305), and the orientation and size of the created 3D image are set to the land. It is recorded in the mark information database 27 (S306).

  As described above, the position of the object can be specified based on a plurality of images by referring to the position information of the imaging units A and B measured by the GPS receiver (reference numeral 50 in FIG. 2). In addition, the shape of the object can be specified based on the captured images of the imaging units A and B and the information on the azimuth angle, roll angle, and pitch angle of the imaging units A and B. Therefore, the position information and the shape information of the object C, which is a landmark necessary for posting on the map, are automatically acquired based on the image information sent from the imaging unit, and the landmark consisting of the 3D image. Can be created.

Next, a series of operation procedures for the entire automatic landmark information creation system in FIG. 1 will be described.
FIG. 7 is a flowchart illustrating an example of a series of procedures from imaging by the imaging unit of the movement information collection device to arrangement of landmarks including 3D images on a map surface.

  First, image information is transferred from the imaging units of the movement information collection devices 16a and 16b attached to the moving body (S101), together with this image, shooting information (information on current position information, azimuth angle, roll angle and pitch angle). Are transferred together (S102). Then, the image and shooting information are stored in the map server 20 on the network (S103).

  On the other hand, in parallel with the above processing, the operator of the map server 20 sets the conditions of the landmarks to be created / placed via the condition setting unit 23 shown in FIG. 1 (S104). In other words, an infinite number of subjects are shown in the image captured by the imaging unit, and it is not realistic to capture the landmarks for all of them, so it is necessary to set conditions in order to narrow down the objects that are landmark candidates. Is to do. In FIG. 7, a point (or area) where a landmark is to be created and the type of landmark are specified. For example, “designate a building of a specific station on an existing map as a creation target of a landmark of a 3D image”, or “specify in a region within a circle with a radius of 200 m centering on a specific intersection 3 “Specify a building having a height higher than the floor as a creation target of the landmark of the 3D image”.

  The position detection unit 24 of the map server 20 refers to the position coordinates of the image capturing unit, extracts the point designated in S104 (captured image near the area (S105), and uses the extracted plurality of images in common. An object that is a candidate for a landmark (an object having a predetermined size or larger and an object having a predetermined feature) is extracted (S106).

  Next, the position and shape of the extracted landmark candidate object are specified by the method described with reference to FIGS. 5 and 6 (S107). Then, a 3D image is generated by CG (S108). The landmark of the generated 3D image is stored in the landmark information database 27.

  Then, the same processing is performed for the next landmark candidate object (S109).

  In S110, the landmarks of the plurality of generated 3D images are taken out, it is determined again whether or not the conditions specified in S104 are met, and only the landmarks that match the conditions are extracted.

  And the landmark which consists of 3D images is arrange | positioned on a map (S111). In this way, the latest visual map including landmarks made up of 3D images is automatically created.

  As described above, in this automatic landmark information creation system, mobile vehicles that may repeatedly travel in a common area, such as delivery vehicles, taxis, and public buses of courier companies and postal collection and delivery companies. The captured image is transmitted from the image capturing unit of the attached movement information collecting apparatus to the map server connected to the communication network together with the posture information and position information of the image capturing unit. In the map server, for example, the captured image is narrowed down under a predetermined condition, a building or the like serving as a common landmark is extracted from a plurality of images after the narrowing down, and the image is analyzed for the building or the like. Detect the position and shape of the building. Further, based on the shape detection, a 3D image is formed by computer graphics (CG), and the 3D image is arranged as a landmark on the map surface. Thereby, the newest map containing the visual landmark by 3D image can be created.

  In addition, according to the present system, information on the azimuth angle, roll angle, and pitch angle of each camera measured by a triaxial geomagnetic sensor mounted on the camera, and position information of the camera measured by a GPS receiver or the like. The position of the object can be specified based on a plurality of images using, for example, the principle of triangulation, and the captured image of each camera and the azimuth, roll angle, and pitch of each camera The shape of an object can be specified based on corner information, and further, the shape of a building or the like can be converted into a 3D image using computer graphics (CG). Can be automatically created based on image information sent from the camera.

(Second Embodiment)
In the map creation system of FIG. 1, in addition to a new map creation process, an existing map update process can also be performed. The map update unit 31 performs the map update process.

  As shown in FIG. 2, the movement information collection devices 16a and 16b of the moving body are provided with a time measuring unit 51 so that the time point (image capturing time information) when the image is captured can be detected. . The movement information collection devices 16a and 16b add the imaging time information to the captured image and wirelessly transmit it.

  The map correction unit 31 in the map server 20 performs the following processing in order to update the created map (stored in the map database 30 in FIG. 1).

  That is, a landmark consisting of a large number of 3D images stored in the landmark information database 27 is searched, and the imaging time information is referred to, from the creation time of the existing map to a predetermined time after the creation time (for example, the current time). ) Extract only the landmarks created so far.

  Next, the extracted landmark and the existing map stored in the map database 30 are compared and compared to determine whether the landmark is also present on the existing map. It is determined whether or not there are landmarks that are no longer present among the subjects present on the map.

  When a landmark that does not exist on the existing map is detected, the landmark is newly arranged on the map surface. If there is a landmark that no longer exists among the existing landmarks on the existing map, the landmark is deleted from the map surface. Thereby, the existing map is updated.

  In this way, imaging date and time information indicating the imaging time point is added to the captured image, and the generated landmark is created from the creation time of the existing map prepared in advance to a predetermined time after the creation time point. When a specific landmark is extracted and the existing landmark existing in the existing map is compared with the specific landmark and there is no existing landmark corresponding to the specific landmark, the specific landmark that does not exist is added to the existing map. If there is no new specific landmark corresponding to the existing landmark, the existing landmark that does not exist is deleted from the existing map. As a result, the landmark detection result can be accurately reflected on the map, and an immediate map update can be automatically performed.

The procedure for updating the map is summarized as shown in FIG. FIG. 8 is a flowchart showing the main procedure of the map update process.
As shown in the figure, first, landmarks created from the time of creation of an existing map to a predetermined time after the creation time are extracted from a number of landmarks (S401).

  Next, the extracted landmark is compared with a landmark on the existing map to detect a landmark to be newly added and a landmark that no longer exists (S402). . Then, the detection result is reflected on the map, and the map is automatically updated (S403).

  As described above, according to the automatic landmark information creation system, imaging time point information is added to image information, and a predetermined number of times from the creation point of an existing map to a predetermined point after the creation point is selected from among many landmarks. Extract the specific landmarks whose position and shape were detected by analyzing the images taken up to the point in time, compare the extracted specific landmarks with the existing landmarks on the existing map, and create a new By detecting the landmarks that should be added as well as the landmarks that no longer exist and reflecting the detection results on the map, an immediate map update (new establishment and disposal of landmarks) is automatically performed. Can be implemented.

  Further, according to the present system, it is possible to automatically collect images of a plurality of objects in a common area without excessive effort without providing a full-time worker who is responsible for taking images.

  As mentioned above, although this invention was demonstrated based on each embodiment, this invention is not limited to this, A various deformation | transformation is possible. For example, the imaging unit may be installed not only outside the moving body but also inside. Further, in each of the above embodiments, the collected map creation information is transmitted to the map server by wireless communication, but without using a dedicated wireless communication device, such as a general mobile phone or a portable terminal device. It may be transmitted via a device connectable to the Internet. Further, the present invention is not limited to this. For example, the mobile information collection device temporarily accumulates the collected map creation information in an information recording medium such as a memory, a hard disk, or an optical recording medium, and brings the information recording medium to the map server side. It is good also as a system which inputs via this input device. In this case, since the wireless communication means becomes unnecessary, the cost of the infrastructure can be reduced. The system can be realized more easily.

It is a schematic block diagram which shows an example of the automatic landmark information creation system of this invention. It is a block diagram which shows the internal structure of the movement information collection apparatus of FIG. It is a block diagram which shows the internal structure of the map server of FIG. (A) is a diagram for explaining a method of specifying the position of an object from the positions of the imaging units of the two movement information collection devices and the attitude angles (azimuth angle, roll angle, pitch angle) of each imaging unit. FIG. 2B is a diagram for explaining a method for specifying the shape of an object from images captured by each imaging unit and the attitude angles (azimuth angle, roll angle, pitch angle) of each imaging unit. It is a flowchart which shows the process sequence for the position detection of the target object used as a landmark. It is a flowchart which shows the process sequence for shape detection and 3D imaging of a landmark (object). It is a flowchart which shows an example of a series of procedures from the imaging by an imaging part to arrangement | positioning on the map surface of the landmark which consists of 3D images. It is a flowchart which shows the main procedures of the update process of a map.

Explanation of symbols

10 Building (object)
16 (16a, 16b) Movement information collection device 18 Relay station 19 Communication network 20 Map creation device (map server)
DESCRIPTION OF SYMBOLS 23 Communication front end 21 Communication part 22 Image storage part 23 Condition setting part 24 Position detection part 26 Shape detection and 3D image creation part 27 Landmark information database 28 Landmark arrangement | positioning part 29 Output part 30 Map database 31 Map correction part 35 ROM
38 Landmark creation / placement program 39 Printing unit 40 Transmission / reception unit 42 Data processing unit 46 Image processing unit 48 Control unit 51 Timekeeping unit 52 Triaxial geomagnetic sensor 54 Attitude angle calculation unit 60 Communication front end 61 Data processing unit 63 Control unit 64 RAM
AN wireless antenna

Claims (7)

An automatic landmark information creation method for generating a landmark composed of a stereoscopic image using captured images obtained by capturing an object from a plurality of directions, and arranging the generated landmark at a predetermined position on a map,
A shape setting step that uses a plurality of captured images of the object imaged in a specific direction from a specific location, and specifies the external shape of the object from the plurality of captured images;
A landmark generation step for generating the landmark from the appearance shape obtained in the shape setting step;
A position setting step of detecting the position of the object from the plurality of captured images;
A landmark placement step of placing the generated landmark at the position of the detected object on the map;
An automatic landmark information creating method characterized by comprising: An automatic landmark information creation method according to claim 1,
The shape setting step uses the imaging position on the map, the azimuth angle representing the imaging direction, the roll angle, and the pitch angle for each of a plurality of captured images obtained by imaging the object, and is commonly captured in the captured image. An automatic landmark information generating method, comprising: detecting a three-dimensional shape of the target object. An automatic landmark information creating method according to claim 1 or claim 2,
The position setting step uses the imaging position on the map, the azimuth angle representing the imaging direction, the roll angle, and the pitch angle for each of a plurality of captured images obtained by imaging the object, and is commonly captured in the captured image. An automatic landmark information creating method, comprising: detecting a position of the target object. An automatic landmark information creation method according to any one of claims 1 to 3,
The automatic landmark information creation method, wherein the captured images are a plurality of captured images obtained from each imaging unit attached to each of a plurality of moving objects. An automatic landmark information creation method according to any one of claims 1 to 4,
Add imaging date and time information indicating the imaging time to the captured image,
Among the generated landmarks, extract a specific landmark created between the creation time of an existing map prepared in advance and a predetermined time after the creation time,
If the existing landmark corresponding to the specific landmark does not exist by comparing the existing landmark existing on the existing map with the specific landmark, the specific landmark that does not exist is newly arranged on the existing map. An automatic landmark information creating method, wherein, when the specific landmark corresponding to the existing landmark does not exist, the existing landmark that does not exist is deleted from the existing map. An automatic landmark information creation system that arranges the landmark at a predetermined position based on the automatic landmark information creation method according to any one of claims 1 to 4,
A position detector for detecting the position of the object from the captured image;
A shape detection unit for detecting the shape of the object from the captured image;
A landmark generating unit that generates a landmark from the detected shape of the object;
A landmark placement unit for placing the landmark obtained by the landmark generation unit at a position detected by the position detection unit;
Automatic landmark information creation system characterized by comprising The automatic landmark information creation system according to claim 6,
An automatic landmark information creation system comprising a map correction unit that updates the existing map based on the automatic landmark information creation method according to claim 5.

Images