KR102441244B1 - Apparatus for generating precise map and method thereof - Google Patents

Abstract

The present invention relates to an apparatus and method for constructing a precise map, comprising: a camera unit for capturing an image in front of a vehicle; a GPS receiver for receiving vehicle location information using the GPS module; and by applying the previously performed learning or image recognition algorithm, extracting road characteristic data (RFD) from the image data captured by the camera unit, and simplifying the extracted road characteristic data (RFD) through a sessionization operation, After removing the remaining areas except for the road characteristic data (RFD), the control unit generates a road characteristic data (RFD) packet encoded only with information corresponding to the road characteristic data (RFD) and transmits it to a designated precision map construction server; include

Description

Precision map construction apparatus and method

The present invention relates to an apparatus and method for constructing a precise map, and more particularly, after extracting road feature data (RFD) from an image photographed using a vehicle camera, it is generated into a small-sized packet and transmitted to a server By doing so, it relates to an apparatus and method for constructing a precise map, allowing a server to build a precise map using the same.

Recently, in preparation for the commercialization of autonomous vehicles, a vehicle equipped with a Mobile Mapping System (MMS) equipment for producing precision maps with an error of 20 cm or less (ie, MMS vehicle) has been introduced and is being put into production of precision maps. .

However, in the existing precision map construction method using a mobile mapping system (MMS) vehicle, many vehicles cannot be put in because expensive equipment must be used. And there is a problem in that real-time performance is poor because a lot of cost must be invested.

Accordingly, a method of collecting road feature data (RFD) from as many vehicles as possible traveling on the road and transmitting it to a server in real time, so that the server builds a map, is being studied.

However, since the existing method transmits the image captured using the camera mounted on the vehicle to the server as it is and the server processes the image data, the efficiency is reduced due to the large amount of data and communication billing, and the server also has many vehicles. Since it is necessary to construct a precise map by analyzing all the video data transmitted from the .

Therefore, there is a need for a method for solving the problems of the existing precision map construction method.

The background technology of the present invention is disclosed in Korean Patent Application Laid-Open No. 10-2006-0003546 (published on Jan. 11, 2006, precision map data acquisition and movement device using a satellite navigation correction system).

According to one aspect of the present invention, the present invention was created to solve the above problems, and after extracting road feature data (RFD) from an image photographed using a vehicle camera, it is combined with a small packet size. It is an object of the present invention to provide an apparatus and method for constructing a precise map, which is generated as a .

A precision map construction apparatus according to an aspect of the present invention includes: a camera unit for photographing an image in front of a vehicle; a GPS receiver for receiving vehicle location information using the GPS module; and by applying the previously performed learning or image recognition algorithm, extracting road characteristic data (RFD) from the image data captured by the camera unit, and simplifying the extracted road characteristic data (RFD) through a sessionization operation, After removing the remaining areas except for the road characteristic data (RFD), the control unit generates a road characteristic data (RFD) packet encoded only with information corresponding to the road characteristic data (RFD) and transmits it to a designated precision map construction server; characterized by including.

The present invention may further include a communication unit for transmitting the road characteristic data (RFD) packet to the specified precision map construction server in a specified communication method under the control of the control unit.

In the present invention, the control unit processes the image data captured by the camera unit to measure the distance to the road characteristic data in front, or the front road using sensing information measured using at least one distance measuring sensor. It is characterized in that at least one of measuring the distance to the feature data is used.

In the present invention, the road characteristic data (RFD) includes road facility information including lanes, traffic lights, traffic signs, and milestones including white solid lines, yellow solid lines, and white dotted lines, and road surface marking information drawn on the floor surface of the road. , characterized in that it comprises at least one or more of.

In the present invention, the controller simplifies the extracted road characteristic data (RFD) through a sessionization operation to generate the road characteristic data (RFD) packet, and excluding the road characteristic data (RFD) After removing the area, it is converted into a top view image, and data corresponding to the ID assigned to the road feature data (RFD) in the top view image, and data corresponding to the part to which the ID is not assigned (0) It is characterized by coding.

In the present invention, in order to generate the road characteristic data (RFD) packet, the controller sequentially reads the coded data in a designated scan direction to generate a road characteristic data (RFD) packet in a preset form, It is characterized in that the RFD ID is first written while scanning in the scan direction, and the number of corresponding data of the same ID as the RFD ID is counted and recorded until another ID is then encoded. do it with

In the present invention, when a road characteristic data (RFD) packet is generated through the encoding in order to generate the road characteristic data (RFD) packet, the controller includes additional information including road surface marking information therein; It is characterized in that the final road characteristic data (RFD) packet is further added to the GPS coordinate information of the point in time at which the camera image is acquired.

According to another aspect of the present invention, there is provided a method of constructing a precise map, the method comprising: receiving, by a control unit, an image of a front camera photographed through a camera unit; recognizing road feature data (RFD) from the camera image by applying a learning or image recognition algorithm previously performed by the controller; After the controller simplifies the extracted road characteristic data (RFD) through a sessionization operation, removes the remaining areas except for the road characteristic data (RFD), and encodes only the information corresponding to the road characteristic data (RFD) generating a road characteristic data (RFD) packet; and transmitting, by the controller, the road characteristic data (RFD) packet to a designated precision map construction server.

In the present invention, after recognizing the road characteristic data (RFD) from the camera image, the controller allocates a preset unique ID (ID) to the recognized road characteristic data (RFD) do.

In the present invention, in order to generate the road characteristic data (RFD) packet, the controller simplifies the extracted road characteristic data (RFD) through a sessionization operation, and excluding the road characteristic data (RFD) After removing the area, it is converted into a top view image, and data corresponding to the ID assigned to the road feature data (RFD) in the top view image, and data corresponding to the part to which the ID is not assigned (0) It is characterized by coding.

In the present invention, in order to generate the road characteristic data (RFD) packet, the controller sequentially reads the coded data in a specified scan direction to generate a road characteristic data (RFD) packet in a preset form, wherein the It is characterized in that the RFD ID is first written while scanning in the scan direction, and the number of corresponding data of the same ID as the RFD ID is counted and recorded until another ID is then encoded. do it with

In the present invention, in order to generate the road characteristic data (RFD) packet, when a road characteristic data (RFD) packet is generated through the encoding, additional information including road surface marking information therein; It is characterized in that the final road characteristic data (RFD) packet is further added to the GPS coordinate information of the point in time at which the camera image is acquired.

In the present invention, the road characteristic data (RFD) includes road facility information including lanes, traffic lights, traffic signs, and milestones including white solid lines, yellow solid lines, and white dotted lines, and road surface marking information drawn on the floor surface of the road. , characterized in that it comprises at least one or more of.

A precision map construction apparatus according to another aspect of the present invention applies a previously performed learning or image recognition algorithm to extract road characteristic data (RFD) from image data captured by a camera unit, and the road characteristic data (RFD) a vehicle terminal including a control unit for generating a road characteristic data (RFD) packet encoded only with information corresponding to , and transmitting it to a precision map construction server; and a server that receives the road characteristic data (RFD) packet and generates a detailed map reflecting the road characteristic data (RFD) included in the packet.

In the present invention, the vehicle terminal simplifies the extracted road characteristic data (RFD) through a sessionization operation, removes the remaining areas except for the road characteristic data (RFD), and displays the top view image , and data corresponding to the ID assigned to the road feature data (RFD) in the top-view image, and data (0) corresponding to the part to which the ID is not assigned is characterized in that the coding is performed.

In the present invention, the vehicle terminal sequentially reads the coded data in a specified scan direction to generate a road characteristic data (RFD) packet in a preset form, and scans in the scan direction to first obtain an RFD ID. It is characterized in that the road characteristic data (RFD) packet is generated by encoding the number of data corresponding to the same ID as the RFD ID (ID) and counting and writing until another ID (ID) appears. .

In the present invention, when a road characteristic data (RFD) packet is generated through the encoding, the vehicle terminal includes additional information including road marking information and GPS coordinate information at the time the camera image is acquired. It is characterized in that the further added final road characteristic data (RFD) packet is generated.

According to one aspect of the present invention, the present invention extracts road feature data (RFD) from an image captured by a vehicle camera, and then generates it as a small packet and transmits it to the server, so that the server uses it Allows you to quickly build precise maps.

1 is an exemplary diagram showing a schematic configuration of an apparatus for constructing a precise map according to an embodiment of the present invention;
2 is a flowchart illustrating a method of constructing a precise map according to an embodiment of the present invention.
3 is a table exemplarily showing IDs to be assigned to road characteristic data (RFD) in FIG. 2 .
FIG. 4 is an exemplary view for explaining the assignment and coding process of IDs for road feature data (RFD) by a controller processing a camera image in FIG. 2 .
5 is an exemplary view illustrating a road characteristic data (RFD) packet generated in a predetermined format by scanning the coded image data in FIG. 4 .
FIG. 6 is an exemplary view illustrating a final RFD packet generated by including additional information and location information at a time point at which a camera image is acquired in the RFD packet of FIG. 5;

Hereinafter, an embodiment of an apparatus and method for constructing a precise map according to the present invention will be described with reference to the accompanying drawings.

In this process, the thickness of the lines or the size of the components shown in the drawings may be exaggerated for clarity and convenience of explanation. In addition, the terms to be described later are terms defined in consideration of functions in the present invention, which may vary according to intentions or customs of users and operators. Therefore, definitions of these terms should be made based on the content throughout this specification.

1 is an exemplary diagram illustrating a schematic configuration of an apparatus for constructing a precise map according to an embodiment of the present invention.

1 , the precision map construction apparatus according to the present embodiment includes a camera unit 110 , a GPS receiving unit 120 , a distance detection unit 130 , a control unit 140 , and a communication unit 150 . .

The precision map construction apparatus according to the present embodiment is a server that transmits road feature data (RFD) through the communication unit 150 and automatically builds a precision map using the road feature data (RFD). 200 (ie, a precision map building server).

At this time, it should be noted that the precision map building apparatus excluding the server 200 may be understood as a vehicle terminal, and the entirety including the server 200 may be understood as a precision map building apparatus (or precision map building system).

As described above, this embodiment processes the image captured by the camera unit 110 to generate small-sized data including only road characteristic data (RFD) necessary for building a precise map and transmit it to the server 200, It has a feature that can improve real-time and precision map construction work efficiency.

Referring to FIG. 1 , the camera unit 110 captures an image in front of a vehicle.

The camera unit 110 is a camera using a digital image sensor (eg, CCD, CMOS, etc.), a dedicated camera for a device for building a precise map, and a mobile terminal (eg, a smartphone, a smart pad, a navigation system, a black box) etc.) may include at least one of the cameras mounted on the camera.

The GPS receiver 110 receives current location information of the vehicle (or the precision map construction apparatus according to the present embodiment) using a Global Positioning System (GPS) module (not shown).

The distance detector 130 includes at least one sensor (eg, an ultrasonic distance sensor, an infrared distance sensor) for measuring a distance from the vehicle (or a precision map building device) to a target in front (eg, road characteristic data (RFD)) , laser distance sensors, etc.).

However, the distance from the vehicle (or the precision map construction device) to the front target (eg, road characteristic data (RFD)) may be measured by processing the image data captured by the camera unit 110 .

Accordingly, the control unit 140 processes the image data captured by the camera unit 110 and calculates the distance to the target (eg, road characteristic data (RFD)), and the target ( For example, any one of the distances to road characteristic data (RFD)) may be selected and used, or the two distance information may be combined and used to detect a more accurate distance.

In addition, the control unit 140 processes the image data captured by the camera unit 110 to extract road feature data RFD.

In this case, the controller 140 may perform deep learning in advance to extract road feature data (RFD) from the image data. Alternatively, road feature data (RFD) may be extracted by applying a known image recognition algorithm.

For reference, in this embodiment, the road characteristic data (RFD) includes information on lanes (eg, solid white line, solid yellow line, dotted white line, etc.) and road facilities (eg, traffic lights, traffic signs, milestones, etc.) information. In addition, although not specifically described in this embodiment, road surface marking information drawn on the floor surface of the road (hereinafter, described as metadata or additional information in this embodiment) is included.

Accordingly, the controller 140 may process the image data captured by the camera unit 110 to extract the pre-learned road characteristic data RFD.

More specifically, the control unit 140 converts the extracted road characteristic data (RFD) through a sessionization operation (that is, an operation of simply converting road characteristic data having a specific thickness and width into the form of lines or dots). A server designated through the communication unit 150 by generating a road characteristic data (RFD) packet filled only with information corresponding to the road characteristic data (RFD) by removing the remaining areas except for the road characteristic data (RFD). 200 (ie, the server that builds the precision map).

Accordingly, the size of the data packet transmitted to the server 200 is reduced by several tens of times or more (eg, 480 Byte → 18 Byte reduction) compared to the case of directly transmitting image data, thereby reducing data communication cost and improving efficiency. can be obtained, and it is possible to obtain the effect of improving real-time performance (real-time for the time of reflecting road feature data extracted from image data on the precise map) by reducing the load on the server 200 that needs to build a precise map. .

The communication unit 150 transmits the road characteristic data (RFD) packet generated by the control unit 140 to a predetermined server 200 .

Hereinafter, an operation of the control unit 140 and a method of constructing a precise map through linkage with the server 200 will be described in more detail with reference to the flowchart of FIG. 2 .

2 is a flowchart illustrating a method of constructing a precise map according to an embodiment of the present invention.

As shown in FIG. 2 , the controller 140 receives a front camera image (or image data) captured through the camera unit 110 ( S101 ).

Accordingly, the controller 140 processes the camera image and recognizes road characteristic data (RFD) (eg, lanes, road markings, traffic signs, traffic lights, etc.) necessary for building a precise map (S102).

In this case, the controller 140 may perform deep learning on the road feature data RFD in advance to recognize the road feature data RFD from the camera image.

In addition, the controller 140 allocates (provides) a preset unique ID to the recognized road characteristic data (RFD) (S103) (refer to FIGS. 3 and 4 (a)).

3 is a table exemplarily showing IDs to be assigned to road characteristic data (RFD) in FIG. It is an exemplary diagram to explain the assignment of ID and coding process for RFD).

As shown in FIG. 3 , for road characteristic data (RFD), for example, a solid white line among lanes is assigned as 1, a solid yellow line is assigned as 2, and a dotted white line is assigned as 3, and among road facilities, a traffic light is 4, a traffic sign is 5, The milestone is an example diagram showing a table in which 6 is assigned as an ID. However, the ID assigned to the road characteristic data RFD is set as an example. Of course, note that in another embodiment, different IDs may be assigned to each road characteristic data RFD.

In addition, in order to reduce the data size as much as possible while maintaining the geometry information for the road characteristic data (RFD) required for the construction of a precise map, the control unit 140, as shown in (a) of FIG. The camera image (or image data) is converted into a top view image as shown in FIG. 4 (b), and the image is simplified through a binarization operation and a sessionization operation based on the center line (S104) , all surrounding objects except for the road characteristic data RFD to which the ID is assigned are removed (S105).

And as shown in (c) of FIG. 4, the part to which the ID is assigned (ie, the part corresponding to road feature data) in the top view image is coded with the corresponding ID (eg, 1 to 6), and the ID is An unassigned portion (ie, a portion corresponding to the road floor area remaining after all surrounding objects have been removed) is coded as 0.

That is, through the sessionization operation, as shown in FIG. 4C , the road characteristic data (RFD) area in the camera image (or image data) is filled with data with a preset unique ID. , and the other areas are filled with 0 data.

When the coding operation is completed in this way, the camera image (or image data) is simplified into an area having an ID value of road feature data (RFD) and an area filled with zeros.

For reference, in order to convert the top view image, the controller 140 converts the recognized lanes to be parallel, and applies the conversion values applied to convert the lanes in parallel to the road characteristic data RFD, respectively. is available. However, the present invention is not limited thereto, and other known methods may be applied as the method for the top view transformation.

At this time, the camera image (or image data) has a predetermined size (eg, 8 * 20), and as shown in FIG. Generate road feature data (RFD) packets (see FIG. 5). That is, in the coded data as shown in (c) of FIG. 4 , road characteristic data (RFD) such as lanes have the same value (eg, corresponding to the same ID) in the vertical direction (ie, the designated scan direction). value), and objects such as milestones can be displayed as a single point by obtaining the center coordinates of the object.

If the coded data is encoded in the form of ID and size while scanning the coded data in the vertical direction by utilizing the characteristics of the road characteristic data (RFD), that is, when the RFD ID is first written and then another ID is displayed. If the same number of data (i.e., Pixel Length) is encoded in a manner of counting and writing up to (S106), as shown in FIG. .

FIG. 5 is an exemplary view showing a road characteristic data (RFD) packet generated in a predetermined format by scanning the coded image data in FIG. 4 , and the image data coded as shown in FIG. 4C When encoding , there are 24 consecutive 0s, 8 consecutive 1s, 24 consecutive 0s, 8 consecutive 3s, 24 consecutive 0s, 8 consecutive 0s, and 18 consecutive 0s. It is possible to generate an RFD packet indicating that consecutive 6s are continuous 1, and 0 are 45 consecutive 0s.

When it is assumed that image data of a specific size (8 * 20) is transmitted as described above, 480 bytes are transmitted in the existing data transmission method of transmitting the image itself as it is, but the method according to this embodiment cannot be applied. In this case, as shown in FIG. 5 , transmission is possible in a minimum form of 18 bytes.

In addition, since the server 200 can know the shape of the road and the existence of objects without performing separate image recognition or analysis to make a precise map from the transmitted road characteristic data (RFD), the image data It will be possible to build a precise map faster and more efficiently than the existing process that requires processing and analysis.

Next, as shown in FIG. 5 , when a road characteristic data (RFD) packet is generated, as shown in FIG. 6, metadata (ie, additional information such as road marking information) and the image data A final road characteristic data (RFD) packet is generated by further adding GPS position data as a reference ( S107 ). That is, the final RFD packet including location information such as longitude and latitude, time, and heading is completed from the RFD packet and the GPS data at the point in time at which the camera image (or image data) is finally acquired.

Next, the controller 140 transmits the generated final road characteristic data (RFD) packet to the server 200 (S108), an existing process in which the server 200 processes and analyzes the image data. Enables you to build precision maps faster and more efficiently.

In this case, the processing order of each of the steps S101 to S108 may be adjusted or changed to improve convenience and efficiency according to embodiments.

As described above, this embodiment is more effective as the size of the input image increases, and when the lane is curved, more data allocation is required, but the existing method (that is, the image data is transmitted to the server as it is, By sending a smaller amount of data to the server compared to the method of constructing a precise map by extracting road feature data through image processing and analysis), it is possible to reduce the cost burden of image data transmission and improve processing efficiency in the server. has the effect of making it

As described above, the present invention has been described with reference to the embodiment shown in the drawings, but this is merely an example, and various modifications and equivalent other embodiments are possible therefrom by those skilled in the art. will understand the point. Therefore, the technical protection scope of the present invention should be defined by the following claims.

110: camera unit 120: GPS receiver
130: distance detection unit 140: control unit
150: communication unit 200: server

Claims (17)

a camera unit for capturing an image in front of the vehicle;
a GPS receiver for receiving vehicle location information using the GPS module; and
By applying the previously performed learning or image recognition algorithm, the road characteristic data (RFD) is extracted from the image data captured by the camera unit, the extracted road characteristic data (RFD) is simplified through a sessionization operation, and the After removing the remaining areas except for the road characteristic data (RFD), a control unit that generates a road characteristic data (RFD) packet encoded only with information corresponding to the road characteristic data (RFD) and transmits it to a designated precision map construction server; includes; A precision map construction device, characterized in that.
The method of claim 1,
The precision map construction apparatus further comprising a; a communication unit for transmitting the road characteristic data (RFD) packet to the specified precision map construction server in a specified communication method under the control of the control unit.
According to claim 1, wherein the control unit,
Measuring the distance to the road characteristic data in front by processing the image data captured by the camera unit, or measuring the distance to the road characteristic data in the front using sensing information measured using at least one distance measuring sensor Precision map construction apparatus, characterized in that using at least one of.
According to claim 1, wherein the road characteristic data (RFD),
lanes including solid white, solid yellow, and dashed white lines;
road facility information, including traffic lights, traffic signs, and milestones; and
A precision map construction apparatus comprising at least one of the road marking information drawn on the floor surface of the road.
According to claim 1, wherein the control unit,
To generate the road characteristic data (RFD) packet,
After simplifying the extracted road characteristic data (RFD) through a sessionization operation, removing the remaining areas except for the road characteristic data (RFD), converting it into a TOP View image, and performing road features in the top view image An apparatus for constructing a precise map, characterized in that the data corresponding to the ID assigned to the data (RFD) is coded as data (0) corresponding to the part to which the ID is not assigned.
According to claim 5, wherein the control unit,
To generate the road characteristic data (RFD) packet,
A road characteristic data (RFD) packet of a preset form is generated by sequentially reading the coded data in a specified scan direction,
While scanning in the scan direction, the RFD ID (ID) is first written, and then, until another ID (ID) appears, the number of corresponding data of the same ID as the RFD ID (ID) is counted and recorded. Features a precision map building device.
The method of claim 6, wherein the control unit,
To generate the road characteristic data (RFD) packet,
When a road characteristic data (RFD) packet is generated through the encoding, final road characteristic data (RFD) to which additional information including road surface marking information and GPS coordinate information at the time the camera image is acquired are further added A precision map building device, characterized in that it generates a packet.
receiving, by the controller, an image of the front camera taken through the camera unit;
recognizing road feature data (RFD) from the camera image by applying a learning or image recognition algorithm previously performed by the controller;
After the controller simplifies the road characteristic data (RFD) through a sessionization operation and removes the remaining areas except for the road characteristic data (RFD), road characteristics encoded only with information corresponding to the road characteristic data (RFD) generating data (RFD) packets; and
and transmitting, by the control unit, the road characteristic data (RFD) packet to a designated precision map construction server.
9. The method of claim 8,
After recognizing road feature data (RFD) from the camera image,
The method for constructing a precise map, characterized in that the controller allocates a preset unique ID to the recognized road characteristic data (RFD).
9. The method of claim 8, wherein to generate the road characteristic data (RFD) packet,
The control unit is
After simplifying the road characteristic data (RFD) through a sessionization operation and removing the remaining areas except for the road characteristic data (RFD),
It is converted into a top view image, and the data corresponding to the ID assigned to the road characteristic data (RFD) in the top view image and the data corresponding to the part to which the ID is not assigned are encoded as data (0). How to build a precision map.
11. The method of claim 10, wherein to generate the road characteristic data (RFD) packet,
The control unit is
A road characteristic data (RFD) packet of a preset form is generated by sequentially reading the coded data in a specified scan direction,
While scanning in the scan direction, the RFD ID (ID) is first written, and then, until another ID (ID) appears, the number of corresponding data of the same ID as the RFD ID (ID) is counted and recorded. A method of constructing a precision map characterized.
12. The method of claim 11, wherein to generate the road characteristic data (RFD) packet,
The control unit is
When a road characteristic data (RFD) packet is generated through the encoding, final road characteristic data (RFD) to which additional information including road surface marking information and GPS coordinate information at the time the camera image is acquired are further added A method of constructing a precision map, characterized in that it generates a packet.
The method of claim 11, wherein the road characteristic data (RFD),
lanes including solid white, solid yellow, and dashed white lines;
road facility information, including traffic lights, traffic signs, and milestones; and
A method of constructing a precise map comprising at least one of the road marking information drawn on the floor surface of the road.
By applying a previously performed learning or image recognition algorithm, road characteristic data (RFD) is extracted from image data captured by the camera unit, and road characteristic data (RFD) encoded only with information corresponding to the road characteristic data (RFD) a vehicle terminal including a control unit for generating a packet and transmitting it to a precision map construction server; and
A server that receives the road characteristic data (RFD) packet and generates a detailed map reflecting the road characteristic data (RFD) included in the packet;
The vehicle terminal,
The image data is processed and converted into a TOP View image, and the road characteristic data (RFD) in the top view image is converted to data corresponding to a part to which an ID is assigned and data corresponding to a part to which an ID is not assigned (0) A precision map construction device, characterized in that the coding is divided into .
15. The method of claim 14, wherein the vehicle terminal,
The coding is performed on the road characteristic data (RFD) in a top-view image from which the area except for the simplified road characteristic data (RFD) is removed by simplifying the road characteristic data (RFD) through a sessionization operation. A precision map building device with
15. The method of claim 14, wherein the vehicle terminal,
A road characteristic data (RFD) packet of a preset form is generated by sequentially reading the coded data in a specified scan direction,
The RFD ID is first written while scanning in the scan direction, and then the number of corresponding data of the same ID as the RFD ID is counted and recorded until another ID appears. A precision map construction apparatus, characterized in that the data is encoded to generate a road characteristic data (RFD) packet.
The method of claim 16, wherein the vehicle terminal,
When a road characteristic data (RFD) packet is generated through the encoding, final road characteristic data (RFD) to which additional information including road surface marking information and GPS coordinate information at the time the camera image is acquired are further added A precision map building device, characterized in that it generates a packet.

Images