JP7262795B2 - Information processing device, information processing method, program - Google Patents

Description

The present invention relates to information processing technology for estimating facility names.

2. Description of the Related Art There is a social demand for collecting information on the movement of people who move on foot or using vehicles such as automobiles, analyzing the accumulated information, and using it to provide various services. For example, Patent Literature 1 discloses a location management system that identifies a user's staying point and action point based on position information acquired from a mobile terminal.

JP 2013-235323 A

By the way, there are large-scale facilities, such as shopping malls, that include a plurality of stores inside the facilities. However, when a person who moves on foot or by car visits this large-scale facility, the collected information may not be recorded as a visit to the large-scale facility. This is because the visit destination is associated with nearby shops based on latitude/longitude information acquired by GPS or the like.
In this case, the moving person may, for example, be mistakenly recorded as visiting an individual store encompassed by the large-scale facility, or a store outside the large-scale facility in the vicinity of where the person was parked. was difficult to grasp.

According to one aspect of the present invention, there is provided an information processing apparatus comprising: selection means for selecting a position information group based on position information; estimating means for estimating a facility name corresponding to the location information based on the location information.

According to the information processing apparatus of the present invention, it is possible to obtain an effect that the facility name can be estimated more appropriately.

FIG. 2 is a block diagram showing an example of the functional configuration of an information processing device; The figure which shows an example of node information. 4 is a flowchart showing an example of the flow of information processing; FIG. 7 is a diagram showing an example of a method of calculating an object surrounding area and node information of a reference target node group selected from the object surrounding area; FIG. 10 is a diagram showing a specific example of creating a histogram from reference target node group target object name information; The figure which shows an example of the parking area determination to a large-scale facility. FIG. 2 is a block diagram showing an example of a functional configuration of a server-client system; FIG. FIG. 2 is a block diagram showing an example of a functional configuration of a processing unit of a server; FIG. 4 is a block diagram showing an example of information stored in a storage unit of a server; FIG. The figure which shows the specific example about a modification. The figure which shows the specific example about a modification. The figure which shows the specific example about a modification.

Hereinafter, an example of a form for carrying out the present invention will be described with reference to the drawings.
In addition, in the description of the drawings, the same elements may be denoted by the same reference numerals, and redundant description may be omitted.
Also, the components described in these embodiments are merely examples, and are not intended to limit the scope of the present invention.

[First embodiment]
A first embodiment for realizing the information processing technology of the present invention will be described below.
The content described in the first embodiment can be applied to any of other embodiments, examples, and other modifications.

In the present embodiment, the store or facility that is the destination of a person who moves on foot or uses a vehicle such as an automobile, or the store or facility where the person who moves has stayed will be referred to as a "target".
In addition, among landmarks, large-scale facilities such as shopping malls, theme parks, and service areas that include multiple shops and facilities are called "large-scale landmarks", and large-scale landmarks are identified. Proper nouns (for example, "AA Shopping Center", "BB Outlet Mall", "CC Land", etc.) are referred to as "large-scale landmark names". Note that the large-scale target name may include a place name for identifying the location of the facility.

In this embodiment, a unit of human movement is called a "node". That is, one node includes information about the position of a person moving toward a target or a position where the moving person stayed (for example, the position of the target) and information about the name of the target. . The former is hereinafter referred to as "stay position information" and the latter is referred to as "object name".

FIG. 1 is a block diagram showing an example of a functional configuration of an information processing device 1 according to one aspect of the first embodiment.
The information processing apparatus 1 includes, for example, a reference target node selection unit 110, a node target name reading unit 120, and a large-scale target name estimation unit . These are, for example, functional units (functional blocks) possessed by a processing unit (processing device) or a control unit (control device) (not shown) of the information processing apparatus 1, and have a processor such as a CPU and an integrated circuit such as an ASIC. It consists of

The information processing device 1 also includes, for example, the node information storage unit 220 as a storage unit (buffer, storage device) (not shown) of the information processing device 1 . The node information storage unit 220 includes, for example, a volatile memory such as a DRAM, or a nonvolatile memory such as an SRAM.

Note that the node information storage unit 220 may be configured as a cache memory installed in a processor such as a CPU. That is, the node information storage unit 220 may be part of a functional unit (functional block) of a processing unit (processing device) or a control unit (control device) (not shown) of the information processing device 1 .

FIG. 2 shows an example of node information stored in the node information storage unit 220. As shown in FIG.
Information related to the movement of people is accumulated and stored in the node information. For example, a node ID, staying position information, and a target name are stored in association with each other.

A node ID is information used as an identifier for identifying each node. For example, each time a person moves to a destination or each time a person stays at a set and stored.

A target object name is a name indicating a target object (destination) of a moving person or a person who has moved. For example, based on the name of a target (destination) input to a terminal (not shown) owned by a person or a character string (for example, a search word) input to the terminal at a position specified from the stay position information Includes the name of the target (destination) displayed on the terminal. In addition, the name of the destination entered by a person traveling in a passenger car into an in-vehicle terminal (for example, a car navigation system) (not shown) mounted in the car, or the name set as the destination in the in-vehicle terminal (for example, "AA Shopping Mall Cinema Yokohama”), etc. Note that, in the target object names illustrated in FIG. 2, alphabetic characters such as "AA" actually represent proper nouns such as store names. That is, it can be said that the target object name is the name of the target object (destination) in which the moving person is interested.

Stay position information is information indicating a position in which a moving person or a person who has moved is associated with a target object name. Here, the stay position information includes, for example, position information in which the name of the target (destination) is input to a terminal (not shown) possessed by a person, and position information in which the name of the target (destination) is displayed on the terminal. Information, including the latitude and longitude for specifying the position information of the target (destination) imaged by the imaging device of the terminal. The stay position information stores, for example, the latitude and longitude acquired by the terminal using a satellite positioning system such as GPS (Global Positioning System).

In the first embodiment, it is assumed that node information is stored in advance in the node information storage unit 220 of the information processing apparatus 1 . A specific method of accumulating node information will be described later.

The reference target node selection unit 110 receives, for example, destination information indicated by latitude and longitude as an input.
For example, based on the input destination information, the reference target node selection unit 110 selects a reference target node group that is information about the node IDs of the nodes that satisfy a predetermined criterion among the nodes accumulated in the node information storage unit 220. has the function of selecting

When the node target name reading unit 120 receives the reference target node group selected by the reference target node selection unit 110 as an input, it associates it with each node ID based on the node information stored in the node information storage unit 220. It has a function of reading target object names obtained and outputting these target object names as reference target node group target object name information.

When the large-scale target name estimating unit 130 receives the reference target node group target name information from the node target name reading unit 120 as an input, it estimates the large-scale target name from the reference target node group target name information, It has a function to output the name as a large-scale target object name that is linked to the destination information.

That is, the information processing device 1 can be said to be a device that, for example, when destination information is input, estimates and outputs the name of a large-scale facility based on a reference target node group associated with the destination information.

[Information processing procedure]
FIG. 3 is a flowchart illustrating an example of image processing procedures in the first embodiment.
The processing in the flowchart of FIG. 3 is realized, for example, by the processing unit of the information processing apparatus 1 reading the code of the information processing program stored in the storage unit (not shown) into the RAM (not shown) and executing the code.

Each symbol S in the flow chart of FIG. 3 means a step.
Further, the flowcharts described below merely show an example of the information processing procedure in this embodiment, and other steps may be added or some steps may be deleted.

First, the reference target node selection unit 110 receives destination information. Then, based on the latitude and longitude of the destination information, the geographical position coordinates (geographical coordinates) of the destination are set (S101).

Note that the destination information is not limited to latitude and longitude. For example, an address (for example, “1 Nihon Odori, Naka-ku, Yokohama”) is used as destination information, and in the reference target node selection unit 110, for example, the center of gravity of the location indicated by the address is converted into latitude and longitude, and the destination You may set the position coordinates of the earth.
Alternatively, for example, a coded character string indicating a specific location in geographical coordinates, such as a map code (registered trademark), may be used as the destination information, and the position coordinates of the destination may be similarly set.

Next, the reference target node selection unit 110 calculates the surrounding area of the destination based on the position coordinates of the destination (S103).

FIG. 4 is a diagram showing an example of a method of calculating an object surrounding area and node information of a reference target node group selected from the object surrounding area.
On the left side of FIG. 4, the position specified by the positional coordinates of the destination is indicated by a black pin on the map information. Also, based on the node information stored in the node information storage unit 220, the position designated by the stay position information of each node is indicated by a white pin. A letter attached to a white pin represents the node ID of that node.

In FIG. 4, the reference target node selection unit 110, for example, selects a circular area around the destination indicated by a dotted line whose center is the position coordinates of the destination and whose radius is a predetermined Euclidean distance (for example, "500 meters"). defined as an area.

Note that the object peripheral region is not limited to the above region. For example, it may be a polygonal area having predetermined side lengths with the positional coordinates of the destination as the center of gravity. Alternatively, an area (an elliptical area in the geographic coordinate system) centered on the position coordinates of the destination and having a radius equal to a predetermined Mahalanobis distance may be set as the destination peripheral area.

Then, the reference target node selection unit 110 determines nodes existing inside the destination peripheral area (for example, “N01, . . . , N10, N12” in FIG. 4) as reference target nodes, The ID is selected as a reference target node group (S105).

Note that the method of selecting a reference target node group is not limited to the above method. For example, k nodes ("k" is a predetermined positive integer) closest to the coordinates of the destination may be selected as the reference target node group.

Also, time information at the time of acquisition of stay position information may be added to node information, and only nodes having time information within a certain period (for example, within one week) may be selected as reference target nodes.

Then, the reference target node selection unit 110 determines whether or not the number of nodes determined as the reference target node group is equal to or greater than a predetermined number (eg, 6) (S107).
If the number is less than the predetermined number (S107: NO), the reference target node selection unit 110 determines that there is no large-scale facility at the location indicated by the destination information, and terminates the process. In this case, the large-scale target name estimation unit 130 does not output the large-scale target name.

If the number exceeds the predetermined number (S107: YES), the reference target node selection unit 110 determines that a large-scale facility exists at the location indicated by the destination information, and outputs the reference target node group to the node target name reading unit. do.

If it is less than the predetermined number (S107: NO), for example, the radius of the circular area is increased by a predetermined distance (eg, "100 meters") to widen the area around the destination, and steps S105 to S107 are performed again. may be executed.

Upon receiving the reference target node group as an input, the node target name reading unit 120 refers to the node information stored in the node information storage unit 220 based on the node IDs of the nodes included in the reference target node group. Read the target object name associated with the node ID. Then, the node target name reading unit 120 outputs these target names to the large-scale target name estimating unit 130 as reference target node group target name information (S109).

An example of readout processing in step S109 is shown on the right side of FIG. On the right side of FIG. 4, the node IDs of the nodes selected as the reference target node group (eg, “N01, . Mall Cinema Yokohama”, .
These target names are output to the large-scale target name estimating unit 130 as reference target node group target name information.

Note that the node target name reading unit 120 is not provided, and after the nodes to be referenced are determined by the reference target node selection unit 110 in step S107, the target names of those nodes are read from the node information and It may be output as node group target name information.

When the large-scale target name estimating unit 130 receives the reference target node group target name information as an input, it performs morphological analysis on the character strings obtained as the respective target object names (S111). Note that if the character string has already been divided into words, for example, by spaces, word spacing processing may be performed instead of the morphological analysis.

Next, the large-scale target name estimating unit 130 creates, for example, a histogram of word appearance frequency based on the word groups obtained by the morphological analysis or the word segmentation processing (S113).

FIG. 5 is a diagram showing a specific example of creating a histogram from reference target node group target name information. The left side of FIG. 5 shows input reference target node group target name information, and the right side shows a histogram obtained from the target name.
For example, the large-scale target name estimating unit 130 morphologically analyzes the target name "AA Shopping Mall Cinema Yokohama", and obtains the results "AA", "shopping", "mall", "cinema", and "Yokohama". are stacked as elements of the histogram. By accumulating other target names as elements of the histogram in the same manner, the histogram shown on the right side of FIG. 5 is completed.
From this histogram, it can be obtained that the appearance frequency of words is high in the order of "Yokohama", "AA", "store", "shopping", .

Then, the large-scale target name estimating unit 130, for example, converts a character string in which the M ("M" is a predetermined positive integer) words with the highest appearance frequency in the histogram created in step S113 are arranged as the destination information. is estimated as a large-scale target name (name of a large-scale facility) linked to the location indicated by (S115). Then, the large-scale target name estimation unit 130 terminates the processing.

In the example of FIG. 5, for example, when M=3, "Yokohama AA Store", which is the top three words in the histogram, is estimated as the large-scale target name linked to the destination information.

Note that the name to be output is not limited to a character string in which M words are arranged in order of appearance frequency in the histogram. For example, information on the WWW (World Wide Web) may be searched using the obtained M words as search keywords, and the name may be estimated based on, for example, the page titles of the top search results.
More specifically, the large-scale target name estimating unit 130 selects a plurality of large-scale target names based on the page titles selected as the top as large-scale target name candidates as a large-scale target name determining unit (not shown). Output to Then, for example, the large-scale target name determination unit determines a large-scale target name to be output from the information processing apparatus 1 from the large-scale target name candidates by a user's selection operation for the large-scale target name candidates. can be

Alternatively, the large-scale target name included in a prepared corpus may be correlated with the word group having the highest appearance frequency in the histogram, and the large-scale target name with the highest correlation may be used as the estimation result.

By repeating the above process while shifting the latitude and longitude of the destination information, a large-scale target object name corresponding to the area corresponding to each destination information is estimated. In addition, an area on the map information corresponding to the common large-scale target name (hereinafter referred to as a "large-scale facility area") is also estimated. It can be said that it is the range of the name on the map information.

It is also possible to give destination information at an arbitrary point and estimate the large-scale target object name corresponding to that point each time.

FIG. 6 is a diagram showing an example of large-scale facility area determination.
The left side of FIG. 6 shows the same map information as the left side of FIG. 4, and the right side of FIG. 6 shows a large-scale facility area and its large-scale target name, which can be repeatedly obtained while changing the destination information. ing.
On the right side of FIG. 6, the map information is based on two sets of points of destination information obtained by clustering large-scale target names linked to individual destination information by, for example, the Ward method. Two bounded areas are shown as large facility areas. The right side of FIG. 6 shows an example in which the large-scale facility area near the center is classified as "Yokohama AA Store" and the large-scale facility area in the lower left is classified as "BB Outlet Yokohama".
From this figure, the destination or place of stay of the people in the large-scale facility area of "Yokohama AA store" near the center is "Yokohama AA store", and the person in the large-scale facility area of "BB Outlet Yokohama" in the lower left. It can be inferred that the destination or place of stay is "BB Outlet Yokohama". It can also be inferred that the destinations or places of stay of people outside these two large-scale facility areas are not large-scale facilities.

[First embodiment]
Next, a first example of a terminal or an electronic apparatus (electronic equipment) to which the information processing apparatus 1 of the first embodiment is applied or provided with the information processing apparatus 1 will be described.
Here, as an example, an embodiment of a server provided with the information processing device 1 and an in-vehicle terminal mounted in a passenger car will be described. However, it goes without saying that the embodiments to which the present invention can be applied are not limited to this embodiment.

FIG. 7 is a diagram showing an example of the system configuration of the server-client system 1000 in the first embodiment.
In the server client system 1000, for example, a server 10 and a plurality of in-vehicle terminals 40 are connected via a network 30. FIG. Note that FIG. 7 shows only one representative component of the in-vehicle terminal 40 . Also, a plurality of servers 10 may be connected.

The server 10 includes, for example, a processing unit 100, a storage unit 200, an operation unit 310, a display unit 320, a sound input unit 330, a sound output unit 340, and a communication unit 350.

The processing unit 100 is a processing device that comprehensively controls each unit of the server 10 according to various programs such as a system program stored in the storage unit 200 and performs various processes related to information processing. or the like and an integrated circuit such as an ASIC.

FIG. 8 shows an example of functional units that constitute the processing unit 100 .
The processing unit 100 has a reference target node selection unit 110, a node target name reading unit 120, and a large-scale target name estimating unit 130 as main functional units. These functional units respectively correspond to the functional units included in the information processing apparatus 1 of FIG.

The storage unit 200 is a storage device including a volatile or nonvolatile memory such as ROM, EEPROM, flash memory, RAM, etc., a hard disk device, and the like.

FIG. 9 shows an example of information stored in the storage unit 200. As shown in FIG.
The storage unit 200 stores, for example, a facility name estimation program 210 and a node information storage unit 220 .

The facility name estimation program 210 is a program that is read by the processing unit 100 and executed as large-scale facility name estimation processing. This large-scale facility name estimation process is executed, for example, as a process based on the flowchart shown in FIG.

The node information storage unit 220 corresponds to the storage unit included in the information processing device 1 in FIG. 1 or the storage unit attached to the information processing device 1, respectively.

The map information storage unit 230 stores, for example, information (map information) on terrain features within a certain area. Here, the features include those that physically exist such as roads and buildings, those that do not exist physically such as place names, administrative boundaries and contour lines, or both. Note that the map information storage unit 230 may store a satellite photograph associated with the position as the map information.

Note that the node information storage unit 220 and the map information storage unit 230 are stored in a database management device (not shown) connected via the communication unit 350 (for example, a SQL (Structured Query Language) server, etc.). may transmit node information and map information to the processing unit 100 in response to a database read command received via the network 30 .

The operation unit 310 includes input devices such as a keyboard and a mouse for the user to input various operations to the server. The operation unit 310 outputs an operation signal to the processing unit 100 according to the user's operation.
Note that the operation unit 310 may have a touch panel (not shown) integrally configured with the display unit 320 , and this touch panel may function as an input interface between the user and the server 10 .

The display unit 320 is a display device including an LCD (Liquid Crystal Display), an OELD (Organic Electro-luminescence Display), or the like, and performs various displays based on display signals output from the processing unit 100. .
Note that the display unit 320 may be configured integrally with a touch panel (not shown) to form a touch screen.

The sound input unit 330 is a sound input device including a microphone, an A/D converter, etc., and performs various sound inputs based on sound input signals input to the processing unit 100 .

The sound output unit 340 is a sound output device including a D/A converter, a speaker, etc., and performs various sound outputs based on sound output signals output from the processing unit 100 .

The communication unit 350 is a communication device for transmitting and receiving information used inside the device to and from an external information processing device via the network 30 . As a communication method of the communication unit 350, a wired connection format via a cable conforming to a predetermined communication standard such as Ethernet or USB (Universal Serial Bus), Wi-Fi (registered trademark) or 5G (5th generation mobile communication A variety of methods can be applied, such as a wireless connection format using a wireless communication technology conforming to a predetermined communication standard such as system) and a connection format using short-range wireless communication such as Bluetooth (registered trademark).

The in-vehicle terminal 40 includes, for example, a processing unit 460, a storage unit 470, an operation unit 410, a display unit 420, a sound input unit 430, a sound output unit 440, a communication unit 450, a clock unit 480, a position A calculation information detection unit 490 is provided.

The processing unit 460 is a processing device that comprehensively controls each unit of the in-vehicle terminal 40 according to various programs such as a system program stored in the storage unit 470 and performs various processes related to information processing. It is composed of a processor such as a DSP and an integrated circuit such as an ASIC.

The storage unit 470 is a storage device including volatile or nonvolatile memory such as ROM, EEPROM, flash memory, and RAM, hard disk device, and the like.

The operation unit 410, the display unit 420, the sound input unit 430, the sound output unit 440, and the communication unit 450 can be configured in the same manner as the respective functional units of the server 10, for example. omitted.

The clock unit 480 is a built-in clock of the in-vehicle terminal 40, and outputs time information (timekeeping information) obtained based on a clock using a crystal oscillator, for example. Note that the clock unit 480 may acquire time information via the communication unit 450 and the network 30 according to the NITZ (Network Identity and Time Zone) standard or the like.

The position calculation information detection unit 490 is a functional unit that detects (measures) information necessary for the processing unit 460 to calculate (measure) the position of the vehicle on which the in-vehicle terminal 40 is mounted.
The position calculation information detection unit 490 includes, for example, a satellite positioning sensor (satellite positioning unit), which is a sensor or unit for calculating the position of the vehicle using a satellite positioning system such as GPS (Global Positioning System), inertial It is composed of inertial measurement sensors (inertial measurement units (IMUs)) that are sensors and units for calculating the position of a vehicle using a navigation system.

The network 30 is, for example, a communication network that connects one or more servers 10 and one or more in-vehicle terminals 40 by wired connection, wireless connection, or both.

[Accumulation of node information]
The processing unit 100 of the server 10 executes node information accumulation processing prior to the large-scale facility name estimation processing.

At the departure point, the processing unit 460 of the vehicle-mounted terminal 40 causes the display unit 420 to display a display prompting the person (user) on the vehicle to enter the destination (object name).
When the destination is input based on the user's operation on the operation unit 410 , the processing unit 460 receives the state of the vehicle from the vehicle (not shown) through the communication unit 450 . Then, the processing unit 460 determines that the vehicle is parked when predetermined conditions (for example, the ignition switch is turned off and the parking brake is turned on) are met.

After the destination is input, the processing unit 460 continues to acquire vehicle position information (eg, latitude and longitude) based on the position calculation information calculated by the position calculation information detection unit 490 . Then, based on the time information obtained from the clock unit 480, the processing unit 460 may determine that the vehicle is parked when there is no change in the position information of the vehicle for a predetermined time (for example, 5 minutes).

When it is determined that the vehicle is parked, the processing unit 460 acquires the staying position information (for example, latitude and longitude) of the vehicle based on the position calculation information calculated by the position calculation information detection unit 490. . Then, the processing unit 460 transmits the input destination (object name) and the staying position information of the vehicle to the server 10 through the communication unit 450 .

The processing unit 100 of the server 10 generates a unique node ID when the communication unit 350 receives the target object name and the staying position information from the vehicle-mounted terminal 40 . Then, the generated node ID is associated with the received object name and stay position information, and stored in the node information storage unit 220 of the storage unit 200 .

Node information is accumulated in the node information storage unit 220 by repeating the above processing.

[Large-scale facility name estimation processing]
The processing unit 100 of the server 10 performs large-scale facility name estimation processing according to the facility name estimation program 210 stored in the storage unit 200 .

Specifically, for example, the processing unit 100 of the server 10 scans the latitude and longitude indicated by the destination information within the map area stored in the map information storage unit 230, for example, the flowchart shown in FIG. , the large-scale facility name estimation process is repeatedly executed.

As a result, a large-scale facility area is estimated within the map area stored in the map information storage unit 230 . Then, the processing unit 100 of the server 10 causes the map information storage unit 230 to store the large-scale facility area as a new map layer. Hereinafter, the generated map layer will be referred to as a "large-scale landmark name layer". Note that the large-scale target name layer is empty in areas where no large-scale target name is output.
Then, the processing unit 100 of the server 10 causes the display unit 320 to display, for example, a large-scale target object name layer superimposed on the map information.

In this way, by referring to the map information in association with the large-scale target name layer, the user of the server 10 can more accurately estimate the destination of a person traveling in a vehicle. It becomes possible to operate marketing analysis services based on information more effectively.

[Actions and effects of the first embodiment]
According to the server 10 of the first embodiment, it is possible to obtain the same functions and effects as those of the first embodiment described above.

[Modification of First Embodiment]
Embodiments to which the present invention can be applied are not limited to the first embodiment. Modifications will be described below.

<Name estimation method other than histogram>
In the first embodiment, the large-scale target name estimating unit 130 creates a histogram regarding the frequency of appearance of words from the reference target node group target name information, and estimates the large-scale target name, but is not limited to this. . For example, a group of words obtained by morphological analysis may be converted into a group of vectors and estimated by vector analysis.

In this case, for example, all target object names accumulated in the node information are morphologically analyzed, and word groups obtained as a result are learned, for example, by word2vec, which is a natural language processing model using a neural network. Then, the word group obtained as the morphological analysis result of the reference target node group object name information is converted into a vector group using word2vec. Then, each reference target node becomes a point distributed in the transformed vector space, and the reference target node group can be regarded as a probability distribution on the vector space.

Therefore, for example, eigenvalue decomposition is applied to this distribution, and the eigenvector with the maximum eigenvalue is converted into a character string using word2vec. Then, the converted character string becomes the estimated large-scale target object name.

It is also possible to let word2vec learn character strings as they are, without executing morphological analysis or word spacing processing on target object names accumulated in the node information. At this time, each object name in the reference target node group object name information can be vectorized as a character string using word2vec without executing morphological analysis or word separation processing. Large target names can be inferred as well.

<Hierarchical name estimation>
In the first embodiment, the destination information and the name of the large-scale target were described as being associated one-to-one, but the present invention is not limited to this. For example, multiple large-scale landmark names may be estimated for a given destination information.

In this case, for example, a class is defined for each word in a prepared corpus. Here, the class is a label for hierarchically identifying the facilities included in the large-scale facility area. A”, “cinema”, “shopping”, and “parking” are classified as “class B”, and items such as “camera”, “motorbike”, and “liquor” are classified as “class C”.

For example, when creating a histogram regarding the appearance frequency of words from the reference object node group target name information, the histogram is created for each of these classes. Then, for some position information, the class A large-scale target name is "AA Yokohama", the class B large-scale target name is "shopping mall", and the class C large-scale target name is "camera". Suppose that an estimation result is obtained.

At this time, the destination of the person who parks at the location indicated by the destination information is hierarchically arranged from "AA Yokohama" to "shopping mall" to "camera" for each class selected from class A to class C. And it is possible to narrow down the scale of the facility and make an estimate (estimate a more detailed destination).

In an embodiment, the large-scale target object name layer may be stored in the map information storage unit 230 as a plurality of layers for each class.

By using this hierarchical name estimation, it is expected that the usefulness of the data analysis of visited destinations based on location information will be further enhanced.

<Use of map information other than node information>
In the first embodiment, the large-scale facility area is estimated by repeatedly shifting the latitude and longitude of the destination information for the entire area of the map information, but the present invention is not limited to this. For example, if the shape of a large-scale facility is obtained in advance from map information, based on that shape, the destination information is set only for the area inside and around the large-scale facility, and the large-scale facility area and large-scale landmark name are set. may be estimated.

By limiting the search area for the destination information in this way, it becomes possible to estimate the large-scale facility area at a higher speed.

<Expansion of target name>
In the first embodiment, the name of the target is described as the name of the store or facility that is the destination, but the name is not limited to this. For example, the target name may include an address and a keyword. Alternatively, an address or a keyword may be used as the target name.

Also, for example, when an address is mixed as a target object name, and when a facility exists in the same building, there is a high possibility that a character string including the building name is estimated as the large-scale target object name. That is, it is possible to further improve the accuracy of the estimated name.

<Application to mobile terminals other than in-vehicle terminals>
In the first embodiment, the node information is acquired by the in-vehicle terminal 40, but the present invention is not limited to this. For example, it may be acquired by a portable terminal such as a smartphone or a tablet used by a person (who is a user) who moves.

In this case, for example, if a change in the position information acquired by the control unit based on the position calculation information calculated by the position calculation information detection unit of the mobile terminal falls below a predetermined range within a certain period of time, the mobile terminal may acquire the location information of the mobile terminal as stay location information and transmit it to the server 10

Alternatively, after a destination is input based on a user operation on the mobile terminal, when a predetermined user operation on the mobile terminal (for example, an imaging operation on the imaging unit) is performed, the control unit of the mobile terminal moves to the target. Determined to have moved. Then, based on the position calculation information calculated by the position calculation information detection unit, the control unit of the mobile terminal may acquire the position information as stay position information and transmit it to the server 10 .

When a person moves in a vehicle, for example, based on the information for position calculation calculated by the position calculation information detection unit, the control unit of the mobile terminal continues to acquire the position information, and the position information changes with time. Based on this, it may be determined that the vehicle has moved to the target (the vehicle has been parked) at a position where the speed has changed from the speed of moving by vehicle to the speed of moving on foot, and acquired as stay position information.

In addition, by installing the in-vehicle terminal 40 linked in advance with the mobile terminal of the user traveling by vehicle in the vehicle, after the destination is input based on the user's operation on the mobile terminal, the movement information of the vehicle (position, posture, acceleration, etc.) are acquired by the in-vehicle terminal 40 and transmitted to the server 10 . Then, the server 10 judges that the vehicle is parked based on the time change of the transmitted position or acceleration falling below a specified value, and stores the latitude and longitude of that point as staying position information. good too.
In this case, the in-vehicle terminal 40 can be a simple terminal configured by a processing section 460 , a storage section 470 , a clock section 480 , a position calculation information detection section 490 and a communication section 450 .

<Use of large-scale facility name estimation processing result by terminal>
In the first embodiment, the user of the server 10 uses the result of the large-scale facility name estimation process, but the present invention is not limited to this. For example, it may be made available to the user of the in-vehicle terminal 40 .

In this case, first, at the departure point, the processing unit 460 of the in-vehicle terminal 40 acquires map information from, for example, the map information storage unit 230 of the server 10 and causes the display unit 420 to display the map. Then, the processing unit 460 of the in-vehicle terminal 40 causes the display unit 420 to display a display prompting the person (user) who has boarded the vehicle to select a destination position from the map.

When a destination position (for example, the position of the destination pin on the map information on the left side of FIG. 4) is selected based on the user's operation on the operation unit 410, the processing unit 460 calculates the selected latitude on the map. and longitude are transmitted to the server 10 by the communication unit 450 as destination information.

In addition, instead of the destination position, the vehicle-mounted terminal 40 transmits a search word (for example, an address, a store name, etc.) input by the user to the server 10, and the server 10 transmits information stored in the map information storage unit 230. You may make it set destination information based on a search word.

When destination information is received by the communication unit 350 from the vehicle-mounted terminal 40, the processing unit 100 of the server 10 executes the large-scale facility name estimation process, for example, according to the flowchart shown in FIG. Then, a large-scale target object name (for example, "Yokohama AA store") corresponding to the destination information is obtained.

Then, the processing unit 100 of the server 10 retrieves advertisement information (for example, “Yokohama AA store 10th anniversary sale ”) is read.

When reading the advertisement information, if the information on the user of the in-vehicle terminal 40 is available, the advertisement information may be searched in association with the information of the user. For example, if there is a large-scale facility (for example, "CC Land") frequently visited by users of a certain age group, the large-scale facility is prioritized when the user of the vehicle-mounted terminal 40 falls under that age group. You may make it transmit the advertisement information of "CC land."
In addition to age, the user's information may include the area of residence, interests of the user, and the like.

Then, the processing unit 100 transmits the advertisement information to the vehicle-mounted terminal 40 through the communication unit 350 . When the communication unit 450 receives the advertisement information from the server 10, the processing unit 460 of the in-vehicle terminal 40 causes the display unit 420 to display the advertisement information.

Note that the processing unit 460 of the in-vehicle terminal 40 does not transmit the destination information to the server 10, but acquires the large-scale target name based on the large-scale target name layer of the map information, and sends the large-scale target name to the server 10. Send target name. Then, the control unit 100 of the server 10 may read the advertisement information from the received large-scale landmark name.

The use of the large-scale facility name estimation processing result by the terminal is not limited to the distribution of advertisement information. For example, the destination information is acquired from a plurality of in-vehicle terminals 40, and the number of vehicles headed for a large-scale facility is grasped. When it is expected that a certain number of people or more will be concentrated in a certain large-scale facility area, the display unit 420 of the vehicle-mounted terminal 40 is caused to display that the large-scale facility area is congested, thereby reducing congestion. It can also be applied to services that avoid

Also, the above contents may be realized by a portable terminal such as a smart phone or a tablet used by a moving person (who is a user) instead of the in-vehicle terminal 40 .

<Accumulation and estimation of node information>
In the first embodiment, the accumulation of node information and the large-scale facility name estimation process were described separately, but the present invention is not limited to this. For example, node information may be accumulated while performing large-scale facility name estimation processing.

In this case, for example, at the departure point, the processing unit 460 of the in-vehicle terminal 40 causes the display unit 420 to display a display prompting the user who has boarded the vehicle to enter the destination using a character string.

Then, the processing unit 460 of the in-vehicle terminal 40 extracts the destination from, for example, a map database (not shown) stored in the storage unit 470 based on the input character string (facility name, address, etc.) about the destination. Calculate the position (latitude and longitude) of Then, the communication unit 450 transmits the calculated latitude and longitude as destination information to the server 10 .

When destination information is received by the communication unit 350 from the vehicle-mounted terminal 40, the processing unit 100 of the server 10 executes large-scale facility name estimation processing, for example, The facility congestion status is transmitted to the in-vehicle terminal 40. - 特許庁

When the communication unit 450 receives the congestion status from the server 10, the processing unit 460 of the in-vehicle terminal 40 causes the display unit 420 to display the congestion status superimposed on the map information. Then, when the processing unit 460 of the vehicle-mounted terminal 40 determines that the vehicle is parked, based on the position calculation information calculated by the position calculation information detection unit 490, the vehicle parking position information (for example, latitude and longitude) is obtained. ), and transmits the input character string about the destination and the parking position information of the vehicle to the server 10 via the communication unit 450 .

When the communication unit 350 receives the target name and parking position information from the vehicle-mounted terminal 40, the processing unit 100 of the server 10 generates a unique node ID, and combines the node ID with the received target name and parking position information. are associated with each other and stored in the node information storage unit 220 of the storage unit 200 .

In this way, the server 10 can also provide a service of referring to location information while accumulating node information.

[Second embodiment]
In the first embodiment, the destination information is described as geographical position coordinates (geographical coordinates) indicated by latitude and longitude, but is not limited to this. The second embodiment is an embodiment in which the destination information is an area.
The content described in the second embodiment can be applied to any of other embodiments, examples, and other modifications.
Moreover, the same code|symbol is attached|subjected about the component same as an already-appearing component, and description for the second time is abbreviate|omitted.

For example, the area of the map information is divided into grids (for example, grids of 50 meters square), and one of the divided areas is used as the destination information.

Note that the method of dividing the area of the map information is not limited to the grid of regular squares. For example, it may be divided into regular hexagonal lattices. Alternatively, the division may be performed using arbitrary delimiters (for example, road information on a map) instead of the grid pattern.

FIG. 10 is a diagram showing an example of large-scale facility name estimation processing in the case where the map information shown on the left side of FIG. 4 is divided into square grid areas, and the center area is used as the destination information. The grid area (divided area) selected as the destination information is hereinafter referred to as a "selected divided area".

In this case, the reference target node selection unit 110 receives the selected divided area as the destination information. Then, the reference target node selection unit 110 selects the selected divided area as the object surrounding area. Then, the processing after the step of S105 in FIG. 3 is executed.

In the left side of FIG. 10, nodes with node IDs of "N01", "N03", "N05", "N06", "N08", and "N17" are inside the object surrounding area (=selection divided area). exists, and a predetermined number (for example, 6) or more nodes exist. Therefore, in step S107 of FIG. 3, the reference target node selection unit 110 determines that a large-scale facility exists in this selected divided area.
Then, the node target name reading unit 120 reads the target name associated with each node ID as shown in the upper right of FIG.

Then, the large-scale target name estimation unit 130 morphologically analyzes these target names and creates a histogram shown in the lower right of FIG. Finally, the large-scale target name estimation unit 130 determines, for example, "Yokohama AA", which is the top two words in the histogram, as a large-scale target name associated with this selected divided area (large-scale facility area). estimated as

By determining whether each divided area is a large-scale facility area or not, it is possible to perform a wide-area search at high speed.

Even if it is determined that there is no large-scale facility in a selected divided area, if there is a large-scale facility area around this selected divided area, large-scale facilities exist in the adjacent area to the large-scale facility area. You are likely to be judged.

In this case, in step S107 of FIG. 3, the number of nodes judged to have a large-scale facility may be reduced, and the judgment may be made again.

Alternatively, the target name of each node in the large-scale facility area may be replaced with the estimated large-scale target name, and the large-scale target name may be estimated at an arbitrary point near the large-scale facility area. .

FIG. 11 is a diagram showing an example of a method of estimating a large-scale target name at a point near a large-scale facility area.
On the left side of FIG. 11, it is assumed that, of the map information partitioned by grids, for example, the center area is determined to be a large-scale facility area, and the name of the large-scale target is "AA Yokohama".
At this time, for example, consider the case of estimating the name of a large-scale target at a neighboring point on the upper right side of the large-scale facility area indicated by a black pin.

First, when the reference target node selection unit 110 receives a neighboring point as destination information, for example, a k-NN (Nearest Neighbor) method is used to select a group of k nodes that are close to the neighboring point as a reference target node group. select. On the left side of FIG. 11, for example, nodes with node IDs "N01", "N03", "N08", "N09", and "N10" are selected as a reference target node group when "k=5". is shown.

Then, the node target name reading unit 120 refers to the node information stored in the node information storage unit 220 based on these node IDs, and reads the target name associated with each node ID. However, for the nodes with node IDs "N01", "N03", and "N08" that have already been determined to be large-scale facility areas, for example, the large-scale target name layer is referred to, and the target name is already set to The presumed "AA Yokohama" is used.
For example, in the upper right part of FIG. 11, the node ID of the node selected as the reference target node group is associated with the target object name read from the node information and replaced as necessary.

After that, the large-scale target name estimating unit 130 executes the processing after the step of S111 in FIG. For example, the lower right portion of FIG. 11 shows the created histogram. From this histogram, it can be seen that, for example, "Yokohama AA", which is the word with the highest appearance frequency, is estimated as the large-scale target name at this neighboring point.

By executing the above processing in a predetermined neighboring area, it becomes possible to estimate the name of the large-scale target in the neighboring area of the large-scale facility area.

After the large-scale facility name estimation process, if the scale of the map is reduced (the map area is expanded) and the large-scale landmark name layer is superimposed on the map information, the large-scale landmark name layer will have a large scale for each grid. Since the scale target name is included, there arises a problem that the map information displayed on the display unit 320 becomes difficult to see.
This solution is shown below.

FIG. 12 is a diagram showing an example of the result of estimating the name of a large-scale target in a region divided into nine grids in the region of the map information, for example.
The numbers and alphabets shown in <> in each grid are identifiers for identifying the grid. In this example, vertical grids are identified by "1", "2" and "3", and horizontal grids are identified by "a", "b" and "c". In the following, for example, the region within the upper left grid is referred to as the <1a> region, and the like.

Below the identifiers in the grid are shown the estimation results of the large-scale target names within each region. When it is determined that there is no large-scale target in the area, "no large-scale facility" is displayed.

Beneath the estimated large-scale target name, a matching frequency score is displayed, which expresses the ratio of the large-scale target name included in the target names selected as reference target nodes. .
Here, each match frequency score takes "0" when the match frequency is the lowest and "1" when the match frequency is the highest.
For example, in the <1b> area, if the number of nodes included in the reference target node group is "10", among the 10 target object names, if the matching frequency score is "0.6", the reference target "AA Yokohama" is included in "6" nodes obtained by multiplying the number of nodes "10" included in the node group by the matching frequency score "0.6".

In FIG. 12, for example, consider a case where the scale of the map is doubled (the area is 1/4). At this time, the large-scale target names to be displayed corresponding to the <1a> area, <1b> area, <2a> area, and <2b> area can be determined as follows, for example.
・When selecting the largest number of large-scale target names included in each grid, “AA Yokohama” is displayed because there are 2 “AA Yokohama” and 1 “BB Outlet”. When selecting the one with the highest match frequency score among the large-scale target names included in each grid, "BB outlet" is displayed because "BB outlet" is "1" and is the highest ・Included in each grid When selecting the one with the highest match frequency score total for each large-scale target name among the large-scale target names that are displayed, "AA Yokohama" is "0.6" + "0.7" = "1.3" , "AA Yokohama" is displayed. These methods may be combined, for example, to select the one with the highest matching frequency score total when the numbers are the same.

Similarly, for example, consider a case where the scale of the map is tripled (the area is 1/9). Then, the names of large-scale targets to be displayed corresponding to the <1a> to <3c> areas (that is, the entire area of the map information in FIG. 12) can be similarly selected.
・When selecting the largest number of large-scale target names included in each grid, “AA Yokohama” is displayed because there are 4 “AA Yokohama” and 3 “BB outlets”. When selecting the one with the highest match frequency score among the large-scale target names included in each grid, "BB outlet" is displayed because "BB outlet" is "1" and is the highest ・Included in each grid When selecting the one with the highest match frequency score total for each large-scale target name among the large-scale target names that are displayed, "BB outlet" is "2.7 (= 1 + 0.9 + 0.8)", "AA Yokohama " is "2.8 (= 0.6 + 0.7 + 0.7 + 0.8)" and "AA Yokohama" is the highest, so "AA Yokohama" is displayed.

It should be noted that the above method can be used not for display selection but for estimating a new large-scale target name for each grid area when the resolution of the grid in the map area is reduced.

[Third Embodiment]
In the first embodiment, the target is a store/facility that is the destination of a person who travels regardless of means of transportation.・Set the facility as the target.
The content described in the third embodiment can be applied to any of other embodiments, examples, and other modifications.
Moreover, the same code|symbol is attached|subjected about the component same as an already-appearing component, and description for the second time is abbreviate|omitted.

In the third embodiment, the movement of a person who moves using a vehicle such as an automobile or a bicycle is considered by dividing it into the following two steps.
(a) Move the vehicle from the departure point to a parking lot, bicycle parking lot, or the like near the destination, and park (stop) the vehicle in the parking lot, bicycle parking lot, or the like.
(b) Move to the destination on foot from the parking lot, bicycle parking lot, or the like.
In the following, as a vehicle, for example, the movement of a person who moves by a passenger car is considered.

In the third embodiment, the position where the passenger car is parked in (a) above is called a "parking position", and the destination store or facility in (b) above is called a "target".

In the third embodiment, a unit of movement of a person who moves using a vehicle, which is completed by the series of steps (a) and (b) above, is referred to as a "node". That is, one node includes at least information about the parking position and information about the target. The former is hereinafter referred to as "parking position information" and the latter as "object name". Note that the node may include identifier information for identifying the vehicle and information on the departure point.

In addition, the vehicle of the moving person is not limited to a passenger car. For example, for a person traveling by bus, the parking location can replace the bus stop location. Also, for a person traveling by helicopter, the parking position can be replaced with the helipad position.

In this case, the information processing device 1 executes the process based on the flowchart shown in FIG. 3 by using the parking position information as the staying position information, for example.
Then, the estimated large-scale facility area is considered to be an area (stop area) where a person heading to the facility identified by the large-scale target name parks a passenger car.

1 information processing device 10 server 30 network 40 on-vehicle terminal 110 reference target node selection unit 120 node target name reading unit 130 large-scale target name estimation unit 220 node information storage unit 1000 server client system

Claims (5)

An information processing device,
a location information acquisition unit that acquires destination location information, which is location information of the destination;
a range setting unit that sets a first range based on the destination position information;
a node information identifying unit that identifies a node information group that is a set of node information associated with the name and location information and that is a set of node information associated with the location information included in the first range;
a name group extraction unit for extracting a name group linked to each node information in the identified node information group;
a determination unit that determines the name of the large-scale facility corresponding to the destination based on the extracted name group when a first condition regarding the number of node information in the identified node information group is satisfied;
An information processing device. The information processing device according to claim 1 ,
If the first condition is not satisfied,
The range setting unit sets a second range different from the first range based on the destination position information,
the node information identifying unit identifies a node information group included in the second range;
Information processing equipment. The information processing device according to claim 1 or 2 ,
The determination unit determines the name of the large-scale facility based on at least one of word spacing and morphological analysis.
Information processing equipment. An information processing method in an information processing device ,
Acquiring destination location information, which is the location information of the destination;
setting a first range based on the destination location information;
Identifying a node information group, which is a set of node information associated with the name and location information and associated with the location information included in the first range;
Extracting a name group linked to each node information in the identified node information group;
Determining a name of a large-scale facility corresponding to the destination based on the extracted name group when a first condition regarding the number of node information in the identified node information group is satisfied;
A method of processing information, comprising: to the computer,
Acquiring destination location information, which is the location information of the destination;
setting a first range based on the destination location information;
Identifying a node information group, which is a set of node information associated with the name and location information and associated with the location information included in the first range;
Extracting a name group linked to each node information in the identified node information group;
Determining a name of a large-scale facility corresponding to the destination based on the extracted name group when a first condition regarding the number of node information in the identified node information group is satisfied;
The program that causes the to run.

Images