JP2016024619A - Information processing method, information processing device and information processing program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve the problem that the geographical range of a point to be clustered becomes too wide.SOLUTION: A generation part 111 generates a new cluster N in the case that an inter-cluster distance L between an element and an element or between an element and a cluster is the smallest in a distance matrix. A specification part 112 registers the inter-cluster distance L in the distance matrix in the case that an inter-cluster distance L between the new cluster N and another cluster is smaller than a prescribed value. An output part 113 makes a cluster whose near-miss degree is equal to or more than a prescribed value an attention point among clusters registered in the distance matrix, and outputs information of the attention point to an attention point list.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an information processing method, an information processing apparatus, and an information processing program.

  There is a technology that warns the driver before driving a dangerous situation when driving a car. In such a technique, every time a situation around a vehicle is detected, data indicating a so-called near-miss event that the driver may have felt dangerous, such as near-miss or close-up, is accumulated. In the accumulated data, data having similar occurrence points are grouped to generate a plurality of groups. Representative data is created for each group, and each representative data is registered in the near-miss map.

  In addition, representative data that matches the situation around the vehicle is retrieved from the near-miss map while the vehicle is being driven. As a result of the search, the driver is warned when representative data that matches the situation around the car is registered in the near-miss map.

International Publication No. 2009/128398 JP 2013-196366 A

  In order for the spot (spot) indicated in the near-miss map to be accurate, it is important that the geographical range of the spot does not spread more than necessary. There is a problem that is larger than necessary.

  For example, when trying to identify a spot using conventional mesh technology, the size of the mesh section should be the size of the expected spot to ensure that the mesh frame encompasses each spot without disruption. I had to set it above the size. If this happens, it will not be possible to easily show on the map the frequent occurrence points of near misses.

  Moreover, if the geographical range of the point (spot) obtained in this way is too wide, for example, if there are a plurality of sudden braking areas in a large intersection, it becomes a large spot that encompasses them. It will be. For example, the entire intersection is regarded as one spot. If this happens, information on where to pay attention in the intersection will be lost, and the value as a near-miss map will be reduced.

  In one aspect, there is provided an information processing method, an information processing apparatus, and an information processing program capable of preventing the geographical range of clustered points from becoming too wide when clustering a plurality of points associated with an event. For the purpose.

  In the first proposal, the information processing method causes a computer to execute a process for generating a cluster including a plurality of points based on the distance between points for a point group associated with a specific event. The information processing method causes a computer to execute a process of specifying the longest distance among the distances between each of a plurality of points included in the generated cluster and a specific point that is not included in the cluster. The information processing method causes a computer to execute a process of generating and outputting a cluster including a plurality of points included in the cluster and the specific point when the specified longest distance is within a predetermined distance.

  According to one embodiment of the present invention, when clustering a plurality of points associated with an event, it is possible to prevent the geographical range of the points to be clustered from becoming too wide.

FIG. 1 is a diagram illustrating an example of the overall configuration of the system according to the first embodiment. FIG. 2 is a diagram for explaining generation of the event DB. FIG. 3 is a diagram for explaining a process executed by the information processing apparatus according to the first embodiment. FIG. 4 is a diagram illustrating an example of the event DB. FIG. 5 is a diagram illustrating an example of the attention point list. FIG. 6 is a diagram for explaining an example of a distance matrix. FIG. 7 is a diagram illustrating an example of the aggregation point table. FIG. 8 is a first diagram for explaining the clustering process. FIG. 9 is a second diagram for explaining the clustering process. FIG. 10 is a third diagram for explaining the clustering process. FIG. 11 is a fourth diagram for explaining the clustering process. FIG. 12 is a fifth diagram for explaining the clustering process. FIG. 13 is a sixth diagram for explaining the clustering process. FIG. 14 is a first diagram for explaining the generated cluster. FIG. 15 is a second diagram for explaining the generated cluster. FIG. 16 is a third diagram for explaining the generated cluster. FIG. 17 is a diagram for explaining identification of a caution point. FIG. 18 is a diagram illustrating an example of the entire processing flow. FIG. 19 is a diagram illustrating an example of the flow of initial processing. FIG. 20 is a diagram illustrating an example of a flow of data reading processing. FIG. 21 is a diagram illustrating an example of a flow of record reading processing. FIG. 22 is a diagram illustrating an example of a flow of a distance matrix creation process. FIG. 23 is a diagram illustrating an example of a flow of clustering processing. FIG. 24 is a diagram illustrating an example of a flow of connection check processing. FIG. 25 is a diagram illustrating an example of the flow of the inter-cluster distance calculation process. FIG. 26 is a diagram illustrating an example of a flow of attention point output processing. FIG. 27 is a diagram illustrating an example of the flow of printer output processing. FIG. 28 is a second diagram for explaining collection of event logs. FIG. 29 is a diagram illustrating a hardware configuration of the information processing apparatus according to the first embodiment. FIG. 30 is a diagram illustrating a configuration example of a program operating on a computer.

  Embodiments of an information processing apparatus disclosed in the present application will be described below in detail with reference to the drawings. This scope of rights is not limited by this embodiment. Each embodiment can be appropriately combined within a range in which processing contents do not contradict each other.

(Configuration of entire system of Embodiment 1)
FIG. 1 is a diagram illustrating an example of the overall configuration of the system according to the first embodiment. As illustrated in the example of FIG. 1, the information processing apparatus 100 includes a communication unit 101, a control unit 110, and a storage unit 120. The storage unit 120 includes an event DB (Data Base) 121 and a caution point list 122. The storage unit 120 corresponds to, for example, a RAM (Random Access Memory), a ROM (Read Only Memory), a semiconductor memory element such as a flash memory, or a storage device such as a hard disk or an optical disk. Hereinafter, each configuration of the storage unit 120 will be described in detail.
(Each component of the storage unit)

  The generation of the event DB will be described with reference to FIG. FIG. 2 is a first diagram for explaining generation of the event DB. As shown in the example of FIG. 2, the computer 11 is connected to the drive recorders 10 </ b> A to 10 </ b> C and the printer 12 so as to communicate with each other. The computer 11 is an example of the information processing apparatus 100.

  The computer 11 reads the SD memory card set in the drive recorders 10A to 10C and collects event logs from the drive recorders 10A to 10C. The computer 11 generates an event DB 121 based on the collected event log. For example, the computer 11 generates the event DB 121 according to FIG. The computer 11 creates an attention point list 122 and outputs it to the printer 12.

  Next, the determination of “near presence / absence” in the event DB 121 will be described. The computer 11 determines whether there is a failure based on each event log. For example, the computer 11 determines whether or not there has been a sudden brake based on a change in acceleration or speed in the traveling direction of the event log. When the drive recorders 10A to 10C determine that there is a sudden brake during traveling, the drive recorders 10A to 10C store the data at that time in the SD card as an event log.

(Processes executed by the information processing apparatus according to the first embodiment)
The overall process executed by the information processing apparatus 100 according to the first embodiment will be described with reference to FIG. FIG. 3 is a diagram for explaining a process executed by the information processing apparatus according to the first embodiment. The driving caution point list creation program 70 has the same functions as the generation unit 111, the specification unit 112, and the output unit 113. The driving caution point list creation program 70 generates an event DB 121 that collects event logs. The driving caution point list creation program 70 receives input data 73a of various parameters such as a reference position (zx, zy), a censoring distance limit_L (m), and a driving caution reference P_hyr. The driving caution point list creation program 70 reads the event log from the event DB 121.

  The work table 72 is a work area secured in a memory included in the information processing apparatus 100. The work table 72 includes a distance matrix and an aggregation point table. The driving caution point list creation program 70 creates a distance matrix that stores the distances between the elements distributed on the map based on the read event log. The driving caution point list creation program 70 generates clusters based on the distance matrix and stores information on each cluster in the aggregation point table. The aggregation point table associates information such as the cluster number, the number of events, the number of near incidents, the latitude and longitude of the center of gravity position of each cluster with each cluster. The driving caution point list creation program 70 calculates the nearness degree of each cluster based on the number of events and the number of near incidents associated with the clusters in the aggregation point table. The driving caution point list creation program 70 specifies the position of the center of gravity of a cluster whose nearness is equal to or higher than the driving caution criterion P_hyr as the caution point. The driving caution point list creation program 70 outputs the cluster information related to each specified caution point to the caution point list 122. Note that the driving caution point list creation program 70 outputs a message 73b such as various parameters at the end of the process to the monitor.

  An example of the data structure of the event DB 121 will be described with reference to FIG. FIG. 4 is a diagram illustrating an example of the event DB. The event DB 121 is a database that records event information collected from each vehicle. As shown in the example of FIG. 4, the event DB 121 includes an ID (Identification), an event number, a date, a time, a vehicle number, a lateral acceleration, a traveling acceleration, a speed, and a longitude. Corresponding latitude, near-missing and comment. “ID” is an identification number uniquely assigned to event information collected from each vehicle. “Event number” is a number assigned to an event when the event is collected. “Date” is the date when the event occurred. “Time” is the time when the event occurred. The “vehicle number” is an identification number of a vehicle for which an event is collected. “Acceleration” is the acceleration of the vehicle. The “lateral direction” is acceleration in the left or right direction with reference to the traveling direction of the vehicle. The “traveling direction” is acceleration in the traveling direction of the vehicle. “Position” is the position of the vehicle when the event occurs. “Latitude” is the latitude at which the vehicle was located when the event occurred. “Longitude” is the longitude of the position where the vehicle was when the event occurred. “Presence / absence” indicates a determination result as to whether or not there was a near-miss when an event occurred. The near-miss represents an event that the driver may have felt dangerous during the operation of the vehicle, for example, something suddenly pops out in front of the vehicle and the brake is applied. When the presence / absence of a near miss is “0”, it means that no near miss has occurred. When the presence / absence of a near miss is “1”, it means that a near miss has occurred. “Comment” represents the specific content of the event.

  Whether each event information corresponds to a near-miss or not is determined by, for example, reproducing and viewing the image data recorded with the occurrence of the event to determine whether or not each event was a near-miss. Judgment may be given and data in the near / missed column may be added. Further, based on values such as acceleration and speed included in the information of each event, it may be determined whether or not it corresponds to a near-miss, and the determined result may be stored in a near-miss / not-available column. In the example of FIG. 4, an example in which the data of each item is stored in association with each other is shown. However, if the relationship between the items associated with each other is maintained in the above description, the data is stored in another storage. It doesn't matter how you do it. The same applies to the examples of FIGS.

  An example of the data structure of the caution point list 122 will be described with reference to FIG. FIG. 5 is a diagram illustrating an example of the attention point list. The caution point list 122 is a list indicating the position and caution level of the caution point specified on the near map. The caution point is, for example, the position of the center of a spot where the risk of an accident is high. That is, the attention point is the center of gravity position of the cluster. The caution point list 122 includes a record representing cluster information and a record representing element information. An element represents an occurrence point of an event included in a cluster. The information processing apparatus 100 attaches the same group number to records related to the same cluster. In the example of FIG. 5, a record in which 0 is stored in “ID” is a record representing cluster information. A record in which a value other than 0 is stored in “ID” is a record representing element information. For example, in the example of FIG. 5, the information processing apparatus 100 stores the cluster information related to the group number “79” in the record in the first effective row. Further, the information processing apparatus 100 stores information on the element relating to the group number “79” in the records in the second to eighth effective rows.

  The caution point list 122 includes a group number, an ID, a number / event number, a date, a time, a vehicle number, a nearness degree, a longitude, a latitude, a distance, a driving caution point, and a comment. Associate. The “group number” is an identification number uniquely assigned to each generated cluster. Records belonging to the same cluster are assigned the same group number. When it is an element of a cluster, “ID” is the same as the “ID” of the event DB, that is, the identification number uniquely assigned to the event information collected from each vehicle. The value is larger than “0” is added to the cluster record. In “number of events / event number”, the number of events is stored in the cluster record, and the event number is stored in the element record. “Date” represents the date on which the event occurred. “Time” represents the time when the event occurred. “Vehicle number” is the identification number of the vehicle that collected the event log.

  The “degree of incident” stores the degree of incident in the record of the cluster. The incident level represents the ratio of the number of incidents to the total number of events included in the cluster. For example, when the total number of events is 10 and the number of near incidents is 5, the near incident degree is 50%. In addition, in the element record, whether or not a near occurrence has occurred is stored. For example, when “1” is stored, it indicates that a near occurrence has occurred, and when “0” is stored, it indicates that a near occurrence has not occurred.

  As for “longitude”, the longitude of the point of interest is stored in the cluster record, and the longitude of the point where the event occurred is stored in the element record. As for “latitude”, the latitude of the point of interest is stored in the cluster record, and the latitude of the point where the event occurred is stored in the element record.

  “Distance” is the distance between the attention point and the point where the event occurred. The “driving caution point” is an index that represents a driving caution level. For example, the “driving caution point” is evaluated in three stages of “0”, “1”, and “2”. The driving caution point “0” indicates that the point is less dangerous when driving. The driving caution point “1” represents a point where danger is likely to occur during driving. The driving caution point “2” represents a point where danger is likely to occur during driving. Details regarding evaluation of driving caution points will be described later. “Comment” represents specific contents of the event.

(Configuration of control unit)
As illustrated in the example of FIG. 1, the control unit 110 of the information processing apparatus 100 includes a generation unit 111, a specification unit 112, and an output unit 113. The function of the control unit 110 can be realized, for example, by a CPU (Central Processing Unit) executing a predetermined program. Moreover, the function of the control part 110 is realizable by integrated circuits, such as ASIC (Application Specific Integrated Circuit) and FPGA (Field Programmable Gate Array).

  The information processing apparatus 100 includes a generation unit 111 that generates a cluster including a plurality of points based on the distance between the points for a point group associated with a specific event. The information processing apparatus 100 includes a specifying unit 112 that specifies the longest distance among distances between a plurality of points included in the generated cluster and a specific point that is not included in the cluster among the point group. . The information processing apparatus 100 includes an output unit 113 that generates and outputs a cluster including a plurality of points included in the cluster and the specific point when the specified longest distance is within a predetermined distance.

  The information processing apparatus 100 includes a generation unit 111 that generates a plurality of clusters each including a plurality of points based on the distance between the points for the point group associated with the specific event. The information processing apparatus 100 includes a plurality of points included in the first cluster among the plurality of generated clusters and a plurality of points included in the second cluster among the plurality of generated clusters. Among the distances, a specifying unit 112 that specifies the longest distance is included. The information processing apparatus 100 generates a cluster including a plurality of points included in the first cluster and a plurality of points included in the second cluster when the specified longest distance is within a predetermined distance. And an output unit 113 for outputting.

  The information processing apparatus 100 further determines the cluster based on the number of points related to the first event associated with the same cluster and the number of points associated with the second event among the points. A calculation unit for calculating an evaluation value for. The output unit 113 outputs cluster information and an evaluation value. Further, for example, the second event is a near-miss, for example. Note that the output unit 113 may have the function of a calculation unit.

  The calculation unit calculates an evaluation value based on a ratio of the number of points associated with the second event to the number of points associated with the first event. Further, the output unit 113 outputs a cluster whose evaluation value is equal to or greater than a predetermined value.

  The information processing apparatus 100 assigns a series of cluster numbers to the event point or a cluster generated by bundling the event points, and which cluster number is valid in the process of adding / deleting clusters. Determine whether. Alternatively, the information processing apparatus 100 stores a series of cluster numbers as an associative array in order to obtain a valid cluster number.

  Further, the distance between one cluster and another cluster or another point is stored as a binary heap configured such that the parent node is smaller than the child node. The above associative array and binary heap are generated and held in the memory area in the storage unit 120. For example, the cluster number list is an example of an associative array, and the distance matrix is an example of a binary heap using a binary tree. Hereinafter, the binary heap is simply referred to as heap. The generated cluster number list and heap are stored in a memory inside the information processing apparatus 100.

  Hereinafter, the configuration of the communication unit 101 and the control unit 110 of the information processing apparatus 100 will be described in detail. The communication unit 101 receives an event log recorded by a drive recorder installed in each vehicle. The communication unit 101 is realized by, for example, a USB (Universal Serial Bus). The communication unit 101 communicates with each drive recorder via a bus.

  The generation unit 111 generates a cluster including a plurality of points based on the distance between the points for the point group associated with the specific event. That is, the generation unit 111 is a processing unit that generates a cluster having a plurality of elements. An element is a point where an event occurs. A procedure for generating a new cluster N will be described. First, the generation unit 111 acquires an intercluster distance L between two elements p and q corresponding to the minimum intercluster distance among the intercluster distances L registered in the distance matrix. The distance matrix is a heap configured such that the parent node is smaller than the child node. The generation unit 111 can acquire the minimum inter-cluster distance L by acquiring the inter-cluster distance registered at the head of the distance matrix.

  An example of the data structure of the distance matrix will be described with reference to FIG. FIG. 6 is a diagram for explaining an example of a distance matrix. The distance matrix holds the element-element, element-cluster, and cluster-cluster intercluster distances. As shown in the example of FIG. 6, the distance matrix is a heap configured such that the inter-cluster distance of the parent node is smaller than that of the child node. The data included in each node represents the element number or cluster number of the element and the intercluster distance. For example, (A number, B number, 0.8) of node 1 (1) indicates that the distance between element A and element B is “8”. Further, (A number, C number, 1.1) of the node 1 (2) indicates that the distance between the element A and the element C is “11”.

  The shortest inter-cluster distance is stored in the head node of the distance matrix. The generation unit 111 acquires the shortest inter-cluster distance by extracting the head node from the distance matrix. Each time the generation unit 111 extracts the first node from the distance matrix, the generation unit 111 reconfigures each node of the distance matrix so that the relationship that the parent node is smaller than the child node is maintained.

  A procedure for creating a distance matrix will be described. First, the information processing apparatus 100 extracts two element numbers stored in the cluster number list. The cluster number list holds all element numbers. The information processing apparatus 100 calculates the inter-cluster distance between the two elements using the distance conversion coefficients ux, uy for the latitude difference and the longitude difference between the two elements extracted from the cluster number list. Details regarding the cluster number list will be described later.

  The distance conversion coefficients ux and ui are coefficients used when calculating the inter-cluster distance from the latitude difference and longitude difference between the positions of two elements or clusters. It is the converted value to the distance (km). The information processing apparatus 100 receives an input of a reference position (zx, zy) that is a latitude and longitude serving as a reference for an area for creating a near-miss map, and calculates a distance conversion coefficient ux, uy based on the reference position (zx, zy). Ask. The information processing apparatus 100 may adjust the value of (zx, zy) so that the center of the area where the near-miss map is created becomes the reference position.

  Next, when the calculated intercluster distance L between the two elements is smaller than “limit_L2”, the information processing apparatus 100 registers the intercluster distance L in the distance matrix in association with the number of the two elements. For example, when limit_L2 is given as “20”, the information processing apparatus 100 extracts the element number “A” and the element number “B” from the cluster number list, and clusters between the element A and the element B. When the distance is “8”, (A number, B number, 8) is registered in the distance matrix.

  In the same manner as described above, the information processing apparatus 100 extracts two other element numbers from the cluster number list and repeats the process of registering the inter-cluster distance L to create a distance matrix. The distance matrix is a heap configured such that the inter-cluster distance of the parent node is smaller than that of the child node. Each time the information processing apparatus 100 registers the inter-cluster distance L in the distance matrix, the information processing apparatus 100 reconfigures each node so that the inter-cluster distance L of the parent node is smaller than the child node.

  Thus, the information processing apparatus 100 registers the inter-cluster distance L in the distance matrix only when the inter-cluster distance L between the two elements is smaller than “limit_L2”. As a result, the information processing apparatus 100 can reduce the size of the distance matrix and save the memory area as compared to the case where the inter-cluster distances between all elements are registered in the distance matrix.

  The generation unit 111 can acquire the two elements p and q corresponding to the minimum inter-cluster distance L by acquiring the two elements registered at the head of the heap that is the distance matrix. For this reason, the generation unit 111 can reduce the time required to calculate and search for the two minimum elements each time in normal clustering. The description of the distance matrix is as described above.

  The generation unit 111 registers the new cluster N generated based on the two elements acquired from the distance matrix in the aggregation point table. Here, an example of the data structure of the aggregation point table will be described with reference to FIG. FIG. 7 is a diagram illustrating an example of the aggregation point table. The aggregation point table is a table that stores information of each generated cluster. As shown in the example of FIG. 7, the aggregation point table associates the cluster number, the number of events, the number of incidents, the latitude, the longitude, the left pointer, and the right pointer. The “cluster number” is a sequential identification number uniquely assigned to each generated cluster. As illustrated in the example of FIG. 7, when the total number of events is 327, the generation unit 111 assigns cluster numbers after 328 to each cluster. “Number of events” is the total number of events included in the cluster. The event number k [N] is, for example, the number of sudden braking. “Number of incidents” is the total number of incidents included in the cluster.

  “Latitude” is the latitude of the center of gravity of the cluster. “Longitude” is the longitude of the center of gravity of the cluster. The generation unit 111 may specify the position of the center of gravity of the cluster based on the position of each element included in the cluster. For example, when the cluster has three elements, the generation unit 111 sets the position of the center of gravity of a triangle formed by the three elements as the latitude and longitude of the cluster.

  “Left pointer” and “Right pointer” are the numbers or cluster numbers of the elements under the cluster, and the left and right are determined so that the cluster number indicated by “Right pointer” is larger than the cluster number indicated by “Left pointer”. And The “left pointer” and “right pointer” will be described with examples. For example, when generating the cluster N ′ by adding the element M to the cluster N having a plurality of elements, the generation unit 111 performs the following process. First, the generation unit 111 acquires the intercluster distance L between the cluster N and the element M. When the intercluster distance L is within a predetermined condition, the generation unit 111 adds the element M to the cluster N and generates a cluster N ′. In this case, the subordinate clusters of the cluster N ′ are the cluster N and the element M. The generation unit 111 sets the cluster number of the cluster N and the number of the element M to “left pointer” and “right pointer”, respectively. In this way, the generation unit 111 sets “left pointer” and “right pointer”. In this case, the relationship of N ′> N, N> M always holds for each cluster and element number.

  When “left pointer” or “right pointer” is a cluster number, the cluster corresponding to the cluster number also has “left pointer” and “right pointer”. As in the example of FIG. 7, for example, the subordinate cluster of the cluster with the cluster number “335” is the cluster with the cluster number “328” and the element with the element number “210”. Further, the subordinate cluster of the cluster having the cluster number “328”, which is a subordinate cluster, is an element having an element number “156” and an element having an element number “141”. The description of the structure of the aggregation point table is as described above.

  Next, a procedure for creating the aggregation point table will be described. First, the generation unit 111 assigns a cluster number larger than the maximum element number to the new cluster N generated based on the two elements p and q extracted from the cluster number list. For example, when the maximum element number is 327, the generation unit 111 assigns a cluster number of 328, 329, 330,... To each new cluster.

  The generation unit 111 sets a pointer for each of the two subordinate clusters of the new cluster N. The subordinate cluster is a cluster included under the new cluster N. For example, the generation unit 111 stores the cluster number of one subordinate cluster in the left pointer of the new cluster N. Further, the generation unit 111 stores the cluster number of the other subordinate cluster in the right pointer of the new cluster N. The cluster numbers stored in the left pointer and the right pointer are registered in the aggregation point table. In the following, the left pointer of the new cluster N is represented by pair1 [N]. The right pointer of the new cluster N is represented by pair2 [N].

  The generation unit 111 calculates the number of events k [N] by adding up the number of events in the subordinate cluster of the new cluster N. For example, when the generation unit 111 generates a new cluster N with two elements p and q, the new cluster N includes two events, so the number of events of the new cluster N is 2. The generation unit 111 stores the counted number of events k [N] in the aggregation point table.

  Further, the generation unit 111 calculates the number of incidents H [N] by adding the number of incidents of the subordinate clusters of the new cluster N. For example, the generation unit 111 sets the number of incidents of the new cluster N to 2 when both elements p and q are near. The generation unit 111 sets the number of incidents of the new cluster N to 1 when one of the two elements p and q is near. In addition, the generation unit 111 sets the number of incidents of the new cluster N to 0 when none of the two elements p and q is missing. The generation unit 111 stores the counted incident number H [N] in the aggregation point table.

  The generation unit 111 calculates the position of the center of gravity of the new cluster N. For example, the generation unit 111 sets the coordinates of the center between two elements p and q of the new cluster N as the position of the center of gravity of the new cluster N. The generation unit 111 registers the calculated position of the center of gravity of the new cluster N in the aggregation point table as latitude and longitude. In this way, the generation unit 111 generates an aggregation point table.

  Next, the generation unit 111 deletes the cluster numbers of the elements p and q from the cluster number list. Further, the generation unit 111 adds the cluster number N of the new cluster to the cluster number list.

  Here, the cluster number list will be described. The cluster number list is a list indicating whether or not an element or a cluster exists. The cluster number list includes the element number and the cluster number of the generated cluster. For example, the cluster number list indicates that an element exists if the element number is stored in the list. The cluster number list indicates that a cluster has been generated if a cluster number is stored in the list. In the following description, element numbers and cluster numbers may be collectively referred to as cluster numbers.

  In the cluster number list, element numbers and cluster numbers are managed by serial numbers. However, deletion or addition of elements or clusters may cause missing numbers. For example, the information processing apparatus 100 assigns a number to each element with a serial number, and assigns a number one or more larger than the number of the largest element to the generated new cluster. For example, in the cluster number list, the initial state is elements A, B, C, and D, and thereafter, B and C are combined to generate a new cluster N1. In such a case, the numbers assigned to the initial elements A, B, C, D and the new cluster N1 are as follows. The initial state is (A, B, C, D) = (1, 2, 3, 4). Yes, then (A, D, N1) = (1, 4, 5). The description of the cluster number list is as described above.

  The identifying unit 112 identifies the longest distance among the distances between each of the plurality of points included in the cluster generated by the generating unit 111 and a specific point that is not included in the cluster in the point group. In addition, the specifying unit 112 includes a plurality of points included in the first cluster among the plurality of clusters generated by the generation unit 111 and a plurality of points included in the second cluster among the plurality of generated clusters. The longest distance among the distances to each point is specified. That is, the specifying unit 112 is a processing unit that obtains an inter-cluster distance between the new cluster N generated by the generating unit 111 and each of the other clusters and registers it in the distance matrix. For example, the distance matrix holds the intercluster distance between all elements that satisfy the condition of the cutoff distance limit_L, but does not hold the intercluster distance between the new cluster N and each of the other clusters. Therefore, every time a new cluster N is generated, the specifying unit 112 obtains an intercluster distance between the new cluster N and each of the other clusters, and performs a process of registering in the distance matrix.

  The identifying unit 112 first performs a connection check process and secondly performs an inter-cluster distance calculation process. Hereinafter, the connection check process and the inter-cluster distance calculation process executed by the specifying unit 112 will be described.

  In the first connection check process, the specifying unit 112 determines whether or not the condition that the intercluster distance L between the new cluster N and each of the other clusters is within “limit_L2” is satisfied. The specifying unit 112 registers the intercluster distance L between the new cluster N and each of the other clusters in the distance matrix 122 when the intercluster distance L satisfies the condition “limit_L2” or less.

  The censoring distance limit_L is a parameter for setting the size of the cluster to be generated. As the censoring distance limit_L increases, the size of the generated cluster on the map increases, and as the censoring distance limit_L decreases, the size of the generated cluster on the map decreases. Further, the information processing apparatus 100 calculates “limit_L2” obtained by multiplying the cutoff distance limit_L by 2. The information processing apparatus 100 uses “limit_L2” when determining whether to add an element to the cluster.

  The procedure of the first connection check process will be described. First, the specifying unit 112 acquires the cluster number i from the cluster number list. Note that the cluster number i is not the number of elements or clusters corresponding to the “left pointer” and “right pointer” that directly constitute the new cluster N, and any number from the top of the cluster number list to the last registered number. Element or cluster number.

  The specifying unit 112 registers the inter-cluster distances L (p, i) and L (q, i) between the elements p and q included in the new cluster N and the cluster associated with the cluster number i, in the distance matrix. It is determined whether or not. L (p, i) represents the inter-cluster distance between element p and cluster i, and L (q, i) represents the inter-cluster distance between element q and cluster i. Although an example in which p and q are elements has been described, one or both of p and q may be a cluster.

  When both of the two inter-cluster distances L (p, i) and L (q, i) are registered in the distance matrix, the specifying unit 112 determines that the element or cluster related to the cluster number i is within “limit_L2”. Judge that there is. On the other hand, when both or one of the two inter-cluster distances is not registered in the distance matrix, the specifying unit 112 determines that the element or cluster related to the cluster number i is outside “limit_L2”. Hereinafter, the cluster associated with cluster number i is referred to as remaining cluster i. As described above, the specifying unit 112 performs a connection check process.

  In the second inter-cluster distance calculation process, the specifying unit 112 determines the inter-cluster distance L for the new cluster N and each remaining cluster i that satisfy the condition that the inter-cluster distance L is within “limit_L2” in the first connection check process. Ask. The procedure of the second inter-cluster distance calculation process will be described. First, the specifying unit 112 determines whether or not the intercluster distance L between the remaining cluster i and the new cluster N is registered in the distance matrix.

  The identifying unit 112 determines that the inter-cluster distances L (i, p) and L (i, q) between the remaining cluster i and the clusters p and q constituting the new cluster N are already registered in the distance matrix. Only when this is done, the inter-cluster distance L (i, N) is registered in the distance matrix. The identification unit 112 determines that the intercluster distances L (i, p) and L (i, q) between the remaining cluster i and the clusters p and q constituting the new cluster N are already registered in the distance matrix. If not, make a decision not to register. Here, both L (i, p) and L (i, q) are already registered in the distance matrix means that the condition that L (i, N) is within “limit_L2” is satisfied. There is nothing else.

  A procedure of processing in which the specifying unit 112 registers the intercluster distance L between the remaining cluster i and the new cluster N will be described. First, the specifying unit 112 uses the left and right pointers (pair1 [i], pair2 [i]) set for the remaining cluster i, and clusters of two clusters s and t included in the remaining cluster i from the aggregation point table. Get each number. The specifying unit 112 acquires two inter-cluster distances between the two clusters s and t and the new cluster N from the distance matrix. The specifying unit 112 sets the longer one of the acquired two inter-cluster distances as the inter-cluster distance L between the remaining cluster i and the new cluster N. Note that the cluster associated with cluster number s is cluster s, and the cluster associated with cluster number t is cluster t. As described above, the specifying unit 112 performs the inter-cluster distance calculation process.

  Next, the specifying unit 112 registers the inter-cluster distance L (i, N) between the remaining cluster i and the new cluster N obtained by the inter-cluster distance calculation process in the distance matrix. The specifying unit 112 also determines the intercluster distance L (i, p) or L (p, i) and L (i, q) or L (q, i) between the remaining cluster i and the elements p and q. Are removed from the distance matrix. The identifying unit 112 performs the inter-cluster distance calculation processing for other remaining clusters that satisfy the condition that the inter-cluster distance L is within “limit_L2”, and registers the inter-cluster distance L in the distance matrix in the same manner as the remaining cluster i. To do.

  In this way, each time the specifying unit 112 registers the intercluster distance between the new cluster N and the remaining cluster i in the distance matrix, the specifying unit 112 calculates the intercluster distance between the remaining cluster i and each subordinate cluster of the new cluster N as a distance matrix. Delete from. Thereby, the information processing apparatus 100 can reduce the capacity of the distance matrix and save the memory area.

  The output unit 113 generates and outputs a cluster including a plurality of points included in the cluster and the specific point when the longest distance specified by the specifying unit 112 is within a predetermined distance. The output unit 113 generates a cluster including a plurality of points included in the first cluster and a plurality of points included in the second cluster when the longest distance is within a predetermined distance. Output. In other words, the output unit 113 is a processing unit that generates and outputs a cluster including the cluster generated by the generation unit 111 and other elements or other clusters. The output unit 113 performs a process of generating a new cluster N when the pair having the smallest inter-cluster distance L in the distance matrix is a cluster generated by the generation unit 111 and another element or another cluster. Do.

  Note that the process of generating a new cluster N performed in the output unit 113 is the same as that of the generation unit 111, and thus description thereof is omitted. The specifying unit 112 also performs processing for registering the intercluster distance L in the distance matrix for the new cluster N generated by the output unit 113.

  Further, the output unit 113 outputs the generated cluster information. For example, the output unit 113 outputs a caution point list 122 in which information on caution points distributed on the near map is collected.

  The output unit 113 calculates a nearness level for each cluster, and identifies a cluster whose nearness level is equal to or higher than the driving attention criterion P_hyr. Here, the driving attention criterion P_hyr is a parameter serving as a reference when evaluating the driving attention point for each cluster. The driving caution point is a risk at the time of driving at the caution point, and is set based on whether or not the near degree is equal to or higher than the driving caution standard P_hyr. In addition, the example regarding evaluation of a driving attention point is mentioned later.

  For example, the output unit 113 determines the position of the center of gravity of each identified cluster as the attention point. The output unit 113 enumerates and outputs information regarding each determined caution point in the caution point list 122. For example, the output unit 113 outputs the attention point list 122 according to FIG. Hereinafter, the output of the attention point list 122 will be described in detail.

  First, the output unit 113 uses item names such as group number, ID, number of events / event number, date, time, vehicle number, nearness, longitude, latitude, distance, driving caution point, and comments as the header of the caution point list 122. Is output. The output unit 113 acquires the cluster number j from the cluster number list. The output unit 113 calculates the nearness P [j] related to the cluster j by the following equation (1).

  Incident degree P [j] = number of incidents H [j] / number of events k [j] (1)

  The output unit 113 determines whether or not the calculated nearness P [j] is equal to or greater than the driving attention criterion P_hyr. The output unit 113 determines that the cluster j is a caution point when the near degree P [j] is equal to or greater than the driving caution criterion P_hyr, and sets the driving caution point to “2”. Further, the output unit 113 determines that the cluster j may be a caution point when the nearness degree P [j] is less than the driving caution standard P_hyr and the near incident number H [j] is one or more, and driving Set the attention point to "1". When the incident number H [j] is 0, the output unit 113 determines that the point is not a caution point, and sets the driving caution point to “0”. For example, in the example of FIG. 5, the output unit 113 sets the driving caution point to “2” because the nearness of the cluster with the group number “79” is “28.60%” and the driving caution standard is 20% or more.

  The output unit 113 outputs information regarding the cluster j to the attention point list 122. For example, the output unit 113 outputs the number of events k [j], the nearness degree P [j], the position of the center of gravity, and the driving caution point of the cluster j to the caution point list 122 in association with one record. Information about the cluster output to the caution point list 122 will be described with reference to FIG. For example, as shown in the first record of FIG. 5, the output unit 113 (group number, number of cases / event number, nearness, longitude, latitude, driving caution point) = (79,7,28.60%, 139.699509,35.531134, 2) is output.

  The output unit 113 searches the elements included in the cluster j with reference to the aggregation point table, and outputs information on each element to the attention point list 122. Processing for outputting information of each element to the attention point list 122 will be described. First, the output unit 113 acquires the cluster numbers under the cluster j from the “left pointer” and “right pointer” of the aggregation point table, and stores the cluster numbers in pair1 [j] and pair2 [j].

  The output unit 113 determines whether each of the cluster numbers stored in the right pointer Pair1 [j] and the left pointer Pair2 [j] is a cluster or an element. The output unit 113 determines a cluster when the cluster number is greater than the total number of events nc. The output unit 113 determines that the element is an element when the cluster number is equal to or less than the total number nc of event logs.

  When determining that one or both of the cluster numbers (Pair1 [j], Pair2 [j]) are elements, the output unit 113 outputs element information. On the other hand, when one or both of the cluster numbers (Pair1 [j], Pair2 [j]) are determined to be clusters, the output unit 113 is subordinate to the cluster associated with the cluster numbers (Pair1 [j], Pair2 [j]). Get the cluster number of. The output unit 113 recursively repeats the process of acquiring the cluster number until it is determined that the acquired cluster number is an element.

  Information of each element output to the attention point list 122 will be described with reference to FIG. The output unit 113 outputs each element (group number, ID, number / event number, date, time, vehicle number, nearness, longitude, latitude, distance, comment) as follows. For example, as shown in the second record in FIG. 5, the output unit 113 (79,99,2011-01_04_00000058_20101230174810,2010 / 12 / 30,17: 48,58,1,139.699553,35.531047,0.010) ) Is output. Moreover, the output part 113 is (79,249,2011-01_10_00000053_20110110014147,2011 / 1 / 10,1: 41,53,1,139.699581,35.531047,0.012, side collision, impact sound as shown in the third record in FIG. 5!) Is output. Note that the nearness level column of the caution point list 122 indicates the presence or absence of each element.

(Specific example of clustering processing)
The clustering process will be described with a specific example with reference to FIGS. FIG. 8 is a first diagram for explaining the clustering process. As shown in the example of FIG. 8, elements A, B, C, and D are distributed in the near miss map 50a. The inter-cluster distance between element A and element B is “8”. The inter-cluster distance between element A and element C is “11”. The intercluster distance between element B and element C is “12”. The inter-cluster distance between the element C and the element D is “13”. The intercluster distance between element B and element D is “18”. The inter-cluster distance between element A and element D is “22”. The distance matrix in this case is shown in FIG. 6, for example. Here, the following description will be made based on the assumption that the value of limit_L2, which is a condition for adding a cluster, is “20”, for example. Note that the intercluster distance between the element A and the element D is larger than limit_L2, and thus is not registered in the distance matrix.

  FIG. 9 is a second diagram for explaining the clustering process. The generation unit 111 acquires the inter-cluster distance “8” between the element A and the element B stored at the head of the distance matrix illustrated in the example of FIG. 6, and generates a cluster N1 including the element A and the element B. . The generated cluster N1 is shown in the near-miss map 50b in the example of FIG. The generation unit 111 deletes the intercluster distance between the element A and the element B from the distance matrix.

  FIG. 10 is a third diagram for explaining the clustering process. The identifying unit 112 obtains an intercluster distance between the new cluster N1 and the element C in the near miss map 50c. Since the distance “11” between the element A and the element C and the distance “12” between the element B and the element C are both registered in the distance matrix, the identifying unit 112 compares them, and compares the longer element B− The distance “12” between the elements C is the intercluster distance between the cluster N1 and the element C. The identifying unit 112 registers the intercluster distance “12” between the cluster N1 and the element C in the distance matrix. Further, the specifying unit 112 deletes the inter-cluster distance between the element A and the element C and between the element B and the element C from the distance matrix.

  In addition, the identifying unit 112 attempts to obtain the inter-cluster distance between the cluster N1 and the element D in the near-miss map 50c, similarly to the element C. However, the distance “22” between the element A and the element D is not registered in the distance matrix, and only the distance “18” between the element B and the element D is registered. In this case, since the intercluster distance “22” between the cluster N1 and the element D is larger than limit_L2, the specifying unit 112 does not register the intercluster distance between the cluster N1 and the element D in the distance matrix. Further, the specifying unit 112 deletes the intercluster distance between the element B and the element D from the distance matrix.

  FIG. 11 is a fourth diagram for explaining the clustering process. FIG. 12 is a fifth diagram for explaining the clustering process. As shown in the example of FIG. 11, the intercluster distance between the cluster N1 and the element C is “12”. The generation unit 111 acquires the intercluster distance “12” between the cluster N1 and the element C stored at the head of the distance matrix, and generates a cluster N2 including the cluster N1 and the element C. The generated cluster N2 is shown in the near-miss map 50e in the example of FIG.

  FIG. 13 is a sixth diagram for explaining the clustering process. The identifying unit 112 attempts to obtain the intercluster distance between the new cluster N2 and the element D. However, the distance between the cluster N1 and the element D is not registered in the distance matrix as described above, and only the distance “13” between the element C and the element D is registered. In this case, the identifying unit 112 does not register the intercluster distance between the cluster N2 and the element D in the distance matrix because the intercluster distance between the cluster N2 and the element D is larger than limit_L2. Further, the specifying unit 112 deletes the intercluster distance between the element C and the element D from the distance matrix. Therefore, the near-miss map 50f shown in the example of FIG. 13 includes the cluster N2 and the element D. When the process proceeds to this state, the distance registered in the distance matrix is lost. That is, it means that there is no combination that satisfies the condition that the distance between clusters is smaller than “limit_L2”. Since the subsequent clustering cannot be executed, the clustering process ends.

(Example of generated cluster)
As described above, the generation unit 111 generates a new cluster N by combining two elements-elements and element-clusters having the shortest inter-cluster distance. On the other hand, the output unit 113 generates a cluster N by combining the cluster generated by the generation unit 111 and another element or another cluster. That is, the generation unit 111 and the output unit 113 generate a larger cluster by repeatedly combining the element-element, the element-cluster, and the cluster-cluster having the smallest inter-cluster distance among the remaining two inter-cluster distances. . The generation unit 111 and the output unit 113 repeat the process of generating clusters until the condition that the distance between clusters is smaller than “limit_L2” is not satisfied.

  An example of the generated cluster will be described with reference to FIGS. FIG. 14 is a first diagram for explaining the generated cluster. Element A to element L are distributed in the near-miss map 51a shown in the example of FIG. The vertical axis and horizontal axis in FIG. 14 are values representing the coordinate values of each element. For example, the value represents a distance (m) from some reference point (origin). The generation unit 111 generates a cluster by combining the element-element or the element-cluster having the shortest inter-cluster distance. The generation unit 111 generates a new cluster by combining the element-element or the element-cluster having the shortest inter-cluster distance among the remaining elements and clusters. Further, the output unit 113 performs a process of combining clusters and generating a new cluster when the combination having the shortest inter-cluster distance among the remaining elements and clusters is a cluster-cluster.

  For example, the generation unit 111 generates a new cluster N1 with the element G-element J having the shortest inter-cluster distance in the near-miss map 51a. Next, the generation unit 111 generates a new cluster N2 with an element D-element H having the shortest intercluster distance among the remaining elements and clusters. As described above, the generation unit 111 and the output unit 113 repeat the process of generating a new cluster based on the combination having the shortest inter-cluster distance.

  FIG. 15 is a second diagram for explaining the generated cluster. The vertical axis and the horizontal axis in FIG. 15 are values representing the coordinate values of each element, as in FIG. For example, the value represents a distance (m) from some reference point (origin). As a result of the generation unit 111 and the output unit 113 repeating the process of generating a new cluster based on the combination having the shortest inter-cluster distance, the near-miss map 51b shown in the example of FIG. 15 is generated. The near-miss map 51b includes an element I, a cluster N2 having an element D and an element H, a cluster N7 having an element A, an element G, an element J, and an element L, an element B, an element C, an element E, an element F, and Cluster N8 having element K is distributed. As shown in the example of FIG. 15, a new cluster is generated by combining elements or clusters whose inter-cluster distance is less than or equal to limit_L2. In this figure, in order to clearly indicate the range of each cluster of N2, N7, and N8, a polygon connecting elements included in each cluster is drawn on the outer periphery, but this indicates a spot on the near-miss map. It is an example for expressing. This is not the only way to represent a cluster in a bundled state.

  With reference to FIG. 16, the generation of a cluster will be described assuming that limit_L2 is “20”. FIG. 16 is a third diagram for explaining the generated cluster, and is referred to as a dendrogram. The vertical axis of the graph 52 is the intercluster distance (m). A to L listed at positions of the intercluster distance of 0 m or less in the graph 52 represent each element, and correspond to the elements A to L in FIGS. 14 and 15. On the other hand, N1 to N8 listed at positions with a distance between clusters of 0 m or more represent clusters, and partially correspond to the clusters N2, N7, and N8 in FIG. In the example of FIG. 16, the generation of the cluster N8 will be described as an example. The generation unit 111 generates the cluster N3 by combining the element E and the element F. The generation unit 111 generates the cluster N5 by combining the element C and the element K. The generation unit 111 generates a cluster N6 by combining the element B and the cluster N5. The output unit 113 generates a cluster N8 by combining the cluster N3 and the cluster N6. Here, the generation unit 111 and the output unit 113 determine the cluster N8 because the intercluster distance becomes larger than limit_L2 “20” when another element or cluster is combined with the cluster N8. In this way, a cluster N8 is generated. Note that the generation unit 111 and the output unit 113 similarly generate a cluster N2 and a cluster N7.

(Identification of caution points)
FIG. 17 is a diagram for explaining identification of a caution point. The example of FIG. 17 represents a near-miss map around the intersection where the road 20 and the road 21 intersect. On the road 21, a pedestrian crossing 22a and a pedestrian crossing 22b are installed. A mark “x” is attached to the point where the near-miss occurred. As shown in the example of FIG. 17, there are no crosswalks on the road 20, and pedestrians often jump out on the road 20, so there are many near-misses.

  The generation unit 111 generates a cluster Na and a cluster Nb in an area where near-misses frequently occur in the example of FIG. The generation unit 111 calculates the positions Nar and Nbr of the centers of gravity of the cluster Na and the cluster Nb based on the coordinates of the points marked with “x” in the cluster Na and the cluster Nb, respectively.

  For example, by registering the center-of-gravity positions Nar and Nbr as attention points in the drive navigation, the drive-navi notifies the driver of the danger when the vehicle enters within 10 m at the approximate radius of the center-of-gravity positions Nar and Nbr. Can do.

  As described above, the information processing apparatus 100 generates clusters according to the so-called hierarchical clustering concept called the longest method. When a cluster is generated using the longest method, it is determined whether or not adjacent clusters are to be combined based on whether or not the distance between the longest elements is within a predetermined value. In the example of FIG. 17, if the shortest method is used, if the shortest distance between elements between the cluster Na and the cluster Nb is shorter than the cutoff distance, the cluster Na and the cluster Nb may be combined. If the cluster Na and the cluster Nb are combined, it becomes one large cluster including the entire intersection, and it becomes difficult to understand information about where near-miss occurs frequently in the intersection.

  On the other hand, since the information processing apparatus 100 according to the first embodiment forms a cluster using the longest method, even if the cluster Na and the cluster Nb are adjacent to each other, if the longest inter-cluster distance is equal to or greater than limit_L2, which is the cutoff distance, the cluster Na and cluster Nb are not bound. In other words, the information processing apparatus 100 can prevent excessive coupling of adjacent clusters by using the longest method. For example, the information processing apparatus 100 can limit the attention points to be pointed out to two on this intersection in the example of FIG.

(Overall processing flow)
The overall processing flow of the information processing apparatus 100 will be described with reference to FIG. FIG. 18 is a diagram illustrating an example of the entire processing flow. The information processing apparatus 100 performs initial processing (step S10). The information processing apparatus 100 executes a data reading process for reading each record stored in the event DB 121 (step S11). The information processing apparatus 100 executes a distance matrix creation process for creating a distance matrix in which the inter-cluster distances between the elements are stored (step S12). The information processing apparatus 100 executes a clustering process for generating a cluster based on the distance matrix (step S13). The information processing apparatus 100 executes a caution point output process for outputting the caution point on the near-miss map to the caution point list 122 (step S14). The information processing apparatus 100 executes end processing (step S15). In the end process, the information processing apparatus 100 displays various counter values at the end of the process on the monitor and performs a process of stopping the processing program.

(Initial processing flow)
With reference to FIG. 19, the flow of the initial process according to step S10 will be described. FIG. 19 is a diagram illustrating an example of the flow of initial processing. The information processing apparatus 100 collects event logs from the drive recorders installed in each vehicle, and stores the collected event logs in the event DB 121 (step S20). The information processing apparatus 100 receives various parameters such as the reference position (zx, zy), the cutoff distance (m) limit_L, and the driving attention reference P_hyr (step S21).

  The information processing apparatus 100 calculates distance conversion coefficients ux and ui for converting the latitude difference and the longitude difference between the two points at the reference position (zx, zy) into the distance between the two points (step S22). The information processing apparatus 100 calculates “limit_L2” by multiplying the cutoff distance (m) limit_L by 2 (step S23), and ends the initial process.

(Data read processing flow)
The flow of the data reading process according to step S11 will be described using FIG. FIG. 20 is a diagram illustrating an example of a flow of data reading processing. The information processing apparatus 100 opens the event DB 121 (step S30). At the same time, nc, which is a variable for the number of events, is initialized and its value is set to 0 (step S31). The information processing apparatus 100 executes a record reading process for reading each record in the event DB 121 (step S32). The information processing apparatus 100 determines whether there is a record that has not been read (step S33). When there is a record that has not been read (step S33 Yes), the information processing apparatus 100 returns to the process of step S32. On the other hand, if there is no record that has not been read (No at Step S33), the information processing apparatus 100 proceeds to the process at Step S34.

  The information processing apparatus 100 closes the event DB 121 (step S34). The information processing apparatus 100 determines whether or not the total number of read records is equal to or greater than the upper limit (step S35). When the total number of records is less than the upper limit (No at Step S35), the information processing apparatus 100 ends the data reading process and proceeds to the process at Step S12. On the other hand, if the total number of records exceeds the upper limit (Yes at Step S35), the information processing apparatus 100 outputs an error (Step S36) and forcibly terminates the process (Step S37).

(Record read processing flow)
The record reading process according to step S31 will be described with reference to FIG. FIG. 21 is a diagram illustrating an example of a flow of record reading processing. The information processing apparatus 100 reads one record from the event DB 121 (step S40). The information processing apparatus 100 determines whether all records have been read from the event DB 121 (step S41). When the information processing apparatus 100 has read all records from the event DB 121 (Yes in step S41), the information reading apparatus 100 ends the record reading process.

  On the other hand, when all the records have not been read from the event DB 121 (step S41 No), the information processing apparatus 100 adds 1 to the event number nc (step S42). The information processing apparatus 100 stores the event log in the memory (step S43). The information processing apparatus 100 adds the element of the cluster number nc to the cluster number list (step S44). The variable nc is incremented by 1 while the data reading process is being performed. The value of nc at this time is used as the ID number assigned to the event point read at that time. That is, the variable nc is the total number of events, and is the ID number of the last point. Here, regarding the relationship between the cluster number and the spot number, for the convenience of creating the program, the cluster number in which the spots are bundled is given a value obtained by sequentially adding 1 from the ID number of the last spot in the order of occurrence. To do. Further, at this stage, it is assumed that the point ID number and the cluster number are not distinguished. That is, when the cluster number is less than or equal to nc, it means that it is an element consisting of a single point, and when the cluster number is greater than nc, it is a cluster formed by combining a plurality of points. It means that. In the following description, on the premise of this, a point that is an element and a number that identifies a cluster that combines the points are collectively referred to as a cluster number.

(Distance matrix creation process flow)
The distance matrix creation process according to step S12 will be described with reference to FIG. FIG. 22 is a diagram illustrating an example of a flow of a distance matrix creation process. The information processing apparatus 100 acquires one combination of two elements from the cluster number list (step S50). The information processing apparatus 100 determines whether all combinations have been acquired from the cluster number list (step S51). If all the combinations have been acquired from the cluster number list (step S51 Yes), the information processing apparatus 100 ends the distance matrix creation process. Here, in actual program implementation, creation of a distance matrix can be realized by the following. That is, in the process of “obtaining one combination of two elements from the cluster number list” (step S50 and step S51), a loop for extracting key elements from the associative array is duplicated. Furthermore, in the inner loop, only the combinations satisfying the magnitude relationship of p <q are processed as extracted elements (p, q).

  On the other hand, when there is a combination that has not been acquired from the cluster number list (step S51 No), the information processing apparatus 100 calculates the inter-cluster distance L between the two elements acquired from the cluster number list (step S52). The information processing apparatus 100 determines whether or not the calculated intercluster distance L is less than “limit_L2” (step S53). When the inter-cluster distance L is equal to or greater than “limit_L2” (No at Step S53), the information processing apparatus 100 returns to the process at Step S50. On the other hand, when the inter-cluster distance L is less than “limit_L2” (Yes in step S53), the information processing apparatus 100 registers the inter-cluster distance L in the distance matrix (step S54), and returns to the process of step S50.

(Flow of clustering process)
The clustering process according to step S13 will be described with reference to FIG. FIG. 23 is a diagram illustrating an example of a flow of clustering processing. The generation unit 111 determines whether or not the distance matrix is empty (step S60). If the distance matrix is empty (No at Step S60), the generation unit 111 ends the clustering process. On the other hand, when the element (p, q) constituting the element or the distance L between clusters remains in the distance matrix (Yes in step S60), the generation unit 111 proceeds to the process of step S61.

  In the clustering process flow of the example of FIG. 23, the processes from step S61 to S67 are processes for generating a new cluster N.

  The generation unit 111 acquires the element component (p, q) having the smallest inter-cluster distance L from the distance matrix by acquiring the two elements registered at the head of the distance matrix (step S61). The element component (p, q) represents a set of two elements registered in the distance matrix, a set of elements and clusters, or a set of two clusters. At this time, the relationship of p <q is always established.

  The generation unit 111 deletes the element component (p, q) from the distance matrix (step S62). The generation unit 111 acquires a cluster number (N = N + 1) obtained by adding 1 to N (step S63). Note that the initial value of the cluster number N is the total number of events nc. The generation unit 111 sets a cluster under the new cluster N in the aggregation point table (step S64). For example, the generation unit 111 stores the cluster number of the subordinate cluster of the new cluster N in the left and right pointers (Pair1 [N], Pair2 [N]) of the new cluster N, so that the subordinate cluster of the new cluster N is stored in the aggregation point table. Set. Here, it is assumed that the left and right subordinate clusters are determined so that the relationship of Pair1 [N] <Pair2 [N] is established. That is, Pair1 [N] = p and Pair2 [N] = q.

  The generation unit 111 calculates the event count k [N] by adding the event counts of the subordinate clusters of the new cluster N (step S65). Further, the generation unit 111 calculates the number of incidents H [N] by adding up the number of incidents of the subordinate clusters of the new cluster N (step S65). Further, the generation unit 111 obtains the position of the center of gravity of the new cluster N based on the position of the subordinate cluster of the new cluster N (step S66).

  The generation unit 111 deletes the cluster numbers p and q from the cluster number list (step S67).

  In the clustering processing flow of the example of FIG. 23, the processing from step S68 to S74 is processing for registering the intercluster distance L between the new cluster N and each remaining cluster i in the distance matrix.

  The specifying unit 112 acquires one cluster number i from the cluster number list (step S68). The cluster number i is one cluster number registered in the cluster number list from the top to the last cluster number, and is a number that does not correspond to either p or q constituting the new cluster N. That is, cluster i is a remaining cluster. The identifying unit 112 determines whether or not all the inter-cluster distances L between p and q constituting the acquired new cluster and other elements or clusters have been compared (step S69). When all the inter-cluster distances L are compared (step S69 Yes), the specifying unit 112 adds the cluster N to the cluster number list (step S75), and returns to the process of step S60. On the other hand, when the comparison of the inter-cluster distance L has not ended (No at Step S69), the specifying unit 112 proceeds to the process at Step S70.

  The identifying unit 112 compares the inter-cluster distances of the cluster i and the element components (p, q), and executes a connection check process for checking whether the cluster i is within “limit_L2” in the longest method (step S1). S70). The identifying unit 112 determines whether or not the cluster i is within “limit_L2” based on the check result of the connection check process (step S71). If the cluster i is within “limit_L2” (Yes at Step S71), the specifying unit 112 proceeds to the process at Step S72. On the other hand, when the cluster i is outside “limit_L2” (No at Step S71), the specifying unit 112 proceeds to the process at Step S74.

  The identifying unit 112 executes a process of calculating the intercluster distance L between the remaining cluster i and the new cluster N (step S72). The identifying unit 112 registers the inter-cluster distance L of (i, N) in the distance matrix (step S73).

  If the inter-cluster distance of (i, p) or (p, i) and (i, q) or (q, i) is registered in the distance matrix, the specifying unit 112 sets the element component as a distance. Delete from the matrix (step S74). Then, the specifying unit 112 returns to the process of step S68.

(Connection check process flow)
The connection check process according to step S70 will be described with reference to FIG. FIG. 24 is a diagram illustrating an example of a flow of connection check processing. The identification unit 112 stores the cluster number of the new cluster N and the cluster number of the remaining cluster i in the variables s and t, respectively (step S80). The identifying unit 112 determines whether or not the variable s is larger than the variable t (step S81). When the variable s is larger than the variable t (Yes in Step S81), the specifying unit 112 switches the variable s and the variable t (Step S82). On the other hand, when the variable s is equal to or smaller than the variable t (No at Step S81), the specifying unit 112 proceeds to the process at Step S83.

  The identifying unit 112 determines whether or not the two inter-cluster distances L between the element component (s, t) and the cluster i are registered in the distance matrix (step S83). When the inter-cluster distance L is registered in the distance matrix (Yes in step S83), the specifying unit 112 determines that there is a remaining cluster i in “limit_L2” (step S84). On the other hand, when the inter-cluster distance L is not registered in the distance matrix (No at Step S83), the specifying unit 112 determines that there is a remaining cluster i outside “limit_L2” (Step S85). That is, the specifying unit 112 determines whether or not there is a remaining cluster i in “limit_L2” based on the longest method.

(Inter-cluster distance calculation processing flow)
The inter-cluster distance calculation process according to step S72 will be described with reference to FIG. FIG. 25 is a diagram illustrating an example of the flow of the inter-cluster distance calculation process. The specifying unit 112 stores the cluster numbers of the new cluster N and the remaining cluster i in the variables s and t, respectively (step S90). The identifying unit 112 determines whether or not the variable s is larger than the variable t (step S91). When the variable s is larger than the variable t (step S91 Yes), the specifying unit 112 switches the variable s and the variable t (step S92). On the other hand, when the variable s is equal to or smaller than the variable t (No at Step S91), the specifying unit 112 proceeds to the process at Step S93. The reason why the variable s is replaced with the variable t when the variable s is larger than the variable t is because the specifying unit 112 executes the inter-cluster distance calculation process for the subordinate cluster related to the cluster t. .

  The identifying unit 112 determines whether or not the intercluster distance (s, t) is registered in the distance matrix (step S93). When the inter-cluster distance (s, t) is registered in the distance matrix (Yes in step S93), the specifying unit 112 determines the element component (s, t) value of the registered distance matrix, that is, (s, t). The value of the distance L is returned (step S94).

  On the other hand, when the inter-cluster distance (s, t) is not registered in the distance matrix (No in step S93), the specifying unit 112 uses the left and right pointers Pair1 [t] and Pair2 [t] to search the cluster t from the aggregation point table. The cluster numbers p and q under the above are acquired (step S95). Next, the specifying unit 112 calls the inter-cluster distance calculation process in step S72 by recursive calling, and calculates the inter-cluster distance between (s, p) and (s, q). The identifying unit 112 sets the larger one of the calculated inter-cluster distances as the inter-cluster distance L between s and t (step S96). Thereafter, the specifying unit 112 proceeds to the process of step S94 in which the intercluster distance L between the clusters s and t is returned.

(Caution point output processing flow)
The caution point output process according to step S14 will be described with reference to FIG. FIG. 26 is a diagram illustrating an example of a flow of attention point output processing. The output unit 113 outputs the form header of the attention point list 122 (step S100). For example, the output unit 113 outputs item names such as group number, ID, number / event number, date, time, vehicle number, nearness, longitude, latitude, distance, driving caution point, and comment as a form header.

  The output unit 113 initializes various counters (step S101). For example, the output unit 113 may be the number of events NG, the number of isolated points NG0, the number of caution points NG1, the total number NC1 of event numbers of caution points, the number of points NG2 that may be caution points, and the caution point. Set to 0 for the total number NC2 of events at the point. Further, the output unit 113 sets 0 for each of the number NG3 of points that are not caution points and the total number NC3 of event numbers of points that are not caution points. Note that the information processing apparatus 100 outputs the counter value of each counter to the monitor in the termination process according to step S15. In addition, the description of the operation of each counter in the flow of attention point output processing according to FIG. 26 is omitted.

  The output unit 113 acquires one cluster number N from the cluster number list (step S102). The output unit 113 determines whether or not all cluster numbers N have been acquired from the cluster number list (step S103). When all the cluster numbers N have been acquired from the cluster number list (step S103 Yes), the output unit 113 ends the attention point output process. On the other hand, when there is a cluster number N that has not been acquired from the cluster number list (step S103 No), the output unit 113 proceeds to the process of step S104.

  The output unit 113 determines whether or not the cluster N is an isolated point (step S104). An isolated point is a point related to a cluster including only one element, and is determined by whether or not the number of events k [N] = 1 in cluster N. When the cluster N is an isolated point (Yes at Step S104), the output unit 113 returns to the process at Step S102. On the other hand, when the cluster N is not an isolated point (No at Step S104), the output unit 113 proceeds to the process at Step S105.

  The output unit 113 substitutes the number of events k [N] and the number of incidents H [N] in the cluster N into the equation (1), and calculates the degree of incident P [N] (step S105). The output unit 113 determines whether or not the calculated nearness degree P [N] is equal to or greater than the driving attention criterion P_hyr and the near miss number H [N] is greater than 1 (step S106). When this condition is satisfied (Yes in Step S106), the output unit 113 determines that the center of gravity position of the cluster N is a caution point (Step S107). The output unit 113 sets the operation caution point of the cluster N to “2”. On the other hand, when this condition is not satisfied (No at Step S106), the output unit 113 proceeds to the process at Step S108.

  The output unit 113 determines whether the near incident number H [N] is one or more (step S108). When the incident number H [N] is 1 or more (Yes in step S108), the output unit 113 determines that the center of gravity of the cluster N is “possibly a caution point” (step S109). The output unit 113 sets the driving attention point of the cluster N to “1”. On the other hand, when the incident number H [N] is 0 (No in step S108), the output unit 113 determines that the center of gravity of the cluster N is not a caution point (step S110). The output unit 113 sets the driving attention point of the cluster N to “0”.

  The output unit 113 outputs information on the cluster N such as the number of events, the degree of nearness, the position of the center of gravity, and driving attention points to the printer (step S111). The output unit 113 outputs the information of the elements under the cluster N to the printer (step S112). Then, the output unit 113 returns to the process of step S102.

(Printer output processing flow)
With reference to FIG. 27, a process of outputting information of elements under the cluster N according to step S112 to the printer will be described. FIG. 27 is a diagram illustrating an example of the flow of printer output processing. The output unit 113 determines whether or not the cluster number N acquired from the cluster number list is greater than the total number of events nc (step S120). If the cluster number N is less than or equal to the total number of events nc (step S120 No), the output unit 113 outputs the cluster N information to the printer (step S122) and ends the printer output process.

  On the other hand, when the cluster number N is larger than the total number of events nc (Yes in step S120), the output unit 113 acquires the cluster numbers (Pair1 [N], Pair2 [N]) under the cluster N (step S121). The output unit 113 replaces each of the acquired Pair1 [N] and Pair2 [N] with the cluster number N, and then recursively calls the printer output processing according to step S112, to the clusters under the cluster N. However, the printer output process according to step S112 is performed. The output unit 113 repeatedly executes recursive calls until the subordinate cluster number of the cluster N that executes the printer output process according to step S112 is equal to or less than the total number of events nc, that is, until the cluster reaches a point.

  In FIG. 26 and FIG. 27, the example in which the processing result by the information processing device 100 is output from the printer has been described. For example, the information on the cluster N to be output to the printer may be displayed on the screen of the computer displaying the map. Alternatively, the information on the cluster N may be transmitted to another computer or a computer mounted on a vehicle via a communication network and displayed on a display connected to the computer.

  As described above, when determining whether to add an element outside the cluster to the cluster, the information processing apparatus 100 first determines the longest inter-cluster distance between each element in the cluster and the element outside the cluster. Let the distance be the intercluster distance L. When the inter-cluster distance L is “limit_L2” or less, the information processing apparatus 100 combines these two clusters and adds them as a new cluster. That is, the information processing apparatus 100 generates a cluster based on the principle of the longest method. Here, since the size of the newly combined cluster is generated using the longest method, it does not exceed “limit_L2”. Accordingly, the information processing apparatus 100 according to the first embodiment can suppress excessive coupling between adjacent clusters while satisfying the upper limit of the size of the generated cluster.

  In addition, if it is determined whether the number of events is k [N] or not, it is more likely that there will be more events such as sudden braking in areas with high traffic volume. A spot is easily determined as a point of caution. On the other hand, the information processing apparatus 100 determines whether or not it is a caution point by using the near degree P [N] obtained by dividing the number of incidents H [N] by the number of events k [N]. It is possible to accurately identify a caution point having a high risk level.

(Effect of Example 1)
The information processing apparatus 100 includes a generation unit 111 that generates a cluster including a plurality of points based on the distance between the points for a point group associated with a specific event. The information processing apparatus 100 includes a specifying unit 112 that specifies the longest distance among distances between a plurality of points included in the generated cluster and a specific point that is not included in the cluster among the point group. . The information processing apparatus 100 includes an output unit 113 that generates and outputs a cluster including a plurality of points included in the cluster and the specific point when the specified longest distance is within a predetermined distance. Thereby, an area (so-called spot) where plots of points on the map are concentrated can be specified with high accuracy.

  In addition, the information processing apparatus 100 includes a generation unit 111 that generates a plurality of clusters each including a plurality of points based on the distance between the points for the point group associated with the specific event. The information processing apparatus 100 includes a plurality of points included in the first cluster among the plurality of generated clusters and a plurality of points included in the second cluster among the plurality of generated clusters. Among the distances, a specifying unit 112 that specifies the longest distance is included. The information processing apparatus 100 generates a cluster including a plurality of points included in the first cluster and a plurality of points included in the second cluster when the specified longest distance is within a predetermined distance. And an output unit 113 for outputting. Thereby, an area (so-called spot) where plots of points on the map are concentrated can be specified with high accuracy.

  The information processing apparatus 100 further determines the cluster based on the number of points related to the first event associated with the same cluster and the number of points associated with the second event among the points. A calculation unit for calculating an evaluation value for. The output unit 113 outputs cluster information and an evaluation value. Thereby, it is possible to carefully select an area to be particularly noted among the specified areas.

  The output unit 113 outputs a map indicating the cluster position because the first event is sudden braking. Thereby, the area where sudden braking frequently occurs can be accurately identified.

  The second event is a near-miss. As a result, it is possible to identify an area with a particularly high degree of risk among the identified areas.

  The information processing apparatus 100 includes a calculation unit that calculates an evaluation value based on a ratio of the number of points associated with the second event to the number of points associated with the first event. Further, the output unit 113 outputs a cluster whose evaluation value is equal to or greater than a predetermined value. For example, even if sudden braking occurs frequently, the sudden braking may not be caused by a near-miss. From the viewpoint of preventing accidents, special attention should be paid to places where sudden braking frequently occurs, and near-miss conditions frequently occur. That is, there are many ratios where it can be estimated that the cause of sudden braking is that a near-miss event has occurred. According to the information processing apparatus 100, the cluster to be output can be selected from the identified areas based on the evaluation value, and thus such an area to be particularly careful can be output.

  The information processing apparatus 100 further determines the distance between one cluster and another cluster or another point, which is a heap configured such that the inter-cluster distance is smaller in the parent node than in the child node. It has the memory | storage part 120 memorize | stored in a matrix. As a result, the speed of searching for the cluster set having the shortest distance can be increased, and the processing speed can be greatly improved.

(Other aspects related to Example 1)
Hereinafter, some of the modifications in the above-described embodiment will be described. Not only the following modifications, but also design changes can be made as appropriate without departing from the spirit of the technology of the present disclosure.

  The information processing apparatus 100 according to the first embodiment has been described with reference to FIG. 2 in which a plurality of drive recorders 10A to 10C are connected to one computer 11 to collect event logs and generate an event DB 121. However, the present invention is not limited thereto. Not. For example, the information processing apparatus 100 may collect event logs from a plurality of drive recorders via the Internet and generate the event DB 121.

  FIG. 28 is a second diagram for explaining collection of event logs. As shown in the example of FIG. 28, the drive recorders 15A to 15C are connected to computers 16A to 16C, respectively. The computers 16 </ b> A to 16 </ b> C are connected to the server 18 via the Internet 17. The server 18 is connected to the printer 19. In the example of FIG. 28, the server 18 is an example of the information processing apparatus 100.

  The computers 16A to 16C read event logs from SD cards included in the drive recorders 15A to 15C, respectively. Each of the computers 16 </ b> A to 16 </ b> C transmits the read event log to the server 18 via the Internet 17. The server 18 generates an event DB 121 based on the transmitted event log. The server 18 generates a caution point list 122 based on the event DB 121 and outputs it to the printer 19.

  Parameters such as the reference position (zx, zy), the cutoff distance (m) limit_L, and the driving attention reference P_hyr can be appropriately changed by the user. The information processing apparatus 100 can adjust the size of the cluster, that is, the size of the spot to be obtained by changing the cutoff distance limit_L. For example, when the stop distance limit_L is 10 m, the information processing apparatus 100 can find a plurality of attention points in the intersection. Further, the information processing apparatus 100 can identify an area with a high traffic risk when the cutoff distance limit_L is 500 m.

  Further, the information processing apparatus 100 can increase or decrease the caution point by changing the driving caution criterion P_hyr. For example, when the driving caution criterion P_hyr is increased, the information processing apparatus 100 carefully selects a caution point to be selected as a caution point having a higher risk level.

  Although it has been described that the information processing apparatus 100 collects event logs from the drive recorder, the present invention is not limited to this. For example, the information processing apparatus 100 may collect event logs from a digital tachograph or a car navigation apparatus.

  In the first embodiment, the information processing apparatus 100 has described that it is determined whether or not there is a near-miss based on the presence or absence of sudden braking and the presence or absence of detection of an object such as a pedestrian or a vehicle, but is not limited thereto. The information processing apparatus 100 may determine whether there is a failure according to the type of the collected event log. For example, when there is an event C1 that is about to collide with a vehicle ahead, an event C2 that a pedestrian or a bicycle has jumped out, or an event C3 that is about to contact a side vehicle, the formula (2) It may be determined whether or not there has been a near miss based on the above.

  C = C1∪C2∪C3 (2)

  The information processing apparatus 100 described above may be mounted on a single computer or may be mounted on a cloud including a plurality of computers. For example, in the information processing apparatus 100, a plurality of computers included in the cloud system exhibit functions similar to those of the generation unit 111, the specification unit 112, and the output unit 113 illustrated in FIG.

  The processing procedure, control procedure, specific name, and information including various data and parameters shown in the first embodiment can be arbitrarily changed unless otherwise specified.

  In the first embodiment, the heap is adopted as the implementation of the distance matrix, but the present invention is not limited to this. This is a data structure that can realize the so-called "priority queue" function, which is the requirement of the data structure of the distance matrix, such as being able to additionally register the distance between clusters, extracting the smallest element, deleting any element, etc. As long as there is no other implementation rejected.

(Hardware configuration of information processing device)
FIG. 29 is a diagram illustrating a hardware configuration of the information processing apparatus 100 according to the first embodiment. As illustrated in the example of FIG. 29, the computer 200 includes a CPU 201 that executes various arithmetic processes, an input device 202 that receives data input from a user, and a monitor 203. The computer 200 also includes a medium reading device 204 that reads programs and the like from a storage medium, an interface device 205 for connecting to other devices, and a wireless communication device 206 for connecting to other devices wirelessly. The computer 200 also includes a RAM 207 that temporarily stores various types of information and a hard disk device 208. Each device 201 to 208 is connected to a bus 209.

  The hard disk device 208 stores, for example, an information processing program having functions similar to the processing units of the generation unit 111, the identification unit 112, and the output unit 113 illustrated in FIG. Furthermore, the hard disk device 208 stores an information processing program having the same functions as the processing units of the generation unit 111, the identification unit 112, and the output unit 113. Also, the hard disk device 208 stores various data for realizing the information processing program.

  The CPU 201 reads out each program stored in the hard disk device 208, develops it in the RAM 207, and executes it to perform various processes. These programs can cause the computer 200 to function as, for example, the generation unit 111, the specification unit 112, and the output unit 113 illustrated in FIG.

  Note that the above information processing program is not necessarily stored in the hard disk device 208. For example, the computer 200 may read and execute a program stored in a storage medium readable by the computer 200. The storage medium readable by the computer 200 corresponds to, for example, a portable recording medium such as a CD-ROM, a DVD disk, a USB (Universal Serial Bus) memory, a semiconductor memory such as a flash memory, a hard disk drive, and the like. Alternatively, the program may be stored in a device connected to a public line, the Internet, a LAN (Local Area Network), etc., and the computer 200 may read and execute the program therefrom.

  FIG. 30 is a diagram illustrating a configuration example of a program operating on a computer. In the computer 200, an OS (operating system) 27 for controlling the hardware group 26 (201 to 208) shown in FIG. The CPU 201 operates in accordance with the procedure according to the OS 27 and controls and manages the hardware group 26, whereby the processing according to the application program 29 and the middleware 28 is executed in the hardware group 26. Further, in the computer 200, the middleware 28 or the application program 29 is read into the RAM 207 and executed by the CPU 201.

  When the information processing function is called by the CPU 201, the function of the control unit 110 is realized by performing processing based on at least a part of the middleware 28 or the application program 29. Each of the information processing functions may be included in the application program 29 itself, or may be a part of the middleware 28 that is executed by being called according to the application program 29.

100 Information processing apparatus 101 Communication unit 110 Control unit 111 Generation unit 112 Identification unit 113 Output unit 120 Storage unit 121 Event DB
122 List of Caution Points

Claims (12)

For a group of points associated with a specific event, generate a cluster containing multiple points based on the distance between the points,
Among the distances between each of the plurality of points included in the generated cluster and a specific point not included in the cluster in the point group, the longest distance is specified,
When the specified longest distance is within a predetermined distance, a cluster including the plurality of points included in the cluster and the specific point is generated and output.
An information processing method characterized by causing a computer to execute processing. For a group of points associated with a specific event, generate multiple clusters each containing multiple points based on the distance between the points,
The distance between each of the plurality of points included in the first cluster of the plurality of generated clusters and each of the plurality of points included in the second cluster of the plurality of generated clusters Identify the longest distance,
When the identified longest distance is within a predetermined distance, a cluster including the plurality of points included in the first cluster and the plurality of points included in the second cluster is generated. Output,
An information processing method characterized by causing a computer to execute processing. Furthermore, based on the number of points related to the first event associated with the same cluster and the number of points associated with the second event among the points, an evaluation value for the cluster The computer executes the process of calculating
The information processing method according to claim 1, wherein the output processing outputs the cluster information and the evaluation value.   The calculating process calculates the evaluation value based on a ratio of the number of points associated with the second event to the number of points associated with the first event. The information processing method according to claim 3, wherein: The first event is a sudden braking;
5. The information processing method according to claim 1, wherein the outputting process outputs a map showing a position of the cluster. 6.   The information processing method according to claim 3, wherein the second event is a near miss.   5. The information processing method according to claim 3, wherein the output process further targets a cluster whose evaluation value is equal to or greater than a predetermined value.   The distance between the one cluster and the other cluster or the other point is stored in a heap configured so that a parent node is smaller than a child node. The information processing method according to any one of 1 to 7. For a point group associated with a specific event, a generation unit that generates a cluster including a plurality of points based on the distance between the points,
A specifying unit for specifying a longest distance among distances between each of the plurality of points included in the generated cluster and a specific point not included in the cluster in the point group;
When the specified longest distance is within a predetermined distance, an output unit that generates and outputs a cluster including the plurality of points included in the cluster and the specific point; and
An information processing apparatus comprising: For a point group associated with a specific event, a generating unit that generates a plurality of clusters each including a plurality of points based on the distance between the points,
The distance between each of the plurality of points included in the first cluster of the plurality of generated clusters and each of the plurality of points included in the second cluster of the plurality of generated clusters Among them, a specific part that specifies the longest distance,
When the identified longest distance is within a predetermined distance, a cluster including the plurality of points included in the first cluster and the plurality of points included in the second cluster is generated. An output section to output,
An information processing apparatus comprising: For a group of points associated with a specific event, generate a cluster containing multiple points based on the distance between the points,
Among the distances between each of the plurality of points included in the generated cluster and a specific point not included in the cluster in the point group, the longest distance is specified,
When the specified longest distance is within a predetermined distance, a cluster including the plurality of points included in the cluster and the specific point is generated and output.
An information processing program for causing a computer to execute processing. For a group of points associated with a specific event, generate multiple clusters each containing multiple points based on the distance between the points,
The distance between each of the plurality of points included in the first cluster of the plurality of generated clusters and each of the plurality of points included in the second cluster of the plurality of generated clusters Identify the longest distance,
When the identified longest distance is within a predetermined distance, a cluster including the plurality of points included in the first cluster and the plurality of points included in the second cluster is generated. Output,
An information processing program for causing a computer to execute processing.

Images