JP2021017080A - Railroad crossing risk determination program - Google Patents

Abstract

To prevent an accident by noticing a risk due to contact with a train inside a railroad crossing in advance.SOLUTION: In a railroad crossing risk determination program for determining a risk at a railroad crossing, a computer executes: an information acquisition step for acquiring voice information inside a railroad crossing for which a risk is newly determined; and a determination step for determining a risk inside the railroad crossing, on the basis of the voice information acquired through the information acquisition step, by referring to association degrees of three or more stages between reference voice information inside the railroad crossing, that is acquired in advance, and risks at the railroad crossing.SELECTED DRAWING: Figure 3

Description

The present invention relates to a railroad crossing risk determination program for determining the risk in a railroad crossing.

Conventionally, accidents caused by contact between moving objects (vehicles, people) and trains at railroad crossings have been a problem. Although the railroad tracks are being elevated, there are still many railroad crossings that pass through many railroad crossings, so it is possible to detect in advance the risk of contact between moving objects and trains inside railroad crossings and prevent accidents. Desired. It is necessary to automatically and accurately perform such a degree of risk.

Japanese Patent Application No. 10-995

In the above-mentioned disclosure technique of Patent Document 1, it is described that the railroad crossing is provided with a learning function of artificial intelligence, but there is no description about determining the degree of danger in the railroad crossing by utilizing artificial intelligence. ..

Therefore, the present invention has been devised in view of the above-mentioned problems, and the purpose of the present invention is to detect in advance the degree of danger due to contact between a moving body in a railroad crossing and a train, and to prevent an accident. In order to prevent this, it is an object of the present invention to provide a railroad crossing risk determination program that automatically determines the risk in a railroad crossing using artificial intelligence.

The railroad crossing risk determination program according to the present invention is a railroad crossing risk determination program for determining the risk of a railroad crossing, and includes an information acquisition step for newly acquiring voice information in the railroad crossing for determining the risk, and an information acquisition step acquired in advance. A discrimination step for determining the risk level in a railroad crossing based on the voice information acquired through the information acquisition step by referring to the reference voice information in the railroad crossing and the degree of association between the level crossing and the risk level of the railroad crossing. It is characterized by having a computer execute and.

In order to detect the degree of danger due to contact with the train inside the railroad crossing in advance and prevent accidents, the degree of danger inside the railroad crossing can be automatically determined using artificial intelligence.

It is a block diagram which shows the whole structure of the railroad crossing risk determination system which implements the railroad crossing risk determination program to which this invention is applied. It is a figure which shows the specific configuration example of the discrimination apparatus. It is a figure for demonstrating operation in a railroad crossing risk determination program to which this invention is applied. It is a figure for demonstrating operation in a railroad crossing risk determination program to which this invention is applied. It is a figure for demonstrating operation in a railroad crossing risk determination program to which this invention is applied. It is a figure for demonstrating operation in a railroad crossing risk determination program to which this invention is applied. It is a figure for demonstrating operation in a railroad crossing risk determination program to which this invention is applied. It is a figure for demonstrating operation in a railroad crossing risk determination program to which this invention is applied. It is a figure for demonstrating operation in a railroad crossing risk determination program to which this invention is applied. It is a figure for demonstrating operation in a railroad crossing risk determination program to which this invention is applied.

Hereinafter, the railroad crossing risk determination program to which the present invention is applied will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing an overall configuration of a railroad crossing risk determination system 1 to which a railroad crossing risk determination program to which the present invention is applied is implemented. The railroad crossing risk determination system 1 includes an information acquisition unit 9, a determination device 2 connected to the information acquisition unit 9, and a database 3 connected to the determination device 2.

The information acquisition unit 9 is a device for a person using this system to input various commands and information, and specifically, is composed of a keyboard, buttons, a touch panel, a mouse, a switch, and the like. The information acquisition unit 9 is not limited to a device for inputting text information, and may be configured by a device such as a microphone that can detect voice and convert it into text information. Further, the information acquisition unit 9 may be configured as an image pickup device capable of capturing an image of a camera or the like. The information acquisition unit 9 may be configured by a scanner having a function of recognizing a character string from a paper-based document. Further, the information acquisition unit 9 may be integrated with the discrimination device 2 described later. The information acquisition unit 9 outputs the detected information to the determination device 2. The information acquisition unit 9 outputs the detected information to the estimation device 2. Further, the information acquisition unit 9 may be configured by means for specifying the position information by scanning the map information. In addition, the information acquisition unit 9 provides video information such as a user terminal, a smartphone, a tablet terminal, a wearable terminal, a digital camera, a video camera, and other mobile terminals that are integrally or partially attached to the worker's head or glasses. It may be composed of a camera or the like capable of acquiring (moving image or still image). An example of a user terminal that is integrally or partially attached to a worker's head or eyeglasses may include a display unit that displays information generated based on image information to be captured in a transparent state. .. The user terminal may be, for example, a holo lens (registered trademark) which is one type of HMD (head mounted display). The user can transparently confirm the display information of the user terminal through the work area and the device to be evaluated through a display unit such as a head-mounted display or a hollow lens that transparently displays the information.

Database 3 accumulates information on the degree of danger in a railroad crossing, such as an accident that occurred before at a railroad crossing, or a case in which an accident did not occur but the degree of risk was high. The database 3 contains reference image information previously captured by a camera that actually constitutes the information acquisition unit 9, reference image information for a train to the railroad crossing at the time of shooting the reference image information, and moving objects that cross the railroad crossing. Reference audio information indicating audio information, reference motion line information extracted from the motion lines of moving objects crossing railroad crossings (including time-series changes in the motion lines), etc. are accumulated. Database 3 contains reference diamond information of the train at the time of shooting the reference image information, reference delay information of the train at the time of shooting the reference image information, and reference audio information in the crossing at the time of shooting the reference image information. Reference opening / closing time zone information of the crossing at the time of shooting the reference image information, reference equipment information of the crossing at the time of shooting the reference image information, and a reference crossing object that traverses the crossing determined based on the reference image information. , Reference motion information that detects the movement of the reference crossing object, reference visibility information to the crossing by shooting an image from each position outside the crossing at the time of shooting the reference image information toward the crossing, etc. are also recorded. Has been done.

The discrimination device 2 is composed of, for example, an electronic device such as a personal computer (PC), but is embodied in any other electronic device such as a mobile phone, a smartphone, a tablet terminal, a wearable terminal, etc., in addition to the PC. It may be converted. The user can discriminate the degree of danger at the railroad crossing by obtaining the search solution by the discriminating device 2.

FIG. 2 shows a specific configuration example of the discrimination device 2. The discrimination device 2 performs wired communication or wireless communication with a control unit 24 for controlling the entire discrimination device 2 and an operation unit 25 for inputting various control commands via operation buttons, a keyboard, or the like. A communication unit 26 for the purpose, a judgment unit 27 for making various judgments, and a storage unit 28 for storing a program for performing a search to be executed represented by a hard disk or the like are connected to the internal bus 21, respectively. .. Further, a display unit 23 as a monitor that actually displays information is connected to the internal bus 21.

The control unit 24 is a so-called central control unit for controlling each component mounted in the discrimination device 2 by transmitting a control signal via the internal bus 21. Further, the control unit 24 transmits various control commands via the internal bus 21 in response to the operation via the operation unit 25.

The operation unit 25 is embodied by a keyboard or a touch panel, and an execution command for executing a program is input from the user. When the execution command is input by the user, the operation unit 25 notifies the control unit 24 of the execution command. Upon receiving this notification, the control unit 24, including the determination unit 27, executes a desired processing operation in cooperation with each component. The operation unit 25 may be embodied as the information acquisition unit 9 described above.

The judgment unit 27 is responsible for various judgments regarding the risk of railroad crossings. The determination unit 27 reads out various information stored in the storage unit 28 and various information stored in the database 3 as necessary information when executing the estimation operation. The determination unit 27 may be controlled by artificial intelligence. This artificial intelligence may be based on any well-known artificial intelligence technique.

The display unit 23 is composed of a graphic controller that creates a display image based on the control by the control unit 24. The display unit 23 is realized by, for example, a liquid crystal display (LCD) or the like.

When the storage unit 28 is composed of a hard disk, predetermined information is written to each address based on the control by the control unit 24, and is read out as needed. Further, the storage unit 28 stores a program for executing the present invention. This program is read by the control unit 24 and executed.

The operation of the railroad crossing risk determination system 1 having the above-described configuration will be described. In the example of FIG. 3, it is assumed that the input data is, for example, reference voice information P01 to P03. The reference voice information as such input data is linked to the output. In this output, the degree of risk in the railroad crossing as an output solution is displayed as a percentage.

The reference audio information is information obtained by recording the audio in the crossing, and is embodied by the loudness level, for example, the frequency information of sound waves and the data obtained by spectrally analyzing the sound in various sound wave bands. You may. Further, if the reference audio information is a moving image, it may be audio data superimposed on the time axis of the moving image.

The reference voice information is associated with each other through three or more levels of association with the risk level as the output solution. Reference voice information is arranged on the left side via this degree of association, and each risk level is arranged on the right side via this degree of association. The degree of association indicates which degree of risk is highly relevant to the reference voice information arranged on the left side. In other words, this degree of association is an index showing what kind of risk each reference voice information is likely to be associated with, and is used to select the most probable risk level from the reference voice information. It shows the accuracy. In the example of FIG. 3, w13 to w19 are shown as the degree of association. These w13 to w19 are shown in 10 stages as shown in Table 1 below, and the closer to 10 points, the higher the degree of relevance of each combination as an intermediate node to the degree of risk as an output. On the contrary, the closer to one point, the lower the degree of each combination as an intermediate node is related to the degree of risk as an output.

The search device 2 acquires in advance the degree of association w13 to w19 of three or more stages shown in FIG. That is, the search device 2 accumulates past data on what kind of risk the reference voice information captured and acquired in the past was in determining the actual search solution, and analyzes and analyzes these. By doing so, the degree of association shown in FIG. 3 is created.

For example, it is assumed that the reference voice information is P01. It is assumed that the risk level for P01 is 60%. By collecting and analyzing such a data set, the degree of association between the reference voice information P01 and the risk level of 60% is strengthened. The data set investigates the voice information actually investigated and the actual risk level for this, and uses each result as a data set. For example, it is assumed that voice information at a certain point in time is represented by a certain waveform with respect to the time axis. At this time, it is extracted from the previous data whether or not an accident has occurred, and whether or not the risk level is such that an accident may occur even if the accident does not occur. These data may be extracted from the past accident data stored at the electric railway company or each station, or the data of the scene where the person was in a hurry. The risk level may be quantified by visually recognizing the reference image information by a plurality of people and collecting questionnaire surveys and the like regarding the risk level.

This analysis and analysis may be performed by artificial intelligence. In such a case, for example, in the case of the reference voice information P01, it is analyzed from the past data whether or not an accident actually occurred, or whether or not the accident did not occur but the risk was high. The more accidents occur, the higher the degree of association that leads to higher-risk output is set, and the fewer accidents occur, the higher the degree of association that leads to lower-risk output. For example, in the example of the reference voice information P01, the output is linked to the output of 90% risk and 60% risk, but since the risk is extremely high from the previous case, w13 leads to 90% risk. The degree of association of w14 is set to 7 points, and the degree of association of w14, which leads to a risk of 60%, is set to 2 points.

Further, the degree of association shown in FIG. 3 may be composed of the nodes of the neural network in artificial intelligence. That is, the weighting coefficient for the output of the node of this neural network corresponds to the above-mentioned degree of association. Further, the network is not limited to a neural network, and may be composed of all decision-making factors constituting artificial intelligence.

Such a degree of association is what is called learned data in artificial intelligence. After creating such trained data, in order to roughly distinguish the article, the solution is searched by using the trained data described above. In such a case, the information acquisition unit 9 measures the voice information at the timing when the solution search for the degree of danger is actually desired.

The risk level is searched based on the newly acquired voice information in this way. In such a case, the degree of association shown in FIG. 3 (Table 1) acquired in advance is used. For example, when the newly acquired voice information is the same as or similar to the reference voice information P02, the risk level is 30% w15 and the risk level 60% is w16 depending on the degree of association. Is associated. In such a case, the risk level of 30%, which has the highest degree of association, is selected as the optimum solution. However, it is not essential to select the one with the highest degree of association as the optimum solution, and it is also possible to select the one with a low degree of association but the degree of risk of 60% that the association itself is recognized as the optimum solution. In addition to this, it goes without saying that an output solution to which the arrows are not connected may be selected, and as long as it is based on the degree of association, it may be selected in any other priority. Further, the risk level to be selected is not limited to one, and two or more may be selected. In such a case, two or more may be selected in order from the highest degree of association, but the present invention is not limited to this, and may be based on the priority of any other degree of association.

In this way, it is possible to search for the optimum degree of risk from the newly acquired voice information and provide it to the user. By looking at the search result, the user can determine the degree of risk of the railroad crossing in the voice information at the present time. In addition to this, the system side can also grasp the degree of risk of railroad crossing from the voice information at the present time.

In the railroad crossing risk determination system 1, as shown in FIG. 4, it is premised that a combination of reference voice information and reference timetable information is formed.

The reference timetable information is information about a train timetable passing through a railroad crossing imaged through reference voice information, in other words, a concept including all information posted on the timetable of the train. In addition to this, the reference timetable information also includes the scheduled passing time for each train to pass the railroad crossing. The scheduled transit time of the train at each railroad crossing may be utilized if the electric railway company suppresses it as data, or may be calculated from the scheduled transit time of each station.

In the example of FIG. 4, for example, P01 to P03, reference timetable information is assumed to be, for example, 8:14, 14:28, ....

As the input data, such reference voice information and reference timetable information are arranged side by side. The intermediate node shown in FIG. 4 is a combination of reference voice information and reference timetable information as such input data. Each intermediate node is further linked to the output. In this output, the degree of risk in the railroad crossing as an output solution is displayed as a percentage.

Each combination (intermediate node) of the reference voice information and the reference timetable information is associated with each other through three or more levels of association with the risk level in the railroad crossing as this output solution. The reference voice information and the reference timetable information are arranged on the left side via this degree of association, and each risk level is arranged on the right side via this degree of association. The degree of association indicates the degree of risk and the degree of relevance to the reference voice information and the reference timetable information arranged on the left side. In other words, this degree of association is an index showing what kind of risk each reference voice information and reference timetable information is likely to be associated with, and is the most from the reference voice information and the reference timetable information. It shows the accuracy in selecting a certain degree of risk. In the example of FIG. 4, w13 to w22 are shown as the degree of association. These w13 to w22 are shown in 10 stages as shown in Table 1 below, and the closer to 10 points, the higher the degree of relevance of each combination as an intermediate node to the degree of risk in the railroad crossing as an output. On the contrary, the closer to one point, the lower the degree of relevance of each combination as an intermediate node to the degree of risk in the railroad crossing as an output.

The discriminating device 2 acquires in advance the degree of association w13 to w22 of three or more stages shown in FIG. That is, the discriminating device 2 accumulates reference voice information, reference timetable information, and data on the degree of risk in that case and analyzes them in determining the actual risk level. , The degree of association shown in FIG. 4 is created by analysis.

For example, it is assumed that the reference voice information is P01. In the reference timetable information at this point, it is assumed that the passing time of the train passing the railroad crossing most recently is 19:21. At this time, it is extracted from the previous data whether or not an accident has occurred and whether or not the risk level is such that an accident may occur even if the accident does not occur. These data may be extracted from the past accident data stored at the electric railway company or each station, or the data of the scene where the person was in a hurry. The risk level may be quantified by visually recognizing the above-mentioned reference voice information by a plurality of people and collecting questionnaire surveys and the like regarding the risk level.

This analysis and analysis may be performed by artificial intelligence. In such a case, for example, in the case of the reference voice information P01 and the reference timetable information P15, whether or not an accident actually occurred, or whether or not an accident did not occur but the risk was high in the past. Analyze from the data of. The more accidents occur, the higher the degree of association that leads to higher-risk output is set, and the fewer accidents occur, the higher the degree of association that leads to lower-risk output. For example, in the example of the intermediate node 61a, the output is linked to the output of 90% risk and 30% risk, but since the risk is extremely high from the previous case, the association of w13 leading to 90% risk. The degree is set to 7 points, and the degree of association of w14, which leads to a risk level of 30%, is set to 2 points.

Further, the degree of association shown in FIG. 4 may be composed of the nodes of the neural network in artificial intelligence. That is, the weighting coefficient for the output of the node of this neural network corresponds to the above-mentioned degree of association. Further, the network is not limited to a neural network, and may be composed of all decision-making factors constituting artificial intelligence.

In the example of the degree of association shown in FIG. 4, the node 61b is a node in which the reference voice information P01 is combined with the reference timetable information P13, and the association degree of 60% is w15 and the risk is 0%. The degree of association is w16. The node 61c is a node in which the reference timetable information P14 and P16 are combined with respect to the reference voice information P02, and the association degree of 30% risk is w17 and the association degree of 70% risk is w18. ..

Such a degree of association is what is called learned data in artificial intelligence. After creating such learned data, when actually determining the risk level at a railroad crossing from now on, the risk level will be determined using the above-mentioned learned data. In such a case, voice information is newly acquired and timetable information is acquired.

The timetable information may be acquired by directly acquiring the train operation status data managed by the electric railway company. How many minutes the train will arrive at the actual railroad crossing is based on the current time, train schedule, delay information, climate and weather. From this information, we calculate how many minutes the train will actually arrive at the railroad crossing.

Based on the newly acquired voice information and the train timetable information in this way, the degree of risk at the time when the newly acquired voice information and the timetable information is actually obtained is obtained. In such a case, the degree of association shown in FIG. 4 (Table 1) acquired in advance is referred to. For example, when the newly acquired voice information is the same as or similar to P02, and the timetable information of the train arriving most recently at the time of acquisition is the same as or similar to the reference timetable information P16. Node 61d is associated with the node 61d via the degree of association, and the node 61d is associated with "risk degree 60%" by w19 and "risk degree 70%" by the association degree w20. In such a case, “risk degree 60%” having a higher degree of association is selected as the optimum solution. However, it is not essential to select the one with the highest degree of association as the optimum solution, and it is also possible to select “70% risk” as the optimum solution, which has a low degree of association but the association itself is recognized. In addition to this, it goes without saying that an output solution to which the arrows are not connected may be selected, and as long as it is based on the degree of association, it may be selected in any other priority.

In addition, Table 2 below shows examples of the degrees of association w1 to w12 extending from the input.

The intermediate node 61 may be selected based on the degree of association w1 to w12 extending from this input. That is, the larger the degree of association w1 to w12, the heavier the weighting in the selection of the intermediate node 61 may be. However, the association degrees w1 to w12 may all have the same value, and the weightings in the selection of the intermediate node 61 may all be the same.

As an alternative to this reference timetable information, reference delay information may be used. This reference delay information indicates how many minutes are behind the train schedule. The reference delay information may be informationized as the delay time itself such as a delay of 4 minutes or a delay of 2 minutes, but the delay time itself is not limited to this, and the delay time is calculated for an accurate timetable. The scheduled passage time of the railroad crossing may be informatized. As the scheduled passage time of the railroad crossing for which the delay time is calculated for an accurate timetable, for example, the scheduled passage time itself such as 7:10 or 15:53 may be informed. The delay time may be fetched directly from the database managed by the electric railway company.

After the learned data is created in this way, voice information is newly acquired and train delay information at that time is acquired. Next, the combination having the reference voice information and the reference delay information of the train and the degree of association of the railroad crossing risk with respect to the combination are referred to at three or more levels. Then, based on the acquired voice information and the delay information of the train, it is possible to determine the degree of danger in the railroad crossing.

As an alternative to this reference timetable information, reference approach information may be used. This reference approach information indicates how many minutes later the train is approaching. For example, the time when the train is approaching the railroad crossing is "down train, 30 seconds later", "up train," It shall be "after 45 seconds" or the like. The approach information may be acquired by directly acquiring the train operation status data managed by the electric railway company. Further, the approach information may be acquired based on any one or more of the train timetable information and the train delay information. How many minutes the train will arrive at the actual railroad crossing is based on the current time, train schedule, delay information, climate and weather. From this information, we calculate how many minutes the train will actually arrive at the railroad crossing. Further, the acquisition of approach information may be based on whether or not the train has passed a point 50 m before the railroad crossing. In other words, if the sensor detects that the train has passed 50 m before the railroad crossing, it determines that the train is approaching, and if it does not detect it, it determines that the train has not yet approached. This may be included in the proximity information. The method of acquiring these approach information is the same when acquiring the above-mentioned reference approach information.

After creating the learned data in this way, the voice information is newly acquired and the approach information of the train at that time is acquired. Next, the degree of association between the combination having the reference voice information and the reference approach information of the train and the degree of risk of the railroad crossing with respect to the combination is referred to. Then, based on the acquired voice information and the approach information of the train, it is possible to determine the degree of danger in the railroad crossing.

FIG. 5 shows an example in which a combination of the above-mentioned reference voice information, reference opening / closing time information, and a degree of association of three or more levels of the degree of risk in the railroad crossing with respect to the combination are set.

As the input data, such reference voice information and reference opening / closing time information are arranged side by side. The intermediate node shown in FIG. 5 is a combination of the reference voice information and the reference opening / closing time information as such input data.

The reference opening / closing time information is information indicating the opening / closing time of the railroad crossing. For example, the opening time of the railroad crossing is registered as 15:28, 15:45, 15:59, etc. The closing time is registered as 15:34 to 36.

The discriminating device 2 acquires in advance the degree of association w13 to w22 of three or more stages shown in FIG. That is, in determining the actual risk level, the discrimination device 2 accumulates reference voice information, reference opening / closing time information, and data on the degree of risk in that case, and stores these. By analyzing and analyzing, the degree of association shown in FIG. 5 is created.

In the example of the degree of association shown in FIG. 5, the node 61b is a node in which the reference voice information P01 is combined with the reference reference opening / closing time information P21, and the association degree of 60% is w15 and the risk degree is w15. The degree of association of 0% is w16.

Similarly, when such a degree of association is set, the degree of risk is obtained based on the newly acquired voice information and the opening / closing time information close to the shooting time. The method of acquiring the opening / closing time information may be directly imported from the database managed by the electric railway company.

In determining the degree of risk, the degree of association shown in FIG. 5 acquired in advance is referred to. For example, when the acquired voice information is the same as or similar to the reference voice information P01 and the acquired opening / closing time information is the same as or similar to the reference opening / closing time information P21, the combination is associated with the node 61b. The degree of risk is calculated from the degree of association with the output at the node 61b.

FIG. 6 shows an example in which a combination of the above-mentioned reference voice information and reference equipment information and a degree of association of three or more levels of the degree of risk in the railroad crossing with respect to the combination are set.

As the input data, such reference voice information and reference equipment information are arranged side by side. The intermediate node shown in FIG. 6 is a combination of reference equipment information and reference voice information as such input data.

The reference equipment information is information about railroad crossing equipment, for example, information indicating whether or not there is a barrier, whether or not there is an alarm, its mounting position and type, and the loudness of the alarm sound. It may be represented by a drawing, an image, a character string, or the like.

The discriminating device 2 acquires in advance the degree of association w13 to w22 of three or more stages shown in FIG. That is, the discrimination device 2 accumulates reference voice information, reference equipment information, and data on the degree of risk in that case in determining the actual risk level, and analyzes these. , The degree of association shown in FIG. 6 is created by analysis.

In the example of the degree of association shown in FIG. 6, the node 61b is a node in which the reference equipment information P26 is combined with the reference voice information P01, and the association degree of 60% is w15 and the risk is 0%. The degree of association is w16.

Similarly, when such a degree of association is set, the degree of risk is obtained based on the newly acquired audio information and the equipment information of the railroad crossing to be photographed. This equipment information can be acquired directly from a database or drawing managed by an electric railway company, or the equipment information can be obtained by analyzing the image of the railroad crossing using artificial intelligence. It may be extracted and specified. Further, the equipment information may be manually input by the user, or may be captured by a scanner or the like.

In determining the degree of risk, the degree of association shown in FIG. 6 acquired in advance is referred to. For example, if the acquired voice information is the same or similar to the reference voice information P01 and the acquired equipment information is the same or similar to the reference equipment information P26, the combination is associated with the node 61b and the node 61b. The degree of risk will be calculated from the degree of association with the output in.

FIG. 7 shows an example in which the combination of the above-mentioned reference voice information, the reference crossing object, and the degree of risk in the railroad crossing with respect to the combination are set to three or more levels of association.

As the input data, such reference voice information and reference crossing objects are arranged side by side. The intermediate node shown in FIG. 7 is a combination of reference crossing objects with reference voice information as such input data.

A crossing object specifically identifies a moving object (passerby, vehicle) crossing a railroad crossing. This crossing object may be roughly classified as a passerby, a vehicle, etc., but if it is a vehicle, it may be a specific type (bus, truck, passenger car, camper, construction machine, two-wheeled vehicle, motorcycle, etc.) and even the specific vehicle type. It may be classified. Passersby may be classified at the level of children, adults, the elderly, physically handicapped people, etc., and even adults may be classified at the level of men and women, office workers, housewives, etc. In addition, elderly people may be classified according to whether or not they are wearing a cane, and physically handicapped people may also be classified according to whether or not they are using a wheelchair. The extraction of this crossing object may be obtained from an image captured of a moving object crossing a railroad crossing. The crossing objects reflected in the captured image of the moving object crossing the railroad crossing are extracted and classified using image analysis and well-known deep learning technology. In such a case, the image data of each classification of these crossing objects is trained in advance as learning data, and the reference audio information is analyzed, the crossing object is specified, and this is crossed for reference while comparing with this. Let it be an object.

The discriminating device 2 acquires in advance the degree of association w13 to w22 of three or more stages shown in FIG. 7. That is, the discrimination device 2 accumulates reference voice information, reference crossing objects, and data on the degree of risk in that case, and analyzes them in determining the actual risk level. , The degree of association shown in FIG. 7 is created by analysis.

In the example of the degree of association shown in FIG. 7, the node 61b is a node in which the reference crossing object P30 is combined with the reference voice information P01, and the degree of association of 60% is w15 and the degree of risk is 0%. The degree of association is w16.

Similarly, when such a degree of association is set, the degree of risk is obtained based on the newly acquired voice information and the crossing objects extracted and classified from the images captured of the moving object. In this method of acquiring a crossing object, the crossing object reflected in the image is extracted and classified by using image analysis or a well-known deep learning technique. In such a case, the image data of each classification of these crossing objects is trained in advance as learning data, and the image information is analyzed and the crossing objects are specified while comparing with the image data.

In determining the degree of risk, the degree of association shown in FIG. 7 acquired in advance is referred to. For example, if the acquired audio information is the same or similar to the reference audio information P01 and the acquired traverse object is the same or similar to the reference traverse object P30, the combination is associated with node 61b, which node 61b. The degree of risk will be calculated from the degree of association with the output in.

FIG. 8 shows a combination of the above-mentioned reference voice information, the reference movement information for detecting the movement of the reference crossing object in addition to the reference crossing object, and the degree of risk in the railroad crossing for the combination. An example is shown in which the degree of association is set to a level or higher.

The reference motion information is the detection of the motion of the reference crossing object. The movement of the crossing object for reference may be specified through image analysis, or the movement of the crossing object reflected in the image information may be extracted by using image analysis or a well-known deep learning technique. Good. The movement of the reference crossing object may be classified by extracting the movement speed, the movement direction, the movement form, the interval or speed at which the object stops in the middle, the regularity, the irregularity, the characteristic movement, and the like.

In such a case, as shown in FIG. 8, the degree of association is expressed as a set of combinations of reference voice information, reference crossing object, and reference motion information as nodes 61a to 61e of intermediate nodes as described above. Will be done.

For example, in FIG. 8, in the node 61c, the reference voice information P02 is associated with the association degree w3, the reference crossing object P30 is associated with the association degree w7, and the reference motion information P34 is associated with the association degree w11.

Similarly, when such a degree of association is set, the degree of danger is obtained based on the newly acquired voice information, the crossing object, and the movement information. The method of acquiring the motion information referred to here is to extract from the cross-sectional object of the newly acquired image information, and is the same as the method of extracting the motion information for reference.

In determining the degree of risk, the degree of association shown in FIG. 8 acquired in advance is referred to. For example, the acquired voice information is the same or similar to the reference voice information P02, the acquired crossing object is the same or similar to the reference crossing object P30, and the acquired motion information is the same or similar to the reference motion information P34. In this case, the combination is associated with a node 61c, which has a risk of 30% associated with a degree of association w17 and a risk of 70% with an association degree of w18. As a result of such a degree of association, the degree of risk is obtained based on w17 and w18.

FIG. 9 shows an example in which a combination of the above-mentioned reference voice information and reference visibility information and a degree of association of three or more levels of the degree of risk in the railroad crossing with respect to the combination are set.

As the input data, such reference voice information and reference visibility information are arranged side by side. The intermediate node shown in FIG. 9 is a combination of reference audio information and reference setting visibility information as such input data.

The reference visibility information is information for confirming the visibility to the railroad crossing by taking an image from each position outside the railroad crossing toward the railroad crossing at the time of shooting the reference voice information. This reference visibility information indicates whether or not a vehicle or a passerby crossing a railroad crossing can easily confirm the visibility of a passing train. The reference visibility information may be represented by the captured image itself, or may be displayed by a character string, a drawing, a rank, a symbol, or the like. When visibility is poor due to a building structure, a wall, etc., for example, the numerical value indicating the rank becomes low. Such reference visibility information may be extracted and analyzed from, for example, an image captured from each position outside the railroad crossing toward the railroad crossing through image analysis, a well-known deep learning technique, or the like.

The discriminating device 2 acquires in advance the degree of association w13 to w22 of three or more stages shown in FIG. That is, the discriminating device 2 accumulates reference voice information, reference visibility information, and data on the degree of risk in that case in determining the actual risk level, and stores these. By analyzing and analyzing, the degree of association shown in FIG. 9 is created.

In the example of the degree of association shown in FIG. 9, the node 61b is a node in which the reference voice information P01 is combined with the reference visibility information P35, and the degree of association of 60% is w15 and the degree of risk is 0%. The degree of association is w16.

Similarly, when such a degree of association is set, the degree of risk is obtained based on the newly acquired voice information and its visibility information. The method of acquiring the visibility information is the same as the method of acquiring the reference visibility information.

In determining the degree of risk, the degree of association shown in FIG. 9 acquired in advance is referred to. For example, if the acquired voice information is the same or similar to the reference voice information P01 and the acquired visibility information is the same or similar to the reference visibility information P35, the combination is associated with the node 61b. The degree of risk is calculated from the degree of association with the output at the node 61b.

In FIG. 10, in addition to the above-mentioned reference voice information and reference timetable information, the combination of the reference equipment information and the degree of risk in the railroad crossing with respect to the combination are set to three or more levels of association. An example is shown.

In such a case, as shown in FIG. 10, the degree of association is expressed as a set of combinations of reference voice information, reference timetable information, and reference equipment information as nodes 61a to 61e of intermediate nodes as described above. Will be done.

Similarly, when such a degree of association is set, the newly acquired voice information, the timetable information, and the equipment information thereof can be acquired, and the degree of risk can be obtained by referring to the above-mentioned degree of association. ..

In the above-mentioned degree of association, the case where the degree of association is formed for the combination of the reference voice information, the reference timetable information, and the reference equipment information has been described as an example, but the description is limited to this. It is not something that is done. Any combination of two or more of voice information for reference, timetable information for reference, delay information for reference, opening / closing time zone information for reference, equipment information for reference, crossing object for reference, motion information for reference, visibility information for reference, etc. However, the above-mentioned degree of association may be associated with the above.

In the above-mentioned degree of association, the degree of association is expressed by a 10-step evaluation, but it is not limited to this, and it may be expressed by a degree of association of 3 or more levels, and conversely, it may be expressed by 3 or more levels. For example, 100 steps or 1000 steps may be used. On the other hand, this degree of association does not include those expressed in two stages, that is, whether or not they are related to each other, either 1 or 0.

The degree of risk has been described by taking the case where it is described as a percentage of 0 to 100% as an example, but the risk is not limited to this. This degree of risk may be expressed in two stages, for example, "high risk" and "low risk". In such a case, when forming the degree of association, when analyzing the case of determining whether or not the risk is high, the combination of the reference voice information and other reference information is "high risk" or "dangerous". It will be analyzed and discriminated in association with the result of "low sex". Of course, these tasks may also be replaced by artificial intelligence.

According to the present invention having the above-described configuration, it is possible to easily determine the degree of risk at a railroad crossing with little effort without requiring special skill. Further, according to the present invention, it is possible to determine the degree of danger at a railroad crossing with higher accuracy than that performed by a human being. Further, by configuring the above-mentioned degree of association with artificial intelligence (neural network or the like), it is possible to further improve the discrimination accuracy by learning this.

Further, according to the present invention, it is characterized in that the optimum physical properties and the generation mechanism are searched for through the degree of association set in three or more stages. The degree of association can be described by, for example, a numerical value from 0 to 100% in addition to the above-mentioned 5 stages, but is not limited to this, and any stage can be described as long as it can be described by a numerical value of 3 or more stages. It may be configured.

By searching for the most probable risk level based on the degree of association represented by the numerical values of three or more levels, the association is considered as a candidate for the possibility of increasing the risk on the railroad crossing. It is also possible to search and display in descending order of degree. If the user can be displayed in descending order of the degree of association in this way, it is possible to preferentially display the more probable risk degree, and it is possible to call attention to the higher degree of risk.

In particular, when the degree of danger is higher, processing operations such as calling attention to moving objects by voice etc. or notifying the train entering the railroad crossing that the danger is high and urging the train to stop voluntarily are preceded. It can be done and the safety of passengers can be protected. Further, according to the present invention, since the accuracy of detecting the degree of danger is high, it is sufficient to stop the train only in the cases where it is really necessary without stopping the train unnecessarily, which adversely affects smooth traffic. It can also be prevented. Further, it may be output as a search solution whether or not to transmit a stop signal to the train in association with the degree of danger.

In addition to this, according to the present invention, it is possible to judge without overlooking the discrimination result of the output having an extremely low degree of association such as 1%. Remind the user that even a judgment result with an extremely low degree of association is connected as a slight sign, and may be useful as the judgment result once every tens or hundreds of times. be able to.

Further, according to the present invention, there is an advantage that the search policy can be determined by the method of setting the threshold value by performing the search based on the degree of association of three or more stages. If the threshold value is lowered, even if the above-mentioned degree of association is 1%, it can be picked up without omission, but it is unlikely that a more appropriate discrimination result can be detected favorably, and a lot of noise may be picked up. is there. On the other hand, if the threshold value is raised, there is a high possibility that the optimum risk level can be detected with a high probability. Sometimes the solution is overlooked. It is possible to decide which one to prioritize based on the ideas of the user side and the system side, but it is possible to increase the degree of freedom in selecting the points to be emphasized.

Further, in the present invention, the above-mentioned degree of association may be updated. This update may reflect information provided via a public communication network such as the Internet. In addition, when new knowledge is discovered about the relationship between the input parameter and the output solution (risk level) based on the camera image independently captured by the electric railway company or the station or various reference information acquired independently. Increases or decreases the degree of association according to the findings.

In other words, this update corresponds to learning in the sense of artificial intelligence. It can be said that it is a learning act because it acquires new data and reflects it in the learned data.

Further, the process of first creating the trained model and the above-mentioned update may use not only supervised learning but also unsupervised learning, deep learning, reinforcement learning, and the like. In the case of unsupervised learning, instead of reading and learning the data set of input data and output data, information corresponding to the input data is read and trained, and the degree of association related to the output data is self-formed from there. You may let it.

This update of the degree of association is artificially performed by the system side or the user side based on the contents of research data, treatises, conference presentations, newspaper articles, books, etc. by experts, except when it is based on information that can be obtained from the public communication network. It may be updated automatically or automatically. Artificial intelligence may be utilized in these update processes.

Further, the degree of association of each combination described above is the degree of association of a combination having one factor and another factor, and it goes without saying that other elements other than these may be associated with the degree of association.

1 Railroad crossing risk discrimination system 2 Discrimination device 21 Internal bus 23 Display unit 24 Control unit 25 Operation unit 26 Communication unit 27 Estimate unit 28 Storage unit 61 Node

Claims (8)

In the railroad crossing risk determination program for determining the risk of railroad crossings
An information acquisition step to acquire voice information in a railroad crossing that newly determines the degree of danger,
The risk level in the railroad crossing is determined based on the voice information acquired through the information acquisition step by referring to the reference voice information in the railroad crossing acquired in advance and the degree of association between the railroad crossing risk level and three or more levels. A railroad crossing risk determination program characterized by having a computer perform the determination steps to be performed. In the above information acquisition step, the train timetable information is acquired and
In the determination step, the combination having the reference voice information, the reference timetable information of the train acquired in advance, and the degree of association of the railroad crossing risk with respect to the combination are referred to, and the information acquisition step is performed. The railroad crossing risk determination program according to claim 1, wherein the risk level in the railroad crossing is determined based on the voice information acquired through the above-mentioned railroad crossing information and the train schedule information. In the above information acquisition step, train delay information is acquired and
In the determination step, the combination having the reference voice information, the reference delay information of the train acquired in advance, and the degree of association of the railroad crossing risk with respect to the combination are referred to, and the information acquisition step is performed. The railroad crossing risk determination program according to claim 1, wherein the risk level in the railroad crossing is determined based on the voice information acquired through the railroad crossing and the delay information of the train. In the above information acquisition step, the opening / closing time zone information of the railroad crossing is acquired.
In the determination step, the combination having the reference audio information, the reference opening / closing time zone information of the railroad crossing acquired in advance, and the degree of association of the railroad crossing risk with respect to the combination are referred to, and the above information is referred to. The railroad crossing risk determination program according to claim 1, wherein the risk level in the railroad crossing is determined based on the voice information acquired through the acquisition step and the opening / closing time zone information of the railroad crossing. In the above information acquisition step, the equipment information of the railroad crossing is acquired.
In the determination step, the combination of the reference audio information, the reference equipment information of the railroad crossing acquired in advance, and the degree of association of the railroad crossing risk with respect to the combination are referred to, and the information acquisition step is performed. The railroad crossing risk determination program according to claim 1, wherein the risk level in the railroad crossing is determined based on the voice information acquired through the railroad crossing and the equipment information of the railroad crossing. In the above information acquisition step, an information acquisition step of acquiring image information by newly taking an image in the railroad crossing and acquiring a crossing object of a moving body traversing the railroad crossing determined based on the image information.
In the determination step, there are three or more stages of a combination having the reference audio information, a reference crossing object that crosses the railroad crossing determined based on the image information acquired in advance, and the risk level of the railroad crossing for the combination. The railroad crossing risk determination program according to claim 1, wherein the level crossing is determined based on the voice information acquired through the information acquisition step and the crossing object with reference to the degree of association. In the above information acquisition step, the motion information obtained by detecting the motion of the crossing object is further acquired.
In the determination step, the information acquisition step is further referred to by three or more levels of association between the combination having the reference movement information that detects the movement of the reference crossing object and the risk level of the railroad crossing with respect to the combination. The railroad crossing risk determination program according to claim 6, wherein the risk level in the railroad crossing is determined based on the voice information, the crossing object, and the movement information acquired through the above. In the above information acquisition step, visibility information to the railroad crossing is acquired by taking an image from each position outside the railroad crossing toward the railroad crossing.
In the determination step, a combination having the reference audio information and a reference visibility to the railroad crossing by taking an image from each position outside the railroad crossing acquired in advance toward the railroad crossing, and a risk of the railroad crossing for the combination. A claim characterized in that the degree of risk in the railroad crossing is determined based on the audio information acquired through the above information acquisition step and the visibility information to the railroad crossing by referring to the degree of association with the degree in three or more stages. Item 1. The railroad crossing risk determination program according to item 1.

Images