Traffic congestion prediction system, terminal device, traffic congestion prediction method, and traffic congestion prediction program

Abstract

The ability to predict traffic congestion at any location by collecting images of moving objects on roads. [Solution] The traffic congestion prediction device 100 collects images of moving objects on the road and location information indicating the location where the images were taken, using multiple terminal devices 110. The image processing unit 101 processes the collected images and extracts traffic information, which is information about moving objects present in the images. The traffic information processing unit 103 uses the traffic information extracted by the image processing unit 101 to calculate information regarding traffic congestion prediction on the road corresponding to the location where the images were taken, generates traffic congestion prediction information, and transmits it to the terminal devices 110.

Description

Technical Field

[0006] , ,

[0005] , , ,

[0001] The present invention relates to a traffic jam prediction system, a terminal device, a traffic jam prediction method, and a traffic jam prediction program that predict traffic jams using images captured from a moving object. However, the use of the present invention is not limited to the traffic jam prediction system, the terminal device, the traffic jam prediction method, and the traffic jam prediction program.

Background Art

[0002] Conventionally, as a technique for predicting traffic jams on roads, a server that determines the traffic jam situation at a vehicle position based on the speed from the vehicle and the average speed of the vehicle position has been disclosed (see, for example, Patent Document 1 below). This technique also describes using an image captured by a vehicle for image analysis to determine traffic jams.

[0003] In addition, there is a system that determines traffic jams using an image captured by a camera fixedly installed on a road.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, in the above conventional technology, it was not possible to predict traffic jams at arbitrary locations. Further, in the configuration for determining traffic jams using images, there is a problem that the installation cost is high and there is a limitation that traffic jams can only be determined at the location where the camera is fixedly installed. <00000​​Furthermore, the system determines whether the section of road the vehicle is traveling on is congested based on information such as speed received from the vehicle, and has the problem that it cannot predict future congestion. The problems that this invention aims to solve include the above-mentioned problems as one example. It can be listed as such. [Means for solving the problem]

[0007] To solve the above-mentioned problems and achieve the objective, the traffic congestion prediction system according to claim 1 is characterized by comprising: an acquisition unit that acquires an image of a moving object on a road and location information indicating the location where the image was taken; an image processing unit that extracts traffic information, which is information relating to the moving object present in the image; and a traffic information processing unit that uses the traffic information extracted by the image processing unit to calculate information relating to traffic congestion prediction on the road corresponding to the location where the image was taken.

[0008] Furthermore, the terminal device according to claim 6 is a terminal device that transmits an image to a traffic congestion prediction device that predicts traffic congestion of moving objects on a road based on an image of a moving object on the road, and is characterized by comprising: an imaging unit that captures an image of a moving object on the road; and a processing unit that adds captured location information to the image captured by the imaging unit and transmits it to the traffic congestion prediction device.

[0009] Furthermore, the terminal device according to claim 7 is a terminal device connected to a traffic congestion prediction device that predicts traffic congestion of moving objects on a road based on images of moving objects on the road, and has a processing unit that requests the traffic congestion prediction device to provide a traffic congestion prediction including a desired location and time, and outputs information regarding the traffic congestion prediction received from the traffic congestion prediction device. To use as a sign.

[0010] Furthermore, the traffic congestion prediction method according to claim 8 is characterized in that, in a traffic congestion prediction method implemented by a traffic congestion prediction device, it includes an acquisition step of acquiring an image of a moving object on a road and location information indicating the location where the image was taken; an image processing step of extracting traffic information, which is information relating to a moving object present in the image; and a traffic information processing step of calculating information relating to traffic congestion prediction on the road corresponding to the location where the image was taken, using the traffic information extracted by the image processing step.

[0011] Furthermore, the traffic congestion prediction program according to claim 9 is characterized by causing a computer to execute the traffic congestion prediction method described above. [Brief explanation of the drawing]

[0012] [Figure 1] Figure 1 is a block diagram showing an example configuration of a traffic congestion prediction system according to an embodiment. [Figure 2] Figure 2 is a flowchart showing an example of the processing procedure of the traffic congestion prediction system according to the embodiment. [Figure 3] Figure 3 is a block diagram showing an example of the hardware configuration of a navigation system. [Figure 4] Figure 4 is a block diagram showing an example of a server hardware configuration. [Figure 5] Figure 5 is a flowchart showing the processing steps for image transmission by the navigation device. [Figure 6] Figure 6 is a flowchart showing the process of generating traffic information by the server. [Figure 7] Figure 7 is a diagram illustrating an example of the content of traffic information. [Figure 8] Figure 8 is a flowchart showing the processing steps for traffic congestion prediction using traffic information. [Figure 9] Figure 9 is a flowchart showing the processing steps from the request for traffic congestion prediction to the transmission of traffic congestion prediction information. [Figure 10] Figure 10 is an explanatory diagram illustrating the conditions under which natural traffic congestion occurs.

Best Mode for Carrying Out the Invention

[0013] (Embodiment) Hereinafter, with reference to the accompanying drawings, preferred embodiments of the traffic congestion prediction system, terminal device, traffic congestion prediction method, and traffic congestion prediction program according to this invention will be described in detail.

[0014] FIG. 1 is a block diagram showing a configuration example of the traffic congestion prediction system according to the embodiment. The traffic congestion prediction system according to the embodiment can be configured by a traffic congestion prediction device 100 and a terminal device 110 that are communicatively connected to each other.

[0015] The traffic congestion prediction device 100 includes an image processing unit 101, an information combining unit 102, a traffic information processing unit 103, a storage unit 104, and a traffic congestion prediction information management unit 105. The terminal device 110 includes an imaging unit 111, a positioning unit 112, and a processing unit 113.

[0016] The traffic congestion prediction device 100 can be configured using, for example, a single server, but is not limited thereto, and may be configured using a plurality of servers distributed according to the functions of the image processing unit 101 and the traffic information processing unit 103.

[0017] The terminal device 110 captures an image of the forward view during movement by the imaging unit 111 and transmits it to the traffic congestion prediction device 100. At this time, the processing unit 113 transmits the captured ID , and the position information at the time of shooting measured by the positioning unit 112 together with the captured image of the imaging unit 111. In addition, the speed and time information during movement may be transmitted. Further, the processing unit 113 requests the traffic congestion prediction device 100 for traffic congestion prediction by a user operation or the like, and processes the traffic congestion prediction information transmitted from the traffic congestion prediction device 100 and outputs it to a display unit (not shown) or the like.

[0018] The terminal device 110 can be configured using a navigation device installed on a mobile device, a smartphone or mobile phone owned by the user, a portable personal computer, etc., to transmit images used for traffic congestion prediction to the traffic congestion prediction device 100, and to request traffic congestion predictions from the traffic congestion prediction device 100.

[0019] The terminal device 110 is not limited to these configurations; it may also be configured using a personal computer or the like, permanently installed in a home or facility (store). In this case, the terminal device 110 omits the function of transmitting images used for traffic congestion prediction to the traffic congestion prediction device 100, and instead requests traffic congestion predictions from the traffic congestion prediction device 100. In such a terminal device 110, the functions of the imaging unit 111 and the positioning unit 112 can be omitted, and the processing unit 113 of the terminal device 110 requests traffic congestion predictions through user operation, etc., and outputs the traffic congestion prediction information transmitted from the traffic congestion prediction device 100 to a display unit or the like.

[0020] The image processing unit 101 of the traffic congestion prediction device 100 processes the captured images transmitted from the terminal device 110. This image processing extracts (detects) the roads and vehicles shown in the captured images. It then detects the lane in which the terminal device 110 (mobile unit) is traveling (its own lane), the number of lanes, the number of vehicles in each lane, and the distance between vehicles. In addition, it can also detect the type of vehicle included in the captured image.

[0021] The information linking unit 102 associates the shooting ID, shooting time, and shooting location information of the captured image transmitted from the terminal device 110 and stores them in the storage unit 104. When using the time from the traffic congestion prediction device 100, the time at which the captured image was received can be used, and it is not necessary for the terminal device 110 to transmit time information. Furthermore, if the traffic congestion prediction device 100 knows the speed of the moving object to which the terminal device 110 is located, the traffic congestion prediction device 100 can determine the speed of the moving object by, for example, calculating based on the amount of latitude and longitude movement of the terminal device 110 and the time required for movement.

[0022] The information combining unit 102 then associates (combines) the lane information, which is the result of image processing by the image processing unit 101, with the same shooting ID and stores it in the storage unit 104 as traffic information. The information combining unit 102 also stores the vehicle information (number of vehicles) for each lane for each shooting ID in the storage unit 104 as lane-specific vehicle information. The traffic information is stored in the storage unit 104 in a state classified by road, location, and time of day.

[0023] The traffic information processing unit 103 uses a statistical function to calculate the probability of congestion occurring at each location and time of day for each piece of traffic information classified by road location and time of day stored in the storage unit 104, and stores this as congestion prediction information in the storage unit 104. The storage unit 104 stores traffic information based on changes in multiple images taken on the same road, and the traffic information processing unit 103 predicts the occurrence of congestion by comparing the traffic information requested from the terminal device 110 with past traffic information (for example, at the same time and location).

[0024] Traffic information is not limited to information about the same road; it can also include information about other roads within a specified range.

[0025] In this process, since traffic information for each lane is obtained based on captured images, the degree of congestion can be determined based on the number of vehicles in each lane (corresponding to the distance between vehicles and vehicle density), and congestion predictions can be made for each lane.

[0026] Traffic congestion forecast information includes future traffic congestion forecasts near the terminal device 110's current location, traffic congestion forecasts along the route from the terminal device 110 to its destination, and traffic congestion forecasts near the destination. This traffic congestion forecast information may be calculated in advance for each road (location), or it may be calculated for the relevant road (location) when requested by the terminal device 110.

[0027] Furthermore, traffic congestion prediction information is not limited to predicting congestion on the road corresponding to the location where the photo was taken. For example, it may also predict congestion on the road closest to the location where the photo was taken, or on the map information that has been map-matched at the time the photo was taken.

[0028] The traffic congestion forecast information management unit 105 manages traffic congestion forecast information. This traffic congestion forecast information management unit 105 transmits location and area traffic congestion forecast information to the terminal device 110 in response to traffic congestion forecast requests from the terminal device 110.

[0029] The terminal device 110 requests, for example, information on the travel route of the mobile vehicle and congestion prediction information within a predetermined area near the destination from the congestion prediction device 100. Similarly, when the mobile vehicle searches for a route, it also requests information on the travel route and congestion prediction information within a predetermined area near the destination. In response, the congestion prediction device 100 can extract information on the travel route of the terminal device 110 (mobile vehicle) and congestion prediction information within a predetermined area near the destination and transmit it to the terminal device 110.

[0030] Figure 2 is a flowchart showing an example of the processing procedure of the traffic congestion prediction system according to the embodiment. It shows the processing content of traffic congestion prediction performed by the traffic congestion prediction device 100. First, the traffic congestion prediction device 100 acquires the captured image, location information at the time of capture, speed, and time from the terminal device 110 (step S201). As mentioned above, speed and time can be acquired without receiving them from the terminal device 110.

[0031] Next, the traffic congestion prediction device 100 processes the captured images (step S202). This image processing detects the lane the terminal device 110 (mobile vehicle) is traveling in (its own lane), the number of lanes, the number of vehicles in each lane, the type of vehicle, and the distance between vehicles. The detected information is associated (combined) with the information acquired in step S201 (location information, speed, time) to generate traffic information, which is then stored in the storage unit 104.

[0032] Multiple identical or different terminal devices 110 transmit multiple captured images to the traffic congestion prediction device 100 for each different location and time. As a result, the traffic congestion prediction device 100 repeatedly executes steps S201 and S202, and stores a large amount of traffic information in a database categorized by road location and time. The traffic congestion prediction device 100 uses this traffic information to perform the following traffic congestion prediction processing.

[0033] Next, the congestion prediction device 100 waits to receive a congestion prediction request from the terminal device 110 (loop of step S203: No). When a congestion prediction request is received (step S203: Yes), congestion prediction processing is performed (step S204). In this congestion prediction processing, the traffic information generated in step S202 is referred to, and based on past congestion occurrences at the same location (range) and time as included in the request, congestion is predicted to occur at this location (range) and time.

[0034] Furthermore, the location where congestion is predicted does not need to be exactly the same as the location included in the request. It may be predicted based on past traffic information and congestion occurrences within the same road or a specified range as the location included in the request. The same applies to the time, including the time included in the request. It is also acceptable to use the designated time slot.

[0035] After this, the traffic congestion forecast information is transmitted to the terminal device 110 (step S205), and the series of traffic congestion forecasting processes is completed.

[0036] According to the above embodiment, since traffic congestion is predicted using images taken at multiple locations and times on the road by a camera on the terminal device, it is possible to predict traffic congestion not only at locations where fixed devices such as cameras and sensors are installed on the road are used, but also at any location on the road.

[0037] Furthermore, by obtaining a large amount of traffic information from images transmitted from numerous terminal devices, it becomes possible to improve the accuracy of traffic congestion predictions at any given location.

[0038] Since the terminal device can be a general-purpose device such as a mobile phone, smartphone, or car navigation system equipped with a camera that can take images and transmit location information, the traffic congestion prediction system can easily acquire a large number of images and reduce system costs.

[0039] Furthermore, while the system uses vehicle-to-vehicle spacing and vehicle count to generate traffic congestion predictions, other data that could contribute to congestion, such as weather, day of the week, traffic information on other roads connected to the requested location, and the presence or absence of events near the requested location, may also be added as parameters to the congestion prediction. This would allow for further improvement in the accuracy of the congestion predictions.

[0040] While congestion was initially judged based on the number of vehicles (distance between vehicles), congestion can also be predicted similarly by comparing images by processing the similarity between an image of a pre-congestion image of any location on the road and a current image transmitted from a terminal device.

[0041] Furthermore, the images captured by the terminal device are not limited to still images; they can also be videos. Videos allow for the determination of the speed of moving objects in other lanes, thus providing more detailed traffic information.

[0042] Furthermore, statistical methods such as multiple regression analysis may be applied to generate traffic congestion prediction data. In this case, the optimal algorithm can be customized for each road. [Examples]

[0043] The following describes an embodiment of the present invention. In this embodiment, a navigation device 300 is mounted on a mobile vehicle, and each user's navigation device 300 accesses the server, which acts as the traffic congestion prediction device 100 described above. Here, the navigation device 300 has the function of the terminal device 110 described above, and photographs the road as the mobile vehicle moves and transmits it to the traffic congestion prediction device 100. The navigation device 300 also requests traffic congestion prediction information for a desired location from the traffic congestion prediction device 100 (server).

[0044] (Hardware configuration of navigation device 300) Figure 3 is a block diagram showing an example of the hardware configuration of a navigation device. In Figure 3, the navigation device 300 includes a CPU 301, ROM 302, RAM 303, magnetic disk drive 304, magnetic disk 305, optical disk drive 306, optical disk 307, audio interface 308, microphone 309, speaker 310, input device 311, video interface 312, display 313, communication interface 314, GPS unit 315, various sensors 316, and camera 317. Each component 301 to 317 is connected by a bus 320.

[0045] The CPU 301 controls the entire navigation system 300. The ROM 302 stores the boot program, traffic congestion prediction program, etc. The RAM 303 is used as the work area for the CPU 301. In other words, the CPU 301 controls the entire navigation system 300 by executing various programs stored in the ROM 302 while using the RAM 303 as its work area.

[0046] The magnetic disk drive 304 controls the reading and writing of data to the magnetic disk 305 according to the control of the CPU 301. The magnetic disk 305 records the data written under the control of the magnetic disk drive 304. For example, the magnetic disk 305 can be an HD (hard disk) or an FD (flexible disk).

[0047] Furthermore, the optical disc drive 306 controls the reading and writing of data to the optical disc 307 according to the control of the CPU 301. The optical disc 307 is a removable recording medium from which data is read according to the control of the optical disc drive 306. The optical disc 307 can also use a writable recording medium. In addition to the optical disc 307, MO disks, memory cards, etc. can be used as removable recording media.

[0048] Examples of information recorded on the magnetic disk 305 and optical disk 307 include map data, vehicle information, road information, and driving history. Map data is used in car navigation systems when searching for routes and is vector data that includes background data representing features such as buildings, rivers, ground surfaces, and energy supply facilities, and road shape data representing the shape of roads with links and nodes.

[0049] The audio interface 308 is connected to a microphone 309 for audio input and a speaker 310 for audio output. The audio received by the microphone 309 is converted from analog to digital within the audio interface 308. The microphone 309 can be installed, for example, on the dashboard of a vehicle, and there may be one or more of them. The speaker 310 outputs audio that has been converted from a predetermined audio signal within the audio interface 308.

[0050] The input device 311 may include a remote control, keyboard, or touch panel equipped with multiple keys for inputting characters, numbers, and various instructions. The input device 311 may be implemented in any one form of a remote control, keyboard, or touch panel, but it can also be implemented in multiple forms.

[0051] The video interface 312 is connected to the display 313. Specifically, the video interface 312 consists of, for example, a graphics controller that controls the entire display 313, a buffer memory such as VRAM (Video RAM) that temporarily stores image information that can be displayed immediately, and a control IC that controls the display 313 based on the image data output from the graphics controller.

[0052] The display 313 displays various data such as icons, cursors, menus, windows, text, and images. For example, the display 313 can be a TFT liquid crystal display or an organic EL display.

[0053] Camera 317 captures images of the road outside the vehicle. The images can be either still images or videos. By capturing images of the road outside the vehicle with camera 317, other vehicles traveling on the road are also captured. After image processing by CPU 301, these images, along with the location information of GPS unit 315, are sent to the traffic congestion prediction device 100 (server 4, described later). It will be sent to 00).

[0054] The communication interface 314 is connected to the network wirelessly and functions as an interface for the navigation device 300 and the CPU 301. Communication networks that function as networks include in-vehicle communication networks such as CAN and LIN (Local Interconnect Network), as well as public telephone networks, mobile phone networks, DSRC (Dedicated Short Range Communication), LAN, and WAN. Examples of communication interfaces 314 include public telephone network connection modules, ETC (Electronic Toll Collection) units, FM tuners, and VICS (Vehicle Information and Communication System: registered trademark) / beacon receivers.

[0055] The GPS unit 315 receives radio waves from GPS satellites and outputs information indicating the vehicle's current position. The output information from the GPS unit 315, along with the output values ​​from the various sensors 316 described later, is used by the CPU 301 to calculate the vehicle's current position. The information indicating the current position is, for example, information that identifies a specific point on map data, such as latitude, longitude, and altitude.

[0056] The various sensors 316, such as a vehicle speed sensor, acceleration sensor, angular velocity sensor, and tilt sensor, output information to determine the vehicle's position and behavior. The output values ​​of the various sensors 316 are used by the CPU 301 to calculate the vehicle's current position and the amount of change in speed and direction.

[0057] The CPU 301 shown in Figure 3 implements the functions of the processing unit 113 of the terminal device 110 shown in Figure 1 by executing a program stored in the ROM 302, etc. The camera 317 in Figure 3 implements the functions of the imaging unit 111 in Figure 1, and the GPS unit 315 in Figure 3 implements the functions of the positioning unit 112 in Figure 1.

[0058] (Example server configuration) Figure 4 is a block diagram showing an example of the server's hardware configuration. The server 400, which constitutes the traffic congestion prediction device 100, has the same configuration as the navigation device 300 shown in Figure 3. Note that the GPS unit 315, various sensors 316, camera 317, etc., shown in Figure 3 are not required for the server 400.

[0059] The traffic congestion prediction device 100 shown in Figure 1 implements the traffic congestion prediction function by having the CPU 401 execute a predetermined program using programs and data recorded in the ROM 402 etc. provided in the server 400. The traffic information and traffic congestion prediction information are stored in the magnetic disk 405 etc. of the server 400. It also communicates with the navigation device 300 via the communication I / F 414 and outputs traffic congestion prediction information to the navigation device 300 in response to requests from the navigation device 300.

[0060] Furthermore, the server 400 may have a function to calculate the transmission time of images transmitted from each navigation device 300 and the speed of the mobile object on which the navigation device 300 is mounted. The server 400 can calculate the speed of the mobile object by calculation based on the amount of latitude and longitude movement of the navigation device 300 and the time required for movement.

[0061] The CPU 401 shown in Figure 4 executes programs stored in the ROM 402, etc., to realize the functions of the image processing unit 101, information coupling unit 102, traffic information processing unit 103, and traffic congestion prediction information management unit 105 of the traffic congestion prediction device 100 shown in Figure 1. The magnetic disk 305 and optical disk 307, etc., in Figure 4 realize the functions of the storage unit 104 shown in Figure 1.

[0062] (Image transmission processing by the navigation device) Figure 5 is a flowchart showing the image transmission process by the navigation device. The navigation device 300 constantly determines the position of the moving object using the GPS unit 315, etc., when the object is moving (step S501). Then, for example, at predetermined intervals (for example, every minute), the camera 317 takes a picture of the road (step S502). For example, by taking a picture of the road ahead, both the road and the moving object on the road are captured in the image.

[0063] Furthermore, the trigger (timing) for the camera 317 to photograph the road can be a request output from the server 400 to the navigation device 300, or instructions from the user of the navigation device 300, etc.

[0064] Then, the navigation device 300 transmits the captured image along with the location information at the time of capture to the server 400 each time an image is taken (step S503). Alternatively, the navigation device 300 may assign a timestamp to each image and transmit multiple images to the server 400 at predetermined intervals.

[0065] The navigation device 300 repeatedly takes images as described above and transmits them to the server 400. This allows the server 400 to collect images of multiple roads taken by multiple navigation devices 300.

[0066] (Server-based generation of traffic information) Figure 6 is a flowchart showing the process of generating traffic information by the server. First, the server 400 receives images and location information (shooting location) transmitted from the navigation device 300 (step S601). Next, it obtains the speed and time of the moving object (navigation device 300) at the time of image reception (step S602).

[0067] Server 400 may obtain the speed and time of these moving objects from the navigation device 300, or it may use data generated internally by Server 400. If generated internally by Server 400, the time used will be the internal time of Server 400 at the time of image reception. The movement of the moving object will be calculated using the speed of the moving object (navigation device 300) based on the amount of latitude and longitude movement and the time required for movement.

[0068] Next, the server 400 detects vehicles, driving lanes, and the number of lanes from the acquired image through image processing by the image processing unit 101 (step S603). At this time, for example, the captured image can be converted into an orthomosaic image to determine the position and distance of the vehicles and to detect the number of vehicles in each lane. In addition, the width of the current lane is derived from the orthomosaic image, and the number of lanes on the road and the distance between vehicles are determined.

[0069] Next, the server 400 associates location information, speed, time, number of vehicles per lane, and current lane information with each acquired image (ID) using the information linking unit 102, and stores this information as traffic information in the storage unit 104 (step S604). The above processing steps S601 to S604 is performed each time an image is received from the navigation device 300.

[0070] (Example of traffic information) Figure 7 is a diagram illustrating an example of traffic information. Figure 7(a) shows the captured image processing information 701 obtained by image processing a single image transmitted by a user. This captured image processing information 701 consists of lane information (number of lanes), vehicle lane information, and lane-specific vehicle information ID. In the example shown, it indicates that the vehicle is the first lane from the left on a three-lane road (lane information).

[0071] The details of the lane-specific vehicle information ID in Figure 7(a) are shown in Figure 7(b) as ID 702. Each lane-specific vehicle information ID 702 is assigned an ID to each image, and the ID is obtained through image processing. It contains information on each lane's position (the lane position of the moving object present in the image) and the distance between vehicles in each lane (the distance between the moving object present in the image and the user's vehicle). In the illustrated example, if ID=1, it indicates that a moving object is located in front of the user's vehicle (lane 1) with a distance of 45m, and that there are moving objects in front of the user's vehicle (lane 3) with distances of 30m and 60m, respectively. In the illustrated example, of the three lanes, the position of the left lane is indicated by "1", the position of the center lane by "2", and the position of the right lane by "3".

[0072] The shooting condition information 703 shown in Figure 7(c) is information that associates the shooting image processing information 701 shown in Figure 7(a) with the shooting conditions, namely the shooting location (latitude and longitude in the illustrated example), speed, and time, using a shooting ID.

[0073] Figure 7(d) shows traffic information 704. This is information accumulated from the shooting condition information 703 shown in Figure 7(c) for different locations and times of the same user, and for different users. By increasing the number of rows in this traffic information 704, it is possible to build a database of time-specific self-lane information, lane information, and lane-specific vehicle information ID (lane position, distance between vehicles) for the desired shooting location.

[0074] (Traffic congestion prediction processing using traffic information) Figure 8 is a flowchart showing the processing steps for traffic congestion prediction using traffic information. First, the traffic information processing unit 103 of the server 400 classifies the traffic information 704 by location, date, and time (step S801), and creates a statistical function for each classified data (step S802). In this state, the traffic information processing unit 103 waits for input of the request conditions included in the traffic congestion prediction request from the navigation device 300 (loop of step S803: No).

[0075] If a request condition is input (Step S803: Yes), the traffic information processing unit 103 calculates the congestion occurrence value using a statistical function corresponding to this request condition (Step S804). Then, the traffic information processing unit 103 determines whether the congestion occurrence value is high (positive value) (Step S805). If the congestion occurrence value is higher than a predetermined value (Step S805: Yes), it generates this congestion occurrence value as congestion prediction information and stores it in the storage unit 104 (Step S806). If the congestion occurrence value is low (Step S805: No), it determines that there is no congestion and terminates the current process.

[0076] A specific example of the above process will be explained. In the above example, the congestion occurrence value is compared with a predetermined threshold as a confidence level to determine whether congestion has occurred. For example, the explanation will be given when the congestion prediction requirements transmitted (input) from the navigation device 300 specify the traffic volume in the left lane, the traffic volume in the right lane, the date (day of the week), and the time of day.

[0077] The navigation device 300 transmits the captured images, and the server 400 analyzes these images to determine the traffic volume in the left lane, the traffic volume in the right lane, the date (day of the week), and the time of day. The date and time are determined by the server's data reception time, etc. One example of traffic volume is the number of cars per unit of time.

[0078] First, the traffic information processing unit 103 references the traffic information 704 corresponding to the date (day of the week) and time period specified in the storage unit 104. For example, it references the traffic information 704 compiled from data for road x, Tuesday, and the period from 12:00 to 13:00.

[0079] The traffic information processing unit 103 then takes the statistical function fn related to congestion occurrence and the required conditions X and Y as input values ​​and calculates the congestion probability using the following calculation formula. fn(X,Y)=an(X)+bn(Y) fn(xn,yn)=an(xn)+bn(yn) Traffic congestion value = max(1 / fn(|X-xn|,|Y-yn|))-tn

[0080] However, fn is a statistical function (n=0,1,...). The statistical function fn is a function used to determine the congestion occurrence value by comparing past congestion prediction data with the current traffic volume. X: input traffic volume for the left lane, Y: input traffic volume for the right lane, xn: statistical traffic volume for the left lane, yn: statistical traffic volume for the right lane, an: prediction coefficient for the left lane, bn: prediction coefficient for the right lane, tn: threshold. Traffic volume is obtained from the speed of moving objects at the time of image capture and the number of moving objects captured.

[0081] Traffic congestion prediction data is generated when congestion is determined to occur on a particular road. It involves statistically analyzing traffic information (e.g., traffic volume, speed, and time in each lane) leading up to the congestion determination, going back a predetermined time (the time back varies for each road) from the time the congestion determination was made. If this statistical data differs significantly from normal traffic information, it becomes traffic congestion prediction data for that road. However, a large number of prediction data sets exist for each road, categorized by parameters such as date and time. The determination of congestion may be made by the system's traffic congestion prediction device 100, or, if information is obtained from another external system, congestion may be determined when the distance between vehicles is narrow or the speed is low.

[0082] Furthermore, regarding the statistical function fn mentioned above, in the example above (a two-lane road), if the road has two characteristics—either "a traffic jam occurs when the right lane becomes extremely congested" or "a traffic jam occurs when both the right and left lanes become somewhat congested"—there are two possible traffic jam patterns, so fn will be n=2. These characteristics are calculated from traffic jam prediction data.

[0083] The input traffic volume for the right lane is the traffic volume for the right lane calculated from the image captured by the navigation device 300, and is indicated by the row number for position "3" in Figure 7(b) Lane-specific vehicle information ID 702 "1" (roads with lane-specific shooting ID "1" are 3-lane roads) (i.e., the number of vehicles present in the right lane of a 3-lane road). The input traffic volume for the left lane is indicated by the row number for position "1" in Figure 7(b) Lane-specific vehicle information ID 702 "1" (i.e., the number of vehicles present in the left lane of a 3-lane road).

[0084] The right (left) lane statistical traffic volume is the average or mode of traffic volume at position "3" in the relevant pattern within the congestion prediction data. The right (left) lane prediction coefficient is a coefficient that weights the difference between the right (left) lane statistical traffic volume and the right lane input traffic volume. For example, if there is a pattern on a road where "extreme congestion in the right lane leads to subsequent congestion," the right lane prediction coefficient will be increased. This will cause the difference in the right lane to be more significantly reflected in the final congestion occurrence value.

[0085] Let's explain a specific example of the above traffic congestion prediction. The traffic congestion prediction device 100 assumes that at time t on a certain road, the statistical data (data in the storage unit 104) for the 30 minutes prior to the occurrence of congestion shows two patterns: "the right lane is extremely congested" and "both the right and left lanes are somewhat congested." This indicates that signs of congestion will appear 30 minutes in advance on this road. These two patterns are stored in the storage unit 104 as traffic congestion prediction data for time t on this road.

[0086] Suppose an image is sent from the user of the navigation device 300 at time t. The traffic congestion prediction device 100 analyzes this image and determines that the right lane is very congested. The traffic congestion prediction device 100 calculates the similarity of the image analysis result to the traffic congestion prediction data using a statistical function and finds that it is very similar to the pattern of "the right lane is extremely congested". In this case, the traffic information processing unit 103 of the traffic congestion prediction device 100 determines that congestion will occur 30 minutes after this image is sent and outputs to the traffic congestion prediction information management unit 105 that this road will be congested in 30 minutes.

[0087] This section explains a specific example of the calculation using the statistical function of the traffic congestion prediction device 100. Since this road has the following two patterns for traffic congestion prediction, Pattern 1: "Traffic jams occur when the right lane becomes extremely congested." Pattern 2: "Traffic jams occur when both the left and right lanes are somewhat congested." Therefore, we will prepare two statistical functions fn. f1(X,Y)=a1(x1)+b1(y1) f2(X,Y)=a2(x2)+b1(y2)

[0088] x1: Left lane traffic volume (average or mode) from the predicted data for Pattern 1 → Statistical traffic volume for the left lane y1: Right lane traffic volume (mean and mode) for Pattern 1 prediction data x2: Left lane traffic volume (average and mode) for Pattern 2 prediction data y2: Right lane traffic volume (mean and mode) for Pattern 2 prediction data a1: Weight of the left lane in Pattern 1, b1: Weight of the right lane in Pattern 1 (In Pattern 1, we want to focus only on the right lane, so we increase the value of a1 and decrease the value of b1) a2: Weight of the left lane in Pattern 2, b2: Weight of the right lane in Pattern 2 (In Pattern 2, we want to focus on both lanes, so a2 = b2 is fine)

[0089] Then, when the traffic volume X and Y for each lane are input to the traffic congestion prediction information management unit 105, Calculate 1 / f1(|x1-X|,|y1-Y|) and 1 / f2(|x2-X|,|y2-Y|). The larger of these values ​​is adopted and compared to the threshold tn. If it is greater than tn, traffic congestion prediction information is generated.

[0090] The traffic congestion forecast information management unit 105 stores and manages traffic congestion forecast information in the storage unit 104, and transmits the traffic congestion forecast information to the navigation device 300 when a traffic congestion forecast request is received from the navigation device 300.

[0091] When the navigation device 300 requests a traffic congestion prediction, if there are specified conditions, such as weather, the traffic information processing unit 103 will refer to the traffic information 704, which includes these parameters. As the number of requested conditions increases, the traffic congestion prediction can be made with higher accuracy in response to the request.

[0092] Furthermore, in the above processing example, the system is configured to calculate congestion prediction information after waiting for a congestion prediction request from the navigation device 300, but this is not the only configuration. For example, the traffic information processing unit 103 may be configured to generate location- and time-specific congestion prediction information in advance based on the traffic information 704. In this case, if a congestion prediction request is received from the navigation device 300, the congestion prediction information for the corresponding location and time can be immediately transmitted to the navigation device 300.

[0093] (Regarding requests for traffic congestion forecasts and the transmission of traffic congestion forecast information) Figure 9 is a flowchart showing the processing steps from the request for traffic congestion prediction to the transmission of traffic congestion prediction information. First, the navigation device 300 sends a traffic congestion prediction request for a specified location or range to the server 400 (step S901).

[0094] The traffic congestion forecast information management unit 105 of the server 400 receives this traffic congestion forecast request (step S902) and transmits (distributes) the traffic congestion forecast for the corresponding specified location or range to the navigation device 300 (step S903).

[0095] The navigation device 300 receives traffic congestion forecast information transmitted from the server 400 (step S904) and displays this traffic congestion forecast information on the display unit (step S905).

[0096] Furthermore, the traffic congestion forecast information management unit 105 of the server 400 may be configured to monitor whether a change in the traffic congestion forecast has occurred after step S903, based on the possibility that the traffic congestion state may change over time (step S906). If no change in the traffic congestion forecast has occurred (step S906: No), the process is terminated, but if a change in the traffic congestion forecast has occurred (step Step S906: Yes), the traffic congestion prediction information is recalculated and transmitted again according to step S903.

[0097] (Regarding traffic congestion) Figure 10 is an explanatory diagram illustrating the conditions under which natural congestion occurs. As shown in Figure 10(a), as the number of moving vehicles increases, the distance between vehicles decreases. Next, as shown in Figure 10(b), when the speed of some vehicles 1001 slows down, the following vehicles 1011 and 1012 successively apply their brakes and slow down to avoid a rear-end collision. As a result, as shown in Figure 10(c), vehicle 1021 comes to a complete stop, and the following vehicles 1031, 1032, 1033, and 1034 also stop one after another, causing natural congestion.

[0098] (Regarding traffic congestion forecasts) The traffic congestion prediction in this embodiment predicts that as the number of moving vehicles increases, the distance between vehicles will decrease, by referring to images captured by the navigation device 300 and past data (traffic information 704). The server 400, by referring to past traffic information 704 corresponding to the requested conditions, determines which of the natural congestion examples (a) to (c) shown in Figure 10 corresponds to the situation, and transmits traffic congestion prediction information corresponding to the requested date and time included in the traffic congestion prediction request to the navigation device 300.

[0099] In addition, information about how traffic congestion will change after the requested time (for example, if the request is for the time shown in Figure 9(a), a prediction that natural congestion as shown in Figure 9(c) will occur 10 minutes later) can also be transmitted to the navigation device 300 as traffic congestion prediction information.

[0100] (Regarding the application of traffic congestion forecast information to route planning) If the aforementioned traffic congestion forecast information indicates a high probability of congestion occurring (high congestion value), the navigation device 300 can provide route guidance by adding traffic congestion forecast information to the roads on the candidate route where congestion is likely to occur.

[0101] The navigation device 300 of this embodiment performs route searching while taking into account the traffic congestion prediction information described above. As a result, since the average travel time on roads (links) where congestion is expected is long, a route that avoids these congested areas is searched for.

[0102] However, in route searching using traffic congestion prediction information in this way, the actual presence or absence of traffic congestion at the relevant location will be determined when the predicted time arrives. Therefore, it is possible that the predicted traffic congestion does not occur (no congestion is occurring) at the predicted time. In such cases, when the predicted time is reached, the navigation device 300 will recalculate the route and will not use the traffic congestion prediction information used previously.

[0103] The above route search is not limited to the navigation device 300; it may also be performed by the server 400. The server 400 can obtain location information such as the current location and destination from the navigation device 300 and perform a route search including other search conditions such as the time.

[0104] In the above embodiment, the system detects traffic volume by processing images to obtain the number of lanes, the number of vehicles, the distance between vehicles, etc. from the captured image. However, it is also possible to configure the system to acquire other parameters through image processing. For example, if there is a construction guidance sign on a road lane in the captured image, there is a possibility of sudden traffic congestion due to the construction. In addition, other parameters that may cause congestion, such as vehicle types (buses, police cars) and pedestrian crossings, can also be used.

[0105] Therefore, by statistically processing the similarities and differences between images before and after traffic congestion occurs, the accuracy of the traffic information 704 can be improved.

[0106] Furthermore, in the above embodiment, the traffic congestion prediction is performed by a traffic congestion prediction device 100 such as a server. However, it is also possible to configure the device so that a single terminal device 110, such as a car navigation system or a personal computer, performs the traffic congestion prediction and transmits it to other terminal devices 110 that have requested the traffic congestion prediction.

[0107] The traffic congestion prediction method described in this embodiment can be implemented by executing a pre-prepared program on a computer such as a personal computer or workstation. This program is recorded on a computer-readable recording medium such as a hard disk, flexible disk, CD-ROM, MO, or DVD, and is executed by being read from the recording medium by the computer. This program may also be transmitted via a network such as the Internet. [Explanation of Symbols]

[0108] 100 Traffic congestion prediction device 101 Image Processing Unit 102 Information connection part 103 Traffic Information Processing Unit 104 Storage Unit 105 Traffic Congestion Forecast Information Management Department 110 Terminal device 111 Imaging Unit 112 Positioning Unit 113 Processing Unit

Claims

[Claim 1] An acquisition unit that acquires images of moving objects on the road and location information indicating the location where the images were taken, An image processing unit that extracts traffic information, which is information about moving objects present in the aforementioned image, A traffic information processing unit calculates information regarding traffic congestion prediction on the road corresponding to the captured location using the traffic information extracted by the image processing unit, A traffic congestion prediction system characterized by having the following features.

Images