Moving position correction system and moving position correction apparatus

Abstract

To provide a moving position correction system capable of acquiring a highly accurate traveling position of a vehicle even when inspection equipment having low positioning accuracy is used.SOLUTION: A moving position correction system according to the present disclosure includes a first moving body, a second moving body, and a moving position correction apparatus that performs communication via a communication network between the first and second moving bodies. The moving position correction apparatus detects a first feature point position in an image of a feature point of a first planimetric feature and a first displacement amount of the first feature point position on the basis of first image information acquired from the first moving body, detects a second feature point position in an image of a feature point of a second planimetric feature and a second displacement amount of the second feature point position on the basis of second image information acquired from the second moving body, compares the first feature point position and the first displacement amount with the second feature point position and the second displacement amount, and corrects a position of the second moving body that moves after the first moving body.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a movement position correction system, and more particularly to a movement position correction system that can correct the traveling position of a vehicle and obtain a highly accurate traveling position. [Background technology]

[0002] In recent years, there has been a demand for technology that can more accurately determine the position of a vehicle on a map. For example, Patent Document 1 discloses a position determination device that stores multiple pieces of reference data obtained by extracting characteristic parts of a landscape image captured in advance by an on-board camera, each piece being associated with a corresponding position on a map, and performs a matching process to determine whether the landscape image captured by the on-board camera matches the multiple pieces of reference data, thereby determining the position of the vehicle on a map. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2013-032954 Summary of the Invention [Problem to be solved by the invention]

[0004] The technology disclosed in Patent Document 1 is aimed at a system capable of matching roads with maps from a navigation device, etc. The navigation device can be used for autonomous navigation using a gyroscope or vehicle speed acquired from the vehicle, and is equipped with a separate GNSS (Global Navigation Satellite System) antenna, providing high positioning accuracy.

[0005] In recent years, in order to address the shortage of personnel for infrastructure inspections and reduce costs, it has become common to install commercially available, inexpensive inspection equipment, such as dashcams (hereafter abbreviated as "dashcams"), on vehicles other than infrastructure inspection vehicles, such as garbage trucks and buses, and conduct daily infrastructure inspections. There is also a need to distinguish between lanes and inbound and outbound lane markings to understand road deterioration information. However, commercially available, inexpensive equipment such as dashcams lack the functionality to acquire vehicle speed data from the vehicle, and because GNSS antennas are built into the equipment, their positioning accuracy is low, with a positioning error of approximately 10 meters. This makes it impossible to identify driving lanes, which are approximately 3 meters wide, while the vehicle is traveling, and therefore fails to meet the above needs.

[0006] Due to the objective of comprehensive road surveys and the frequent update of inexpensive equipment, there is a tendency for multiple types of equipment to be used. When multiple models of dashcams are used, there is a high possibility that positional errors will increase when comparing images with previously acquired images due to differences in resolution and scale. While it is possible to understand the characteristics of the dashcam camera and preprocess the images, it is not realistic to understand all commercially available dashcam models and estimate the camera parameters for all models.

[0007] One possible method is to use an algorithm that is robust to changes in viewpoint and scale using features such as SIFT (Scale-invariant feature transform) or SURF (Speed-Upped Robust Feature), but it is difficult to track the exact position of the same feature point in consecutive image frames such as those used in a dashcam.

[0008] The present disclosure has been made to solve the above-mentioned problems, and aims to provide a moving position correction system that can obtain a vehicle's running position with high accuracy even when inspection equipment with low positioning accuracy is used. [Means for solving the problem]

[0009] A movement position correction system according to the present disclosure is a movement position correction system including a first moving body, a second moving body, and a movement position correction device that communicates between the first and second moving bodies via a communication network, wherein the first moving body has a first positioning device capable of sub-meter or centimeter-level positioning and a first imaging device that takes a first image of the periphery of the first moving body, and has a function of acquiring first position information including the positioning result of the first positioning device and first image information including the first image and transmitting them to the movement position correction device via the communication network, and the second moving body has a second positioning device capable of meter-level positioning and a second imaging device that takes a second image of the periphery of the second moving body, and has a function of acquiring second position information including the positioning result of the second positioning device and second image information including the second image and transmitting them to the movement position correction device via the communication network, and The moving position correction device includes a first feature point analysis unit that detects a first feature point position in an image of a feature point of a first feature and a first displacement amount of the first feature point position in each image frame based on the first image information acquired from the first moving body; a feature point storage unit that stores the first feature point position and the first displacement amount of the first feature point position; a second feature point analysis unit that detects a second feature point position in an image of a feature point of a second feature and a second displacement amount of the second feature point position in each image frame based on the second image information acquired from the second moving body; a feature point storage unit that stores the second feature point position and the second displacement amount of the second feature point position; and a moving position correction unit that compares the first feature point position and the first displacement amount of the first feature point position with the second displacement amount of the second feature point position and the second feature point position, and corrects the position of the second moving body that moves after the first moving body. [Effects of the Invention]

[0010] According to the moving position correction system of the present disclosure, the position of a second moving body having a low-precision second positioning device is corrected using the results of positioning by a first moving body having a high-precision first positioning device, thereby making it possible to increase the precision of the position of the second moving body. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a block diagram showing a configuration of a movement position correction system according to the present disclosure. [Figure 2] 4 is a diagram schematically showing the amount of displacement of feature point positions of an image in the first embodiment. FIG. [Figure 3] 4 is a diagram schematically showing the amount of displacement of feature point positions of an image in the first embodiment. FIG. [Figure 4] 4 is a flowchart illustrating an operation flow of a preliminary traveling vehicle according to the first embodiment. [Figure 5] 4 is a flowchart illustrating an operation flow of a post-traveling vehicle according to the first embodiment. [Figure 6] 10 is a flowchart illustrating an operation flow regarding a pre-traveling vehicle in the movement position correction device according to the first embodiment. [Figure 7] 10 is a flowchart illustrating an operation flow regarding a subsequent traveling vehicle in the movement position correction device according to the first embodiment. [Figure 8] FIG. 10 is a diagram illustrating an operation for identifying whether the vehicle is traveling in a driving lane or an overtaking lane in a modification of the first embodiment. [Figure 9] FIG. 10 is a diagram illustrating an operation for identifying whether the vehicle is traveling in a driving lane or an overtaking lane in a modification of the first embodiment. [Figure 10] 10 is a flowchart illustrating an operation flow regarding a subsequent traveling vehicle in the movement position correction device of the second embodiment. [Figure 11] FIG. 10 is a diagram illustrating lane discrimination in the second embodiment. [Figure 12] 11 is a flowchart illustrating an operation flow relating to a pre-traveling vehicle in the movement position correction device of the third embodiment. [Figure 13] 11 is a flowchart illustrating an operation flow relating to a subsequent traveling vehicle in the movement position correction device of the third embodiment. [Figure 14] FIG. 10 is a diagram schematically showing detected deterioration information of a road in an image. [Figure 15] FIG. 13 is a diagram illustrating lane discrimination in the fourth embodiment. [Figure 16] 13 is a flowchart illustrating an operation flow relating to a pre-traveling vehicle in the movement position correction device of the fourth embodiment. [Figure 17] 13 is a flowchart illustrating an operation flow relating to a subsequent traveling vehicle in the movement position correction device of the fourth embodiment. [Figure 18] FIG. 13 is a diagram illustrating lane discrimination in the fifth embodiment. [Figure 19] 13 is a flowchart illustrating an operation flow relating to a pre-traveling vehicle in the movement position correction device of the fifth embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0012] <First Embodiment> Fig. 1 is a block diagram showing the configuration of a movement position correction system 100 according to the present disclosure. As shown in Fig. 1, the movement position correction system 100 includes a probe vehicle 10 (first moving body) as a pre-traveling vehicle, a probe vehicle 20 (second moving body) as a post-traveling vehicle, and a movement position correction device 30 as a server device.

[0013] The probe vehicle 10 has an image information acquisition unit 11, a probe information generation unit 12, a communication unit 13, an imaging device 14 (first imaging device), a position information acquisition unit 15, and a high-precision locator device 16 (first positioning device).

[0014] The imaging device 14 is a camera such as an infrared camera or an RGB camera installed inside or outside the vehicle, and the image information acquisition unit 11 acquires the image output from the imaging device 14 and outputs the image and timestamp to the probe information generation unit 12.

[0015] The high-precision locator device 16 is, for example, a GNSS receiver, and outputs satellite positioning information. The position information acquisition unit 15 acquires the high-precision satellite positioning information output from the high-precision locator device 16, and outputs the high-precision position, driving lane (movement lane), and timestamp of the probe vehicle 10 to the probe information generation unit 12.

[0016] The high-precision locator device 16 outputs satellite positioning information with a positioning error of sub-meter or centimeter order, and the position information acquisition unit 15 uses this information to output highly accurate latitude and longitude coordinates and driving lane of the probe vehicle 10.

[0017] The probe information generation unit 12 inputs the image and timestamp input from the image information acquisition unit 11, and the high-precision position, driving lane, and timestamp input from the position information acquisition unit 15 to the communication unit 13, and the communication unit 13 transmits these data as probe information to a communication network NW such as the Internet.

[0018] The probe vehicle 20 includes an image information acquisition unit 21, a probe information generation unit 22, a communication unit 23, an imaging device 24 (second imaging device), a position information acquisition unit 25, and a locator device 26 (second positioning device).

[0019] The imaging device 24 is a camera such as an infrared camera or an RGB camera installed inside or outside the vehicle, and the image information acquisition unit 21 acquires the image output from the imaging device 24 and outputs the image and timestamp to the probe information generation unit 22.

[0020] The locator device 26 is, for example, a GNSS receiver, and outputs low-accuracy satellite positioning information. The position information acquisition unit 25 acquires the low-accuracy satellite positioning information output from the locator device 26, and outputs the low-accuracy position and timestamp of the probe vehicle 20 to the probe information generation unit 22.

[0021] The locator device 26 outputs meter-level satellite positioning information with a positioning error on the order of GPS accuracy, and the position information acquisition unit 25 uses this information to output the latitude and longitude coordinates of the probe vehicle 20 .

[0022] The probe information generation unit 22 inputs the image and timestamp input from the image information acquisition unit 21 and the low-accuracy location and timestamp input from the location information acquisition unit 25 to the communication unit 23, and the communication unit 23 transmits this data as probe information to a communication network NW such as the Internet.

[0023] The movement position correction device 30 has a probe information input unit 31, a probe information storage unit 32, a feature point analysis unit 33 (first feature point analysis unit), and a feature point storage unit 34 (first feature point storage unit) related to the probe information from the probe vehicle 10, a probe information input unit 41, a probe information storage unit 42, a feature point analysis unit 43 (second feature point analysis unit), and a feature point storage unit 44 (second feature point storage unit) related to the probe information from the probe vehicle 20, and a movement position correction unit 51 (movement position correction unit).

[0024] The probe information input unit 31 is a communication unit that receives input of probe information transmitted from the probe vehicle 10 via the communication network NW, and the input probe information is stored in the probe information storage unit 32.

[0025] The feature point analysis unit 33 acquires the probe information of the probe vehicle 10, such as the high-precision position, driving lane, image, and timestamp, from the probe information storage unit 32, and determines the feature points of features in the image and their positions within the image by image processing. These feature points and their positions within the image can be identified by, for example, object detection technology using deep learning.

[0026] Additionally, as past information, the high-precision position, driving lane, image, and feature point positions of the immediately previous timestamp are obtained from the feature point storage unit 34, and the amount of displacement of the feature point positions between each image frame is calculated. More specifically, the feature point positions of the feature captured in the image frame of the immediately previous timestamp are obtained from the feature point storage unit 34, and the feature point positions of the same feature captured in the image frame of the new timestamp obtained from the probe information storage unit 32 are obtained, and the amount of displacement between the two feature point positions is calculated.

[0027] These can be determined using machine learning and deep learning. To do this, a model is used to infer the degree to which the positions of feature points in an image will displace as the vehicle moves forward. The explanatory variables (input) are the vehicle's GNSS-based position, vehicle speed, vehicle acceleration, time, etc., and the objective variable (output) is the amount of displacement of the feature position in the image. For example, if a vehicle travels a certain distance from a specific position, say 100 meters, at a specific vehicle speed and acceleration, it is assumed that the amount of movement of the feature position in the image will differ between the left and right lanes. These explanatory variables and objective variables are labeled and trained as training data, and a model is created using artificial intelligence (AI). This makes it possible to infer the amount of displacement of the feature position from the vehicle's speed and acceleration when traveling a certain distance from a certain position.

[0028] The feature point analysis unit 33 outputs the feature points of the image and the amount of time-series displacement of the feature point positions from the past (one timestamp before), and stores them in the feature point storage unit 34. The feature point analysis unit 33 also outputs the high-precision position, driving lane, image, and timestamp, which are probe information from the probe vehicle 10, and stores them in the feature point storage unit 34 in chronological order.

[0029] The probe information input unit 41 is a communication unit that receives input of probe information transmitted from the probe vehicle 20 via the communication network NW, and the input probe information is stored in the probe information storage unit .

[0030] The feature point analysis unit 43 acquires the low-accuracy position, image, and timestamp, which are the probe information of the probe vehicle 20, from the probe information storage unit 42, and determines the feature points of the features in the image and the positions of the feature points in the image by image processing. The method of determining these is the same as that of the feature point analysis unit 33.

[0031] Furthermore, as past information, the low-accuracy positions, images, and feature point positions at the immediately previous timestamp are obtained from the feature point storage unit 44, and the amount of displacement of the feature point positions between each image frame is calculated. The method for calculating this is the same as that used by the feature point analysis unit 33.

[0032] The feature point analysis unit 43 outputs the feature points of the image and the amount of time-series displacement of the feature point positions from the past (one timestamp before), and stores them in the feature point storage unit 44. The feature point analysis unit 43 also outputs the low-accuracy positions, images, and timestamps, which are probe information from the probe vehicle 20, and stores them in the feature point storage unit 44 in chronological order.

[0033] The movement position correction unit 51 acquires the time series displacement amounts of the feature points and feature point positions of the image from the past (one timestamp before) from the feature point memory unit 34, acquires the time series displacement amounts of the feature points and feature point positions of the image from the past (one timestamp before) from the feature point memory unit 44, compares the time series displacement amounts of the feature points and feature point positions of the image from the past (one timestamp before), and outputs a value obtained by correcting the low-accuracy position from the probe vehicle 20 to a high-accuracy position.

[0034] For example, if the amount of displacement of the feature point position in the image obtained by the probe vehicle 10 (pre-travel vehicle) differs from the amount of displacement of the feature point position in the image obtained by the probe vehicle 20 (subsequent travel vehicle), the low-precision position from the probe vehicle 20 is corrected so that the amount of displacement of the feature point position in the image obtained by the subsequent travel vehicle matches the amount of displacement of the feature point position in the image obtained by the pre-travel vehicle.

[0035] This process will be explained using Figures 2 and 3. Figures 2 and 3 are diagrams that respectively show the amount of displacement of the feature point positions of the image obtained by the probe vehicle 10 and the amount of displacement of the feature point positions of the image obtained by the probe vehicle 20. In Figures 2 and 3, the probe vehicle 10, which is the preceding vehicle, travels in the right lane, which is the passing lane, and then the probe vehicle 20, which is the following vehicle, actually travels in the left lane, which is the driving lane, but the position information is recorded as if it were traveling in the right lane, which is the passing lane. This represents the case.

[0036] 2 and 3 show the change in the position of the tower TW in the image frame when the feature point is the tower TW in front of the vehicle, that is, the change over time in the amount of displacement of the feature point position.

[0037] FIG. 2 shows an image R1 at time t, an image R2 at time t+Δ1, and an image R3 at time t+Δ2, and the probe vehicle 10 at the time each image was acquired is indicated by the symbols R1, R2, and R3.

[0038] 3 shows an image Ra1 at time t, an image Ra2 at time t+Δ1, and an image Ra3 at time t+Δ2, and the probe vehicles 20 at the times when each image was acquired are denoted by symbols Ra1, Ra2, and Ra3. In addition, in the right lane, which is the passing lane, the traveling positions of vehicles on the images recorded using low-accuracy position information from the probe vehicle 20 are denoted by Ra1i, Ra2i, and Ra3i.

[0039] In Figures 2 and 3, the displacement amount DV of the tower TW shown in the time-series images R1 to R3 is different from the displacement amount DVa of the tower TW shown in the time-series images Ra1 to Ra3. This is because low-precision position information from the probe vehicle 20 is used. By correcting this with the displacement amount of the tower TW in images R1 to R3, the low-precision position from the probe vehicle 20 is corrected. This makes it possible to improve the accuracy of the traveling position of the probe vehicle 20 and obtain position information as if the probe vehicle 20 were traveling in the traveling lane in Figure 3. In other words, by running a probe vehicle 10 equipped with equipment capable of acquiring high-precision satellite positioning information in advance and running a probe vehicle 20 equipped with equipment capable of only acquiring low-precision satellite positioning information later, the traveling position of the probe vehicle 20 can also be improved in accuracy. This makes it possible to perform accurate infrastructure inspections even with vehicles other than infrastructure inspection vehicles, such as garbage trucks or buses.

[0040] Next, the operation flow of the probe vehicle 10, the probe vehicle 20, and the movement position correction device 30 will be described.

[0041] Fig. 4 is a flowchart illustrating the operation flow of the probe vehicle 10. As shown in Fig. 4, when the probe vehicle 10 starts traveling, first, the high-accuracy locator device 16, such as a GNSS receiver, installed in the vehicle acquires vehicle position information (high-accuracy position), and the position information acquisition unit 15 associates the position information with the time of acquisition of the position information, and outputs the associated information to the probe information generation unit 12 (step S120).

[0042] Next, the imaging device 14 acquires an image of the surroundings of the vehicle, and the image information acquisition unit 11 associates the image with the image acquisition time, and outputs the associated information to the probe information generation unit 12 (step S121).

[0043] Next, the probe information generation unit 12 time-synchronizes the vehicle position information (high-precision position) with the image, creates probe information by combining it with information on the driving lane, and transmits the created probe information to the movement position correction device 30 (server device) (step S122).

[0044] Next, it is checked whether the vehicle has finished traveling (step S123), and if the vehicle has finished traveling (if YES), the series of operations is terminated, and if the vehicle has not finished traveling (if NO), the operations from step 120 onwards are repeated.

[0045] Fig. 5 is a flowchart illustrating the operation flow of probe vehicle 20. As shown in Fig. 5, when probe vehicle 20 starts traveling, first, locator device 26 such as a GNSS receiver installed in the vehicle acquires vehicle position information (low-accuracy position), and position information acquisition unit 25 associates the position information with the time of position information acquisition and outputs the associated information to probe information generation unit 22 (step S130). Note that locator device 26 installed in probe vehicle 20 is a device with low positioning accuracy, such as a dashcam or smartphone, called an edge terminal.

[0046] Next, the imaging device 24 acquires an image of the surroundings of the vehicle, and the image information acquisition unit 21 associates the image with the image acquisition time, and outputs the associated information to the probe information generation unit 22 (step S131).

[0047] Next, the probe information generating unit 22 generates probe information by time-synchronizing the vehicle position information (low-accuracy position) with the image, and transmits the generated probe information to the movement position correcting device 30 (server device) (step S132).

[0048] Next, it is checked whether the vehicle has finished traveling (step S133), and if the vehicle has finished traveling (if YES), the series of operations is terminated, and if the vehicle has not finished traveling (if NO), the operations from step 130 onwards are repeated.

[0049] Fig. 6 is a flowchart illustrating an operation flow regarding a pre-traveling vehicle in the movement position correction device 30. As shown in Fig. 6, when the movement position correction device 30 starts operation, first, the probe information input unit 31 receives probe information from the probe vehicle 10 (step S140).

[0050] Next, the probe information is stored in the probe information storage unit 32 (step S141).

[0051] Next, the feature point analysis unit 33 calculates feature points of features in the image and the positions of each feature point in the image (step S142).

[0052] Next, the feature point analysis unit 33 calculates the amount of displacement of the feature point positions of each image frame (step 143). This is a process of comparing the feature point positions of a feature photographed in the image frame with the previous timestamp with the feature point positions of the same feature photographed in the image frame with the new timestamp, and determining the amount of displacement between the two feature point positions.

[0053] Next, the feature point positions and the displacement amounts of the feature point positions are stored in the feature point storage unit 34 (step S144).

[0054] Fig. 7 is a flowchart illustrating an operation flow regarding a subsequent traveling vehicle in the movement position correction device 30. As shown in Fig. 7, when the movement position correction device 30 starts operation, first, the probe information input unit 41 receives probe information from the probe vehicle 20 (step S150).

[0055] Next, the probe information is stored in the probe information storage unit 42 (step S151).

[0056] Next, the feature point analysis unit 43 calculates feature points of features in the image and the positions of each feature point in the image (step S152).

[0057] Next, the feature point analysis unit 43 calculates the amount of displacement of the feature point positions of each image frame (step 153). This is a process of comparing the feature point positions of a feature photographed in the image frame with the previous timestamp with the feature point positions of the same feature photographed in the image frame with the new timestamp, and determining the amount of displacement between the two feature point positions.

[0058] Next, the feature point positions and the displacement amounts of the feature point positions are stored in the feature point storage unit 44 (step S154).

[0059] Next, the moving position correcting unit 51 corrects the vehicle position in the subsequent traveling by comparing it with the feature point positions and the displacement amounts of the feature point positions in the previous traveling (step S155).

[0060] <Modification> When the same feature is photographed from different vehicle positions, the positions of the feature's feature points in the image will differ. For example, as the vehicle approaches a feature such as a sign that was visible in the distance in the direction of travel, the feature appears to shift outward from the field of view (image). The amount of this shift in position varies depending on the vehicle's position in the image, for example, whether it is in the driving lane or the overtaking lane. This is because the appearance of the feature differs when the vehicle is driving in the driving lane (left lane) and when the vehicle is driving in the overtaking lane (right lane).

[0061] Therefore, by comparing the degree of change in the amount of displacement with the degree of change in the amount of displacement already acquired by the preceding vehicle, it is possible to identify whether the vehicle is traveling in the driving lane or the overtaking lane.

[0062] 8 and 9 are diagrams for explaining the operation of identifying whether the vehicle is traveling in the driving lane or the passing lane. Figures 8 and 9 are diagrams for showing the amount of displacement of the feature point positions in the image obtained by the probe vehicle 10 and the amount of displacement of the feature point positions in the image obtained by the probe vehicle 20, respectively.

[0063] 8 shows the change in the position of the tower TW in the image frame when the feature point is the tower TW ahead of the vehicle, that is, the change over time in the displacement of the feature point position, and the central diagram shows a probe vehicle 10L traveling in the driving lane and a probe vehicle 10R traveling in the passing lane. The probe vehicles 10L and 10R do not necessarily have to travel at the same time.

[0064] 8, the left diagram shows image L1 at time t, image L2 at time t+Δ1, and image L3 at time t+Δ2, with the probe vehicle 10L at the times when each image was acquired being labeled L1, L2, and L3. The right diagram shows image R1 at time t, image R2 at time t+Δ1, and image R3 at time t+Δ2, with the probe vehicle 10R at the times when each image was acquired being labeled R1, R2, and R3.

[0065] 8, the tower TW looks different when traveling in the driving lane (left lane) and when traveling in the overtaking lane (right lane), and the displacement amount DVR of the tower TW shown in images R1 to R3 is different from the displacement amount DVL of the tower TW shown in images L1 to L3. These displacement amounts DVR and DVL of the tower TW, which differ depending on the lane, are linked to the traveling position and lane and stored in the feature point storage unit 34 of the movement position correction device 30.

[0066] 9 shows an image Ra1 at time t, an image Ra2 at time t+Δ1, and an image Ra3 at time t+Δ2 obtained by the probe vehicle 20. The displacement amount DVa of the tower TW in these images Ra1 to Ra3 is linked to the traveling position and stored in the feature point storage unit 44 of the movement position correction device 30.

[0067] By comparing whether the change in the displacement amount DVa of the tower TW in these images Ra1 to Ra3 matches with the change in the displacement amounts DVL and DVR obtained by the probe vehicles 10L and 10R, respectively, it is possible to determine which lane the probe vehicle 20 (subsequent traveling vehicle) is traveling in. From the examples of Figures 8 and 9, it can be seen that the probe vehicle 20 is traveling in the passing lane because the displacement amount DVa is close to the displacement amount DVR.

[0068] <Embodiment 2> In the modified example of the first embodiment, an operation for identifying whether the vehicle is traveling in the driving lane or the passing lane has been described, but in the second embodiment, an operation for more easily determining the lane will be described.

[0069] 10 is a flowchart illustrating an operation flow relating to a subsequent traveling vehicle in the movement position correction device 30 of the movement position correction system 100 of the embodiment 2. Note that the operation flow relating to a pre-traveling vehicle in the movement position correction device 30 is the same as the flowchart explained using FIG. 6, and therefore the explanation will be omitted.

[0070] The operational flow of the probe vehicles 10 and 20 is the same as the flow chart explained using Figures 4 and 5, so the explanation will be omitted. The configurations of the probe vehicles 10, 20, and movement position correction device 30 of the movement position correction system 100 are the same as the configuration shown in Figure 1, so the explanation will be omitted.

[0071] As shown in FIG. 10, when the movement position correction device 30 starts its operation, first, the probe information input unit 41 receives the probe information from the probe vehicle 20 (step S210).

[0072] Next, the probe information is stored in the probe information storage unit 42 (step S211).

[0073] Next, the feature point analysis unit 43 calculates feature points of features in the image and the positions of each feature point in the image (step S212).

[0074] Next, the feature point analysis unit 43 calculates the amount of displacement of the feature point positions of each image frame (step 213).

[0075] Next, the feature point positions and the displacement amounts of the feature point positions are stored in the feature point storage unit 44 (step S214).

[0076] Next, the movement position correction unit 51 performs image recognition of a vehicle traveling alongside or on the opposite side traveling in a lane adjacent to the lane in which the probe vehicle 20 is traveling, and determines the lane in which the probe vehicle 20 is traveling (step S215).

[0077] That is, the feature point of the "vehicle" traveling next to the probe vehicle 20 detected by the feature point analysis unit 43 and the position of the "vehicle" in the image included in the probe information from the probe vehicle 20 are detected by the movement position correction unit 51 via the feature point storage unit 44. The movement position correction unit 51 determines whether the "vehicle" is detected on the left or right side of the position of the probe vehicle 20, and determines the lane in which the probe vehicle 20 is traveling from that result. Therefore, unlike the modified example of the first embodiment, there is no need to use the analysis results of the image captured by the probe vehicle 10, and it is possible to more easily determine the lane in which the probe vehicle 20 is traveling.

[0078] FIG. 11 is a diagram showing a probe vehicle 20 traveling in the passing lane and a vehicle MVX traveling alongside the probe vehicle 20 in the adjacent lane, and shows the positions of each vehicle in chronological order as they head towards tower TW.

[0079] <Third Embodiment> As explained above, there is a need to distinguish between road lanes and grasp road deterioration information. The movement position correction system 100 according to the present disclosure can satisfy this need, and a method for grasping road deterioration information will be explained below as a third embodiment using Figures 12 and 13.

[0080] FIG. 12 is a flowchart illustrating the operation flow of the movement position correction device 30 with respect to the pre-traveling vehicle, and FIG. 13 is a flowchart illustrating the operation flow with respect to the post-traveling vehicle.

[0081] The operational flow of the probe vehicles 10 and 20 is the same as the flowchart explained using Figures 4 and 5, so the explanation will be omitted. Also, the configurations of the probe vehicles 10, 20, and movement position correction device 30 of the movement position correction system 100 are the same as the configuration shown in Figure 1, so the explanation will be omitted.

[0082] As shown in FIG. 12, when the movement position correction device 30 starts its operation, first, the probe information input unit 31 receives the probe information from the probe vehicle 10 (step S310).

[0083] Next, the probe information is stored in the probe information storage unit 32 (step S311).

[0084] Next, the feature point analysis unit 33 detects road deterioration information in the image (step S312). Road deterioration information includes cracks on the road surface, the crack rate, potholes, and blurred white lines. In addition to images, road deterioration information can also be acquired from an acceleration sensor mounted on the vehicle's equipment. That is, road irregularities can be detected from vertical acceleration information of the vehicle, which corresponds to, for example, passing over a hole (pothole) in the road.

[0085] Next, the deterioration information of the road in the image is stored by the feature point storage unit 34. That is, the deterioration information of the road is stored in association with the high-accuracy position information (step S313).

[0086] In this case, the deterioration information is categorized by weather or season and stored, and the deterioration information acquired by the probe vehicle 20 is compared with the deterioration information acquired by the probe vehicle 10 on a category basis, thereby improving the accuracy of position correction of the probe vehicle 20.

[0087] If a vehicle is equipped with a camera such as a dashcam and records the road while the vehicle is driving, the road deterioration information can be detected in post-processing. Therefore, both the probe vehicle 10 and the probe vehicle 20 can detect road deterioration information in post-processing. The probe vehicle 10 can link more accurate location information with road deterioration information.

[0088] Furthermore, road deterioration information looks different depending on the weather, season, and time of day, so simply comparing the deterioration information of the same road may result in different levels of deterioration. For example, a crack photographed on a day after rain may look different when photographed on a sunny day. This is because cracks tend to be more clearly visible when the road is wet after rain. If the road deterioration information does not match between the pre-driving and post-driving, the vehicle may be deemed to be in a different location even though it is actually driving in the same location. For this reason, the accuracy of the comparison can be improved by categorizing and storing road deterioration information and comparing images with the same weather, season, time of day, etc.

[0089] As shown in FIG. 13, when the movement position correction device 30 starts its operation, first, the probe information input unit 41 receives the probe information from the probe vehicle 20 (step S320).

[0090] Next, the probe information is stored in the probe information storage unit 42 (step S321).

[0091] Next, the feature point analysis unit 43 detects deterioration information of the road in the image (step S322).

[0092] Next, the feature point storage unit 44 stores the deterioration information of the road in the image (step S323). That is, the road deterioration information is stored in association with the low-accuracy position information. At this time, the deterioration information is categorized by weather or season and stored.

[0093] Next, the moving position correcting unit 51 compares the road deterioration information linked to the high-precision position information obtained during the preliminary travel and corrects the vehicle position for the subsequent travel (step S324). That is, the road deterioration information linked to the high-precision position information obtained during the preliminary travel is compared with the road deterioration information obtained during the subsequent travel, and if they match, the low-precision vehicle position obtained during the subsequent travel is corrected with the high-precision position information obtained during the preliminary travel, thereby improving the position accuracy of the vehicle position for the subsequent travel. Therefore, even in a probe vehicle 20 that can only detect vehicle positions with low accuracy, it is possible to satisfy the need to distinguish between road lanes and inbound and outbound roads and to grasp road deterioration information.

[0094] Fig. 14 is a diagram that schematically shows road deterioration information in a detected image, in which cracks CR in the image frame are shown surrounded by a frame F1 in an image seen from the driver's seat of a vehicle. The crack rate within this frame F1 is calculated and stored as road deterioration information in the feature point storage units 34 and 44. Note that the frame F2 in Fig. 14 indicates the detection range of cracks CR.

[0095] <Fourth Embodiment> In the modified example of the first embodiment, an operation for identifying whether the vehicle is traveling in the driving lane or the passing lane has been described. In the fourth embodiment, an operation for more easily determining the lane will be described.

[0096] The appearance, i.e., the inclination, of letters and figures on road markings and signs varies depending on the lane of the vehicle. Fig. 15 shows a road marking visible from each of the driving lane and the passing lane, in this case, the letters indicating the speed limit SI written on the driving lane. In Fig. 15, the central diagram shows a probe vehicle 10L traveling in the driving lane and a probe vehicle 10R traveling in the passing lane, the left diagram shows an image LD acquired by the probe vehicle 10L, and the right diagram shows an image RD acquired by the probe vehicle 10R.

[0097] As shown in Fig. 15, the speed limit SI appears to slope downward in image LD, whereas it appears to slope upward in image RD.

[0098] Since the probe vehicle 10 can identify the lane from high-precision position information, by linking the lane information with the differences in how the characters and figures appear, it becomes possible to identify the driving lane from how the characters and figures appear on the probe vehicle 20. In addition, the differences in how the characters and figures appear can also be used to correct the vehicle position for subsequent travel.

[0099] Hereinafter, as a fourth embodiment, an operation for correcting the vehicle position during subsequent travel using differences in how characters and figures appear will be described with reference to FIGS. 16 and 17. FIG.

[0100] FIG. 16 is a flowchart illustrating the operation flow of the movement position correction device 30 with respect to the pre-traveling vehicle, and FIG. 17 is a flowchart illustrating the operation flow with respect to the post-traveling vehicle.

[0101] The operational flow of the probe vehicles 10 and 20 is the same as the flowchart explained using Figures 4 and 5, so the explanation will be omitted. Also, the configurations of the probe vehicles 10, 20, and movement position correction device 30 of the movement position correction system 100 are the same as the configuration shown in Figure 1, so the explanation will be omitted.

[0102] As shown in FIG. 16, when the movement position correction device 30 starts its operation, first, the probe information input unit 31 receives the probe information from the probe vehicle 10 (step S410).

[0103] Next, the probe information is stored in the probe information storage unit 32 (step S411).

[0104] Next, the feature point analysis unit 33 detects characters in the image (step S412). Characters to be detected include characters and figures on road signs and billboards, as described above.

[0105] Next, the shape of the characters of the road in the image is stored by the feature point storage unit 34 (step S413). That is, the shape of the characters of the road is stored in association with the high-accuracy position information.

[0106] As shown in FIG. 17, when the movement position correction device 30 starts its operation, first, the probe information input unit 41 receives the probe information from the probe vehicle 20 (step S420).

[0107] Next, the probe information is stored in the probe information storage unit 42 (step S421).

[0108] Next, the feature point analysis unit 43 detects characters in the image (step S422).

[0109] Next, the shape of the characters in the image is stored by the feature point storage unit 44 (step S423). That is, the shape of the characters on the road is stored in association with the low-accuracy position information.

[0110] Next, the moving position correction unit 51 compares the shape of the characters linked to the high-precision position information obtained in the preliminary travel and corrects the vehicle position in the subsequent travel. That is, the shape of the characters linked to the high-precision position information obtained in the preliminary travel is compared with the shape of the characters obtained in the subsequent travel, and if they match, the vehicle position obtained in the subsequent travel is corrected using the high-precision position information obtained in the preliminary travel, thereby improving the position accuracy of the vehicle position in the subsequent travel.

[0111] <Fifth Embodiment> In a modified example of the first embodiment, as shown in FIG. 8, a probe vehicle 10L travels in the driving lane, and a probe vehicle 10R travels in the passing lane, and each vehicle acquires images. However, it is not possible to travel in the passing lane at all times, and it may not be possible to acquire images of all lanes.

[0112] The movement position correction system 100 of the fifth embodiment can identify whether the vehicle is traveling in the driving lane or the passing lane by traveling in only one lane during preliminary traveling.

[0113] FIG. 18 is a diagram that schematically illustrates the operation of identifying whether a vehicle is traveling in the driving lane or the passing lane simply by traveling in the driving lane. In FIG. 18, the central diagram shows a probe vehicle 10 traveling in the driving lane and an imaginary probe vehicle 10I traveling in the passing lane. The left diagram shows an image LD captured from the viewpoint of the probe vehicle 10, and the right diagram shows an image RDI generated from the image LD captured by the probe vehicle 10 from the viewpoint of the probe vehicle 10I on the imaginary passing lane. As the viewpoint changes, the position of the tower TW, which is a feature point on the image, changes.

[0114] Hereinafter, as a fifth embodiment, an operation for generating an image with a shifted viewpoint from an image obtained by traveling in only one lane during preliminary traveling will be described with reference to FIG.

[0115] Figure 19 is a flowchart explaining the operation flow of the movement position correction device 30 with respect to the pre-traveling vehicle. The operation flow of the probe vehicle 10 and the probe vehicle 20 is the same as the flowchart explained using Figures 4 and 5, and the flowchart explaining the operation flow of the movement position correction device 30 with respect to the post-traveling vehicle is the same as the flowchart explained using Figure 7, so explanation will be omitted. Furthermore, the configurations of the probe vehicle 10, the probe vehicle 20, and the movement position correction device 30 of the movement position correction system 100 are the same as the configuration shown in Figure 1, so explanation will be omitted.

[0116] As shown in FIG. 19, when the movement position correction device 30 starts its operation, first, the probe information input unit 31 receives the probe information from the probe vehicle 10 (step S510).

[0117] Next, the probe information is stored in the probe information storage unit 32 (step S511). Here, the probe vehicle 10 is equipped with cameras that can take images from multiple viewpoints, and the probe information includes images from multiple viewpoints.

[0118] Next, the feature point analysis unit 33 creates a three-dimensional (3D) model using images from multiple viewpoints taken from the probe vehicle 10 (step S512). To generate the 3D model, for example, Neural Radiance Fields (NeRF) can be used. NeRF can generate a free viewpoint image from multiple images taken from various angles. Similarly, photogrammetry technology can be used to generate a 3D model by photographing a subject from various angles.

[0119] Next, the feature point analysis unit 33 generates an image from the viewpoint of a different lane, for example, an express lane, by moving the viewpoint of the 3D model (step S513). That is, the appearance of features differs between the driving lane and the express lane, so an image of the vehicle while driving in the driving lane is actually captured, and the image of the vehicle while driving in the express lane is generated as a generated image from the image of the vehicle while driving in the driving lane. Then, the amount of displacement of the feature points (features) in the generated image is set as the amount of displacement of the features when driving in the express lane.

[0120] Next, the amount of displacement of the feature when traveling in the passing lane is stored in the feature point storage unit 34 together with the generated images from different viewpoints (step S514).

[0121] The moving position correction unit 51 can identify the driving lane of the subsequent traveling vehicle by comparing the amount of displacement of the feature when traveling in the overtaking lane stored in the feature point memory unit 34 with the amount of displacement of the feature points (features) of the probe vehicle 20 stored in the feature point memory unit 44.

[0122] As described above, according to the movement position correction system 100 of the fifth embodiment, it is possible to identify whether the vehicle is traveling in the driving lane or the passing lane simply by traveling in the driving lane. Therefore, even if the vehicle cannot always travel in the passing lane, the subsequent traveling vehicle can identify whether the vehicle is traveling in the driving lane or the passing lane.

[0123] In addition, since there is no need to drive on multiple lanes, driving distance can be reduced by up to 50% or more.

[0124] <Other application examples> In the movement position correction system of the present disclosure described above, an example has been shown in which the system is configured using vehicles such as probe vehicle 10 and probe vehicle 20, but this is not limited to this, and it is also possible to use, for example, a moving body such as a mobility robot for delivery.

[0125] Furthermore, as explained above, the movement position correction system of the present disclosure can be applied to infrastructure inspections such as roads, but it can also be applied to identifying lanes through high-precision positioning, for example, by companies that operate mobility robots for delivery, or by companies that maintain high-precision maps using the driving history of road operation companies.

[0126] It should be noted that the present disclosure allows modifications and omissions to be made to the embodiments as appropriate within the scope of the disclosure.

[0127] It should be noted that, within the scope of the present disclosure, the embodiments can be freely combined, modified, or omitted as appropriate.

[0128] The present disclosure described above will be summarized as an appendix.

[0129] (Appendix 1) a first moving object; a second moving object; a movement position correction device that communicates with the first and second moving bodies via a communication network, The first moving body is a first positioning device capable of sub-meter or centimeter level positioning; a first imaging device that captures a first image of the periphery of the first moving object; First position information including a positioning result by the first positioning device; and a function of acquiring first image information including the first image and transmitting the acquired first image information to the movement position correction device via the communication network; The second moving body is a second positioning device capable of meter-level positioning; a second imaging device that captures a second image of the periphery of the second moving object; a function of acquiring second position information including a positioning result by the second positioning device and second image information including the second image, and transmitting the second position information and the second image information to the movement position correction device via the communication network; The movement position correction device a first feature point analysis unit that detects a first feature point position in an image of a feature point of a first feature and a first displacement amount of the first feature point position in each image frame based on the first image information acquired from the first moving object; a feature point storage unit that stores the first feature point position and the first displacement amount of the first feature point position; a second feature point analysis unit that detects a second feature point position in an image of a feature point of a second feature and a second displacement amount of the second feature point position in each image frame based on the second image information acquired from the second moving object; a feature point storage unit that stores the second feature point position and the second displacement amount of the second feature point position; a movement position correction unit that compares the first feature point position and the first displacement amount of the first feature point position with the second feature point position and the second displacement amount of the second feature point position, and corrects the position of the second moving object that moves after the first moving object.

[0130] (Appendix 2) the first moving body and the second moving body travel along a plurality of parallel moving lanes; The first moving body is passing through the plurality of travel lanes ahead of the second moving body, and acquiring the first position information and the first image information in each travel lane; The movement position correction unit the first feature point position and the first displacement amount of the first feature point position are compared with the second feature point position and the second displacement amount of the second feature point position, and when the second feature point position and the second displacement amount and the first feature point position and the first displacement amount match, the travel lane through which the second moving body is traveling is identified among the multiple travel lanes.

[0131] (Appendix 3) the first moving body and the second moving body travel along a plurality of parallel moving lanes; The movement position correction unit A movement position correction system as described in Appendix 1, which identifies the movement lane in which the second moving body is traveling among the plurality of movement lanes by performing image recognition on other moving bodies in lanes adjacent to the movement lane in which the second moving body is traveling.

[0132] (Appendix 4) The first feature point analysis unit a function of detecting first road deterioration information from the first image; The first feature point storage unit storing the first location information and the first road deterioration information in association with each other; The second feature point analysis unit a function of detecting second road deterioration information from the second image; The second feature point storage unit storing the second location information and the second road deterioration information in association with each other; The movement position correction unit A moving position correction system as described in Appendix 1, which compares the first road deterioration information with the second road deterioration information, and if they match, uses the first position information to correct the position of the second moving body.

[0133] (Appendix 5) The first moving body is a first acceleration sensor, acquiring first road deterioration information from vertical acceleration information of the first moving body, and transmitting the first road deterioration information to the moving position correction device via the communication network; The second moving body is a second acceleration sensor, which acquires second road deterioration information from vertical acceleration information of the second moving body and transmits the second road deterioration information to the moving position correction device via the communication network; The first feature point analysis unit a function of detecting the first road deterioration information; The first feature point storage unit storing the first location information and the first road deterioration information in association with each other; The second feature point analysis unit a function of detecting the second road deterioration information; The second feature point storage unit storing the second location information and the second road deterioration information in association with each other; The movement position correction unit A moving position correction system as described in Appendix 1, which compares the first road deterioration information with the second road deterioration information, and if they match, uses the first position information to correct the position of the second moving body.

[0134] (Appendix 6) The first feature point storage unit storing the first road deterioration information by categorizing it by weather or season; The second feature point storage unit storing the second road deterioration information by categorizing it by weather or season; The movement position correction unit 6. The movement position correction system according to claim 4 or 5, wherein the first road deterioration information and the second road deterioration information are compared on a category basis.

[0135] (Appendix 7) The first feature point analysis unit a function of detecting a first character included in the first image; The first feature point storage unit storing the first position information and the first character shape in association with each other; The second feature point analysis unit a function of detecting a second character included in the second image; The second feature point storage unit storing the second position information and the second character shape in association with each other; The movement position correction unit A moving position correction system as described in Appendix 1, which compares the shape of the first character with the shape of the second character, and if they match, uses the first position information to correct the position of the second moving body.

[0136] (Appendix 8) the first moving body and the second moving body travel along a plurality of parallel moving lanes; The first moving body is passing through the plurality of travel lanes ahead of the second moving body, and acquiring the first position information and the first image information in each travel lane; The first feature point analysis unit a function of detecting a first character included in the first image; The first feature point storage unit storing the first position information and the first character shape in association with each other; The second feature point analysis unit a function of detecting a second character included in the second image; The second feature point storage unit storing the second position information and the second character shape in association with each other; The movement position correction unit A movement position correction system as described in Appendix 1, which compares the shape of the first character with the shape of the second character and identifies the movement lane through which the second moving body is traveling among the multiple movement lanes by finding a match.

[0137] (Appendix 9) the first moving body and the second moving body travel along a plurality of parallel moving lanes; the first image includes a multi-view image; The first feature point analysis unit a function of generating a three-dimensional model using the images from the multiple viewpoints, generating images from different viewpoints by moving the viewpoint of the three-dimensional model, and detecting the first displacement amount from the generated images; The movement position correction unit The movement position correction system of appendix 1, which compares the first displacement amount with the second displacement amount, and identifies the movement lane through which the second moving body is traveling among the plurality of movement lanes when the first displacement amount matches.

[0138] (Appendix 10) a first positioning device capable of sub-meter or centimeter level positioning; a first image capturing device configured to capture a first image of the surroundings; First position information including a positioning result by the first positioning device; a first feature point analysis unit that detects a first feature point position in an image of a feature point of a first feature and a first displacement amount of the first feature point position in each image frame based on first image information including the first image and the first image information acquired via a communication network from a first mobile object that acquires the first image; a feature point storage unit that stores the first feature point position and the first displacement amount of the first feature point position; a second positioning device capable of meter-level positioning; a second image capture device that captures a second image of the surroundings; Second location information including a positioning result by the second positioning device; a second feature point analysis unit that detects second feature point positions in the image of a feature point of a second feature and a second displacement amount of the second feature point positions in each image frame based on second image information including the second image and the second image information acquired via the communication network from a second mobile body that acquires the second image; a feature point storage unit that stores the second feature point position and the second displacement amount of the second feature point position; a movement position correction unit that compares the first feature point position and the first displacement amount of the first feature point position with the second feature point position and the second displacement amount of the second feature point position, and corrects the position of the second moving body that moves after the first moving body. [Explanation of symbols]

[0139] 10,20 Probe vehicle, 14,24 Imaging device, 16 High-precision locator device, 26 Locator device, 30 Movement position correction device, 33,43 Feature point analysis unit, 34,44 Feature point storage unit, 51 Movement position correction unit.

Claims

1. a first moving body; a second moving body; a movement position correction device that communicates with the first and second moving bodies via a communication network, The first moving body is a first positioning device capable of sub-meter or centimeter level positioning; a first imaging device that captures a first image of the periphery of the first moving object; First position information including a positioning result by the first positioning device; and a function of acquiring first image information including the first image and transmitting the acquired first image information to the movement position correction device via the communication network; The second moving body is a second positioning device capable of meter-level positioning; a second imaging device that captures a second image of the periphery of the second moving object, Second location information including a positioning result by the second positioning device; and second image information including the second image, and transmitting the second image information to the movement position correction device via the communication network; The movement position correction device a first feature point analysis unit that detects a first feature point position in an image of a feature point of a first feature and a first displacement amount of the first feature point position in each image frame based on the first image information acquired from the first moving object; a feature point storage unit that stores the first feature point position and the first displacement amount of the first feature point position; a second feature point analysis unit that detects a second feature point position in the image of a feature point of a second feature and a second displacement amount of the second feature point position in each image frame based on the second image information acquired from the second moving object; a feature point storage unit that stores the second feature point position and the second displacement amount of the second feature point position; a movement position correction unit that compares the first feature point position and the first displacement amount of the first feature point position with the second feature point position and the second displacement amount of the second feature point position, and corrects the position of the second moving object that moves after the first moving object.

2. the first moving body and the second moving body travel along a plurality of parallel moving lanes; The first moving body is passing through the plurality of travel lanes ahead of the second moving body, and acquiring the first position information and the first image information in each travel lane; The movement position correction unit 2. The movement position correction system according to claim 1, wherein the first feature point position and the first displacement amount of the first feature point position are compared with the second feature point position and the second displacement amount of the second feature point position, and when the second feature point position and the second displacement amount and the first feature point position and the first displacement amount match, the movement lane through which the second moving body is traveling is identified among the plurality of movement lanes.

3. the first moving body and the second moving body travel along a plurality of parallel moving lanes; The movement position correction unit 2. The movement position correction system according to claim 1, wherein the movement lane in which the second moving body is traveling is identified among the plurality of movement lanes by performing image recognition on other moving bodies in lanes adjacent to the movement lane in which the second moving body is traveling.

4. The first feature point analysis unit a function of detecting first road deterioration information from the first image; The first feature point storage unit storing the first location information and the first road deterioration information in association with each other; The second feature point analysis unit a function of detecting second road deterioration information from the second image; The second feature point storage unit storing the second location information and the second road deterioration information in association with each other; The movement position correction unit 2. The moving position correction system according to claim 1, wherein the first road deterioration information is compared with the second road deterioration information, and if they match, the first position information is used to correct the position of the second moving body.

5. The first moving body is a first acceleration sensor, acquiring first road deterioration information from vertical acceleration information of the first moving body, and transmitting the first road deterioration information to the moving position correction device via the communication network; The second moving body is a second acceleration sensor, which acquires second road deterioration information from vertical acceleration information of the second moving body, and transmits the second road deterioration information to the moving position correction device via the communication network; The first feature point analysis unit a function of detecting the first road deterioration information; The first feature point storage unit storing the first location information and the first road deterioration information in association with each other; The second feature point analysis unit a function of detecting the second road deterioration information; The second feature point storage unit storing the second location information and the second road deterioration information in association with each other; The movement position correction unit 2. The moving position correction system according to claim 1, wherein the first road deterioration information is compared with the second road deterioration information, and if they match, the first position information is used to correct the position of the second moving body.

6. The first feature point storage unit storing the first road deterioration information by categorizing it by weather or season; The second feature point storage unit storing the second road deterioration information by categorizing it by weather or season; The movement position correction unit 6. The movement position correction system according to claim 4, wherein the first road deterioration information and the second road deterioration information are compared on a category basis.

7. The first feature point analysis unit a function of detecting a first character included in the first image; The first feature point storage unit storing the first position information and the shape of the first character in association with each other; The second feature point analysis unit a function of detecting a second character included in the second image; The second feature point storage unit storing the second position information and the second character shape in association with each other; The movement position correction unit 2. The moving position correction system according to claim 1, wherein the shape of the first character is compared with the shape of the second character, and if they match, the position of the second moving body is corrected using the first position information.

8. the first moving body and the second moving body travel along a plurality of parallel moving lanes; The first moving body is passing through the plurality of travel lanes ahead of the second moving body, and acquiring the first position information and the first image information in each travel lane; The first feature point analysis unit a function of detecting a first character included in the first image; The first feature point storage unit storing the first position information and the shape of the first character in association with each other; The second feature point analysis unit a function of detecting a second character included in the second image; The second feature point storage unit storing the second position information and the second character shape in association with each other; The movement position correction unit The movement position correction system according to claim 1 , wherein the shape of the first character is compared with the shape of the second character, and the movement lane through which the second moving body is traveling is identified among the plurality of movement lanes by finding a match between the shape of the first character and the shape of the second character.

9. the first moving body and the second moving body travel along a plurality of parallel moving lanes; the first image includes a multi-view image; The first feature point analysis unit a function of generating a three-dimensional model using the images from the multiple viewpoints, generating images from different viewpoints by moving the viewpoints of the three-dimensional model, and detecting the first displacement amount from the generated images; The movement position correction unit 2. The movement position correction system according to claim 1, wherein the first displacement amount and the second displacement amount are compared, and when the first displacement amount matches, the movement lane through which the second moving body is traveling is identified among the plurality of movement lanes.

10. a first positioning device capable of sub-meter or centimeter level positioning; a first image capturing device configured to capture a first image of the surroundings; First position information including a positioning result by the first positioning device; a first feature point analysis unit that detects a first feature point position in an image of a feature point of a first feature and a first displacement amount of the first feature point position in each image frame based on first image information acquired via a communication network from a first mobile body that acquires the first image; a feature point storage unit that stores the first feature point position and the first displacement amount of the first feature point position; a second positioning device capable of meter-level positioning; a second image capture device that captures a second image of the surroundings; Second location information including a positioning result by the second positioning device; a second feature point analysis unit that detects second feature point positions in the image of a feature point of a second feature and a second displacement amount of the second feature point positions in each image frame based on second image information acquired via the communication network from a second mobile body that acquires the second image; a feature point storage unit that stores the second feature point position and the second displacement amount of the second feature point position; a movement position correction unit that compares the first feature point position and the first displacement amount of the first feature point position with the second feature point position and the second displacement amount of the second feature point position, and corrects the position of the second moving body that moves after the first moving body.

Images