Vehicle position estimation device and vehicle position estimation method

Abstract

To reliably extract a traffic lane line from an image and accurately determine which lane on a road a vehicle is traveling along.SOLUTION: A vehicle position estimation device 50 comprises: a line detection unit 503 for detecting a plurality of lines drawn on a road from an image obtained by capturing a prescribed imaging range including a road which the vehicle is traveling along; a filtering unit 504; and an own lane detection unit 505 for detecting the lane on a road which the vehicle is traveling along as the own lane. The filtering unit 504 extracts a plurality of lane boundary lines from among the plurality of lines on the basis of the interval between the plurality of lines in the image. The own lane detection unit 505 identifies a pair of lane boundary lines located to the left and right of the vehicle from the plurality of lane boundary lines depending on whether the inclination of the plurality of lane boundary lines in the image is positive or negative, and detects the own lane on the basis of the pair of lane boundary lines.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an apparatus and method for estimating the position of a vehicle traveling on a road. [Background technology]

[0002] Conventionally, a technology has been widely used that estimates the position of a vehicle traveling on a road based on measurement information from various sensors mounted on the vehicle, and uses the position information to navigate the vehicle. A representative method is a position estimation method that uses a technology called GNSS (Global Navigation Satellite System). That is, a sensor capable of receiving signals transmitted from an artificial satellite orbiting the Earth is mounted on the vehicle, and the vehicle's position can be estimated by calculating the position from the signal received using the sensor.

[0003] However, the vehicle position information estimated by the above-mentioned conventional method generally contains an error of about several meters, and depending on the usage of the position information, the accuracy may be insufficient. Therefore, various methods have been proposed to improve the accuracy of vehicle position estimation. For example, a technology is known that uses a camera to capture an image of the vehicle's surroundings, detects white and yellow lines (lane boundary lines) that divide each lane on the road using a method such as Hough transform, and determines which lane the vehicle is traveling on the road, and uses the determination result to improve the vehicle position accuracy (Patent Documents 1 and 2). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] U.S. Patent No. 7,295,682 [Patent Document 2] US Patent Application Publication No. 2003 / 0103650 Summary of the Invention [Problem to be solved by the invention]

[0005] With the technology of Patent Document 1, when there are many straight lines in an image, it is difficult to reliably extract lane boundary lines from them, resulting in a problem that it is not possible to accurately determine which lane on the road the vehicle is traveling in. [Means for solving the problem]

[0006] A vehicle position estimation device according to the present invention estimates the position on a road on which a vehicle is traveling, and comprises a straight line detection unit that detects a plurality of straight lines drawn on the road from an image obtained by photographing a predetermined shooting range including the road, a filtering unit that extracts lane boundaries separating each lane on the road from the plurality of straight lines detected by the straight line detection unit, and a self-lane detection unit that detects the lane on the road on which the vehicle is traveling as the own lane based on the lane boundaries extracted by the filtering unit, wherein the filtering unit extracts a plurality of lane boundaries from the plurality of straight lines based on the distance between the plurality of straight lines in the image, and the own lane detection unit identifies a pair of lane boundaries located on the left and right of the vehicle from the plurality of lane boundaries based on the positive or negative slope of the plurality of lane boundaries in the image, and detects the own lane based on the pair of lane boundaries. The vehicle position estimation method of the present invention is a method of estimating the position on a road on which a vehicle is traveling by a computer, in which the computer detects a plurality of straight lines drawn on the road from an image obtained by photographing a predetermined shooting range including the road, extracts lane boundary lines separating each lanes on the road from the plurality of straight lines based on the spacing between the plurality of straight lines in the image, identifies a pair of lane boundary lines located on the left and right of the vehicle from the plurality of lane boundary lines based on the positive or negative slope of the plurality of lane boundary lines in the image, and detects the lane on the road on which the vehicle is traveling as the vehicle's own lane based on the identified pair of lane boundary lines. Effect of the Invention

[0007] According to the present invention, lane boundary lines can be reliably extracted from an image, and it is possible to accurately determine in which lane on the road a vehicle is traveling. [Brief description of the drawings]

[0008] [Figure 1] 1 is a block diagram showing a configuration of a vehicle position estimation system including a vehicle position estimation device according to an embodiment of the present invention. [Diagram 2] FIG. 2 is a sequence diagram showing a processing flow of the vehicle position estimation system. [Diagram 3] 5 is a flowchart showing details of a filtering process in the first embodiment of the present invention. [Figure 4] FIG. 13 is a diagram showing an example of a straight-line interval table obtained in the filtering process. [Diagram 5] 3 is a diagram showing an example of an image transmitted from an in-vehicle device to a vehicle position estimation device in the first embodiment of the present invention. FIG. [Figure 6] FIG. 11 is a diagram showing an example of the inclination of each straight line. [Figure 7] 10 is a flowchart showing details of a process of detecting an own lane. [Figure 8] 10 is a flowchart showing details of a filtering process in the second embodiment of the present invention. [Figure 9] 13 is a diagram showing an example of an image transmitted from an in-vehicle device to a vehicle position estimation device in the second embodiment of the present invention. FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. For clarity of explanation, the following description and drawings have been omitted or simplified as appropriate. The present invention is not limited to this embodiment, and any application example that meets the concept of the present invention is included in the technical scope of the present invention. Unless otherwise specified, each component may be plural or singular.

[0010] (First embodiment) Fig. 1 is a block diagram showing a configuration of a vehicle position estimation system including a vehicle position estimation device according to an embodiment of the present invention. As shown in Fig. 1, the vehicle position estimation system 1 is configured by connecting one or more in-vehicle devices 40 and a vehicle position estimation device 50 to each other via a communication network 10. The communication network 10 includes, for example, the Internet, a LAN, a WAN, a public line network, a mobile phone network, and the like. Note that, although Fig. 1 shows an example in which the vehicle position estimation system 1 has two in-vehicle devices 40, the number of in-vehicle devices 40 is not limited to this.

[0011] The in-vehicle device 40 is an information device that is mounted on a vehicle and used, and includes a communication unit 401 , an image capturing unit 402 , a position detection unit 403 and a storage unit 404 .

[0012] The communication unit 401 communicates with the vehicle position estimation device 50 via the communication network 10. The communication unit 401 can transmit information about each vehicle acquired by the photographing unit 402 and the position detection unit 403 to the vehicle position estimation device 50 by connecting to the communication network 10 by, for example, wireless communication and communicating with the vehicle position estimation device 50.

[0013] The photographing unit 402 is configured using, for example, a camera, and photographs the surroundings (for example, the front) of each vehicle equipped with the in-vehicle device 40. The photographing range of the photographing unit 402 is set in advance in the in-vehicle device 40 so as to include at least the road on which the vehicle is traveling. As a result, an image photographed within a predetermined photographing range including the road on which each vehicle is traveling is acquired by the photographing unit 402. The image information acquired by the photographing unit 402 is transmitted to the vehicle position estimation device 50 by the communication unit 401.

[0014] The position detection unit 403 detects the position of each vehicle equipped with the in-vehicle device 40, for example, by using technology such as GNSS. Vehicle position information detected by the position detection unit 403 is transmitted to the vehicle position estimation device 50 by the communication unit 401. Note that the speed and traveling direction of each vehicle may be obtained from the history of the positions detected by the position detection unit 403 and used for navigation, various safety functions, and the like.

[0015] The storage unit 404 temporarily or permanently stores various information used in the in-vehicle device 40, such as image information acquired by the photographing unit 402 and vehicle position information detected by the position detection unit 403. The storage unit 404 can be configured using various storage media, such as an HDD, an SSD, or a memory card.

[0016] The vehicle position estimation device 50 is installed at a location separate from each in-vehicle device 40, and is configured using information devices such as a server or a PC (personal computer). The vehicle position estimation device 50 includes a communication unit 501, a preprocessing unit 502, a straight line detection unit 503, a filtering unit 504, a vehicle lane detection unit 505, a vehicle position estimation unit 506, a driving state extraction unit 507, and a storage unit 508.

[0017] The communication unit 501 communicates with each in-vehicle device 40 via the communication network 10. The communication unit 501 is connected to the communication network 10, for example, by wire, and communicates with each in-vehicle device 40, thereby being able to acquire image information obtained by capturing images of the surroundings of each vehicle by the imaging unit 402, and position information of each vehicle detected by the position detection unit 403.

[0018] The pre-processing unit 502 performs a predetermined pre-processing for removing noise, errors, and the like, on the position information and image information acquired by the communication unit 501. Using the image information pre-processed by the pre-processing unit 502, each process described below is performed by the line detection unit 503, the filtering unit 504, and the own lane detection unit 505. Note that each process by the line detection unit 503, etc. may be performed without performing the pre-processing by the pre-processing unit 502.

[0019] The line detection unit 503 detects straight lines drawn on the road on which each vehicle is traveling from the image information acquired by the communication unit 501. The line detection unit 503 can extract each straight line portion in the image using a known image processing technique such as a Hough transform. The straight lines detected by the line detection unit 503 include lane boundary lines for dividing each lane on the road. In the following description, the lane boundary lines include straight lines (solid lines or dashed lines) drawn between each lane on the road and a center line drawn between oncoming lanes.

[0020] The filtering unit 504 extracts lane boundary lines that separate the lanes on the road from among the multiple straight lines detected by the straight line detection unit 503. The processing by the filtering unit 504 makes it possible to extract correct lane boundary lines from among the various straight lines detected from the image by the straight line detection unit 503. Details of the processing by the filtering unit 504 will be described later.

[0021] The own lane detection unit 505 detects the lane on the road in which each vehicle is traveling as the own lane of the vehicle based on the lane boundary lines extracted by the filtering unit 504. Here, for each vehicle equipped with the on-vehicle device 40, the own lane is detected by determining in which lane on the road the vehicle is located at predetermined time intervals. The process by the own lane detection unit 505 will be described in detail later.

[0022] The vehicle position estimation unit 506 estimates the position and traveling direction of each vehicle on the road based on the position information of each vehicle acquired by the communication unit 501 and the detection result of the own lane of each vehicle by the own lane detection unit 505. The vehicle position estimation unit 506 corrects the position information of each vehicle according to the detection result of the own lane, for example, and estimates the traveling direction of each vehicle according to the traveling direction of the own lane. This makes it possible to accurately estimate the traveling position of each vehicle on the road from the vehicle position detected by the in-vehicle device 40 of each vehicle.

[0023] The driving state extraction unit 507 extracts various indices related to the driving state of each vehicle using the estimation results of the position and traveling direction of each vehicle by the vehicle position estimation unit 506. The driving state extraction unit 507 extracts indices such as fuel efficiency, safety, and comfort of each vehicle, and transmits information based on the extraction results to the in-vehicle device 40 of each vehicle to provide advice and warnings to the driver of the vehicle. In addition to the information obtained by the vehicle position estimation unit 506, the processing of the driving state extraction unit 507 may use other sensor information and the like.

[0024] The storage unit 508 temporarily or permanently stores various information used in the vehicle position estimation device 50, such as position information and image information acquired by the communication unit 501 from each in-vehicle device 40, position information of each vehicle estimated by the vehicle position estimation unit 506, and various indicators extracted by the driving state extraction unit 507. The storage unit 508 can be configured using various storage media, such as an HDD, an SSD, or a memory card.

[0025] It should be noted that some or all of the functions of the vehicle position estimation device 50 described above may be mounted on the in-vehicle device 40. That is, each in-vehicle device 40 can detect the vehicle's own lane based on image information acquired by itself, and estimate the vehicle's position and traveling direction based on the detection result.

[0026] Fig. 2 is a sequence diagram showing a flow of processing in the vehicle position estimation system 1. In the vehicle position estimation system 1, each in-vehicle device 40 and the vehicle position estimation device 50 performs the processing shown in the sequence diagram of Fig. 2, for example, at each predetermined processing cycle.

[0027] The in-vehicle device 40 acquires image information of a predetermined shooting range by the shooting unit 402 (step S41), and acquires vehicle position information by the position detection unit 403 (step S42). Then, the acquired information is transmitted to the vehicle position estimation device 50 by the communication unit 401 (step S43).

[0028] When the vehicle position estimation device 50 receives image information and position information transmitted from the in-vehicle device 40, the preprocessing unit 502 performs preprocessing on the received image information (step S51), and then the line detection unit 503 detects straight lines that exist within the shooting range from the preprocessed image (step S52).Then, the filtering unit 504 filters the detected straight lines (step S53) to extract lane boundary lines, and the own lane detection unit 505 detects the own lane using the extracted lane boundary lines (step S54).

[0029] If the vehicle's own lane is detected in step S54, the vehicle position estimation unit 506 estimates the vehicle's position on the road using the detection result and the position information transmitted from the in-vehicle device 40 (step S55), and the driving state extraction unit 507 extracts an index of the driving state based on the estimation history of the vehicle position (step S56). Then, the communication unit 501 transmits the extracted index of the driving state to the in-vehicle device 40 (step S57).

[0030] Upon receiving the driving state indicator transmitted from the vehicle position estimation device 50, the in-vehicle device 40 uses the received driving state indicator to provide information to the driver (step S44). Here, for example, advice and warnings regarding fuel efficiency, safety, comfort, etc. are given to the driver of the vehicle equipped with the in-vehicle device 40 using a display device and / or audio output device (not shown).

[0031] FIG. 3 is a flowchart showing details of the filtering process performed by the filtering unit 504 in step S53 of FIG. 2 in the first embodiment of the present invention.

[0032] In step S701, each straight line detected by the straight line detection unit 503 in step S52 in Fig. 2 is interpolated to the top of a region (ROI: Region Of Interest) set in advance in the image. Here, each straight line that does not reach the top of the ROI among the multiple straight lines detected by the straight line detection unit 503 is extended to be interpolated to the top of the ROI. The ROI is an area in the image corresponding to the road on which the vehicle is traveling, and is set in advance in the filtering unit 504 in accordance with the installation position, shooting direction, shooting range, etc. of the shooting unit 402 in the vehicle equipped with the in-vehicle device 40.

[0033] In step S702, the X-coordinate value of the top of the ROI is obtained for each straight line detected by line detection unit 503 in step S52 in Fig. 2 and further interpolated in step S701 as necessary. Here, for example, the intersection points between the top (upper side) of the ROI set in the image and each straight line are found, and the X-coordinate value (horizontal coordinate value in the image) of each intersection point is obtained. Note that the top of the ROI from which the X-coordinate value is obtained corresponds to the part of the road located farthest in the image as seen from the vehicle, i.e., the position of the horizon.

[0034] In step S703, the X coordinate value of each line acquired in step S702 is stored in a preset list. Note that the list used here is a list in which X coordinate values ​​are temporarily stored to calculate the interval between each line in the process of step S709 described below, and is empty at the start of step S703.

[0035] In step S704, the number of X coordinate values ​​stored in the list in step S703 is acquired as the number N stored in the list.

[0036] In step S705, a counter i is initialized to i = 1. This counter i is a variable for specifying one of two straight lines between which an interval is calculated in the process of step S709 described later, and is counted up by one from the initial value 1.

[0037] In step S706, it is determined whether the current value of counter i is smaller than the number of items stored in the list N set in step S704. If the value of counter i is smaller than the number of items stored N, the process proceeds to step S707, and if it is equal to or greater than the number of items stored N, the process proceeds to step S713.

[0038] In step S707, a counter j is initialized to j = i + 1. This counter j is a variable for specifying the other of the two straight lines between which the interval is calculated in the process of step S709 described later, and is counted up by one from the initial value i + 1.

[0039] In step S708, it is determined whether the current value of counter j is equal to or less than the number of items stored in the list set in step S704, N. If the value of counter j is equal to or less than the number of items stored in the list, N, the process proceeds to step S709. If the value of counter j is greater than the number of items stored in the list, the process proceeds to step S712.

[0040] In step S709, the distance between the i-th straight line Li corresponding to the current counter i value and the j-th straight line Lj corresponding to the current counter j value is calculated. Here, the X-coordinate values ​​of the straight lines Li and Lj are respectively acquired from the X-coordinate values ​​stored in the list in step S703, and the difference between them is calculated, so that the distance between the straight lines Li and Lj can be calculated. Note that the straight line Ln (n=1 to N) is the n-th straight line counted from a predetermined reference position (e.g., the left end) in the image among the straight lines detected by the straight line detection unit 503 in step S52 in FIG. 2. Since the X-coordinate values ​​of each straight line at the top of the ROI are stored in the list in this order, it is possible to calculate the distance between the straight lines Li and Lj by finding the difference between the X-coordinate values ​​stored in the j-th and i-th positions in the list.

[0041] In step S710, the calculation result of the interval between the straight lines Li and Lj obtained in step S709 is recorded in the corresponding field of a predetermined straight line interval table. The straight line interval table is a table for temporarily recording the interval between each straight line, and a corresponding field is set for each combination of two straight lines. The straight line interval table is set according to the number N of stored X coordinate values ​​in the above-mentioned list, for example, when the process of step S710 is executed for the first time.

[0042] In step S711, the value of counter j is incremented by 1. After step S711 is executed, the process returns to step S708, and the above-mentioned processing is repeated.

[0043] In step S712, the value of counter i is incremented by 1. After step S712 is executed, the process returns to step S706, and the above-mentioned processing is repeated.

[0044] The above-described processes of steps S706 to S710 are repeatedly performed while counting up the values ​​of counters i and j by 1 each, until it is determined in step S706 that the value of counter i is equal to or greater than the number of stored values ​​N. As a result, for all combinations of lines detected by line detection unit 503 in step S52 of Fig. 2, line intervals at the top of the ROI are calculated and recorded in the line interval table.

[0045] If it is determined in step S706 that the value of counter i is equal to or greater than the number of stored values ​​N, then in step S713, the mode of the line intervals in the line interval table is obtained. Here, the mode can be obtained by identifying the line interval value recorded in each column of the line interval table that has the largest number of columns in which the same value is recorded. Note that a predetermined error range may be set at this time, and if the difference between line interval values ​​recorded in different columns in the line interval table falls within that error range, the values ​​are considered to be the same.

[0046] In step S714, the lines detected by the line detection unit 503 are filtered using the mode acquired in step S713. Here, the combination of lines whose line spacing is the mode is kept, while the combination of lines whose line spacing is not the mode is excluded from the following processing. This makes it possible to select the lines detected by the line detection unit 503 based on the spacing between the lines in the image. In other words, by utilizing the fact that the spacing between lane boundary lines is the same at the top of the ROI, the combination of lines whose spacing is different from the others can be regarded as not being a lane boundary line and can be excluded from the processing when detecting the own lane.

[0047] In step S715, the inclination of each straight line within the image is calculated for each straight line that was not excluded in step S714. Here, the inclination of each straight line can be calculated by determining the angle of the extension direction of each straight line within the image.

[0048] In step S716, for each straight line whose slope was calculated in step S715, it is determined whether or not the slope is within a predetermined range assumed to be a lane boundary line. If the slope is within the predetermined range, the process proceeds to step S717, and the straight line is set as a lane boundary line, whereas if the slope is outside the predetermined range, the process proceeds to step S718, and the straight line is determined not to be a lane boundary line and is excluded from subsequent processing. This makes it possible to exclude straight lines with slopes that cannot be lane boundary lines and extract lane boundary lines from the multiple straight lines detected in the image.

[0049] After performing the process of step S717 or step S718 for all straight lines, the filtering unit 504 ends the filtering process of FIG.

[0050] Fig. 4 is a diagram showing an example of a line interval table obtained in the filtering process described above. The value stored in each column of the table in Fig. 4 represents the interval (pixel value) between two lines corresponding to that column among six lines L1 to L6 detected from the image. Note that columns marked with "NA" represent that there is no line combination corresponding to that column, or that the combination overlaps with a combination in another column, and therefore is not applicable in the table in Fig. 4.

[0051] In the example of the line interval table in Fig. 4, the line interval values ​​stored in the three corresponding columns for the combination of lines L1 and L3, the combination of lines L3 and L5, and the combination of lines L5 and L6 are the same and are the most frequent values, so the lines L1, L3, L5, and L6 are extracted as lane boundary lines, and the other lines L2 and L4 are excluded from processing.

[0052] 5 is a diagram showing an example of an image transmitted from the in-vehicle device 40 to the vehicle position estimation device 50 in the first embodiment of the present invention. This image includes lane boundary lines 801, 803, 805, and 806 drawn on the road ahead of the vehicle. In this embodiment, the filtering process performed by the filtering unit 504 extracts these lane boundary lines as the straight lines L1, L3, L5, and L6 in FIG. 4, respectively, thereby excluding straight lines that do not correspond to lane boundary lines from an image including various straight line shapes, and thus making it possible to reliably extract the lane boundary lines 801, 803, 805, and 806.

[0053] Fig. 6 is a diagram showing an example of the inclination of each straight line calculated in step S715 of Fig. 3 for the above-mentioned straight lines L1, L3, L5, and L6 corresponding to the lane boundary lines 801, 803, 805, and 806 in Fig. 5, respectively. Fig. 6 shows the inclination of each straight line forming the boundary line of these lanes K1 to K3 in the image for the lane K1 existing between the straight line L1 (lane boundary line 801) and the straight line L3 (lane boundary line 803) on the road, the lane K2 existing between the straight line L3 and the straight line L5 (lane boundary line 805), and the lane K3 existing between the straight line L5 and the straight line L6 (lane boundary line 806). Note that in Fig. 6, when the sign of the inclination is positive, it indicates that the extension direction of the straight line rises to the right, and conversely, when the sign of the inclination is positive, it indicates that the extension direction of the straight line falls to the right.

[0054] In step S54 in FIG. 3, information on the slope of each straight line as shown in FIG. 6 obtained by filtering unit 504 is input to current lane detection unit 505, and using this information, current lane detection unit 505 performs current lane detection processing.

[0055] FIG. 7 is a flowchart showing the details of the own lane detection process performed by the own lane detection unit 505 in step S54 of FIG.

[0056] In step S1001, each straight line after the filtering process performed by the filtering unit 504 in step S53 in Fig. 2 is input. Here, each straight line set as a lane boundary line in step S717 is acquired as a target for subsequent processing.

[0057] In step S1002, the slope of each straight line input in step S1001 is obtained. Here, the slope value calculated in step S715 of Fig. 3 for each input straight line is obtained.

[0058] In step S1003, the slope of each line acquired in step S1002 is stored in a preset list. Note that the list used here is a list in which the slope value of each line is temporarily stored in order to perform the determination process in step S1007 described later, and is empty at the start of step S1003.

[0059] In step S1004, the number of inclinations stored in the list in step S1003 is acquired as the number M stored in the list.

[0060] In step S1005, a counter i is initialized to i = 1. This counter i is a variable for specifying the slopes of two straight lines to be determined in the determination process in step S1007 described below, and is counted up by one from the initial value 1.

[0061] In step S1006, it is determined whether the current value of counter i is equal to or less than the number of items stored in the list, M, which was set in step S1004. If the value of counter i is equal to or less than the number of items stored in the list, the process proceeds to step S1007. If the value of counter i is greater than the number of items stored, M, the process in FIG. 7 ends.

[0062] In step S1007, it is determined whether the slope of the i-th line corresponding to the current value of counter i is a positive value and the slope of the i+1-th line is a negative value. Here, the slope values ​​stored in the i-th and i+1-th lines in the list in step S1003 are obtained, and the signs of these values ​​are determined, so that it is possible to determine whether the slope values ​​of the i-th and i+1-th lines are positive or negative. If at least one of the conditions that the slope of the i-th line is a positive value and the slope of the i+1-th line is a negative value is not met, the process proceeds to step S1008, and if both are met, the process proceeds to step S1009.

[0063] In step S1008, the value of counter i is incremented by 1. After step S1008 is executed, the process returns to step S1006, and the above-mentioned processing is repeated.

[0064] If it is determined in step S1007 that both of the above two conditions are satisfied, then in step S1009, the i-th lane Ki corresponding to the current value of counter i is set as the own lane. That is, the i-th and (i+1)-th straight lines in the image are identified as a pair of lane boundary lines that satisfy the determination condition in step S1007, and the lane Ki between these lane boundary lines is detected as the lane in which the own vehicle is currently traveling. This makes it possible to detect the own lane based on the positive and negative inclinations of multiple lane boundary lines in the image.

[0065] In step S1010, the width of the own vehicle lane set in step S1009 is calculated. Here, the value of the line interval corresponding to the combination of the i-th line and the (i+1)-th line whose slope is determined to satisfy the condition in step S1007 among the line intervals calculated in step S709 of Fig. 3 and recorded in the line interval table in step S710 is obtained, and the width of the own vehicle lane can be calculated based on the value of the line interval. That is, the actual width of the own vehicle lane on the road can be calculated by multiplying the obtained line interval value by a predetermined coefficient according to the shooting magnification of the shooting unit 402, etc.

[0066] Once the own lane is set in step S1009 and the width of the own lane is calculated in step S1010, own lane detection unit 505 ends the own lane detection process of FIG.

[0067] According to the first embodiment of the present invention described above, the following advantageous effects are achieved.

[0068] (1) The vehicle position estimation device 50 includes a line detection unit 503 that detects a plurality of straight lines drawn on a road from an image obtained by photographing a predetermined photographing range including the road on which the vehicle is traveling, a filtering unit 504 that extracts lane boundaries dividing each lane on the road from the plurality of straight lines detected by the line detection unit 503, and an own lane detection unit 505 that detects the lane on the road on which the vehicle is traveling as the own lane based on the lane boundaries extracted by the filtering unit 504. The filtering unit 504 extracts a plurality of lane boundaries from the plurality of straight lines based on the intervals between the plurality of straight lines in the image (steps S713 to S717). The own lane detection unit 505 identifies a pair of lane boundaries located on the left and right of the vehicle from the plurality of lane boundaries based on the positive and negative inclinations of the plurality of lane boundaries in the image (step S1007), and detects the own lane based on the pair of lane boundaries (step S1009). In this way, lane boundary lines can be reliably extracted from the image, and it is possible to accurately determine in which lane on the road the vehicle is traveling.

[0069] (2) The filtering unit 504 extracts lane boundary lines based on the distance between multiple straight lines at the top of the ROI, which is a region that is preset in the image corresponding to a road. Specifically, the distance is calculated for each combination of straight lines (step S709), and combinations of multiple straight lines with the same distance are extracted as lane boundary lines for each lane (steps S713 and S714). In this way, lane boundary lines can be accurately extracted by utilizing the fact that the distances between lane boundary lines are consistent at the top of the ROI.

[0070] (3) The filtering unit 504 outputs the calculated interval value for each lane boundary line. The own lane detection unit 505 calculates the width of the own lane based on the interval value output from the filtering unit 504 (step S1010). In this way, the width of the own lane can be calculated accurately.

[0071] (4) The filtering unit 504 interpolates the straight lines that do not reach the top of the ROI in the image to the top (step S701) and calculates the line intervals. By doing so, it is possible to calculate the intervals between the straight lines that do not reach the top of the ROI in the image and other straight lines.

[0072] Second Embodiment Next, a second embodiment of the present invention will be described. In the above-mentioned first embodiment, an example was described in which the intervals between the lane boundary lines match at the top of the ROI when the vehicle is traveling on a straight road, and the X-coordinate values ​​of the lines at the top of the ROI are used to calculate the intervals between the lines. However, when the vehicle is traveling on a curved road, the intervals between the lane boundary lines do not necessarily match at the top of the ROI, so the method described in the first embodiment cannot be applied. Therefore, in this embodiment, an example is described in which the X-coordinate values ​​of the lines at the bottom of the ROI, rather than at the top of the ROI, are used to calculate the intervals between the lines.

[0073] FIG. 8 is a flowchart showing details of the filtering process performed by the filtering unit 504 in step S53 of FIG. 2 in the second embodiment of the present invention.

[0074] In step S1101, each straight line detected by the straight line detection unit 503 in step S52 in FIG. 2 is interpolated to the bottom of the ROI, which is a region set in advance within the image.

[0075] In step S1102, the X coordinate value at the bottom of the ROI is obtained for each line detected by line detection unit 503 in step S52 in Fig. 2 and further interpolated in step S1101 as necessary. Here, the intersections between the bottom (bottom side) of the ROI set in the image and each line are obtained, and the X coordinate value (horizontal coordinate value in the image) of each intersection is obtained, for example, by a procedure similar to that of step S702 in Fig. 3 described in the first embodiment.

[0076] In steps S1103 and S1104, the interval between each line is calculated based on the X coordinate value of each line acquired in step S1102, and the calculation result is recorded in the line interval table. Here, the interval between each line is calculated and recorded in the line interval table, for example, by a procedure similar to steps S703 to S712 in FIG.

[0077] In step S1105, the most frequent value of the line interval in the line interval table is obtained. Here, the same process as in step S713 in FIG.

[0078] In step S1106, the lines detected by line detection unit 503 are filtered using the mode value acquired in step S1105. Here, the same process as in step S714 in Fig. 3 is performed. This makes it possible to extract lane boundary lines from the multiple lines detected in the image. After performing the process in step S1106, the same processes as in steps S715 to S718 in Fig. 3 may be performed as necessary.

[0079] After performing the process of step S1106, the filtering unit 504 ends the filtering process of FIG.

[0080] 9 is a diagram showing an example of an image transmitted from the in-vehicle device 40 to the vehicle position estimation device 50 in the second embodiment of the present invention. This image is obtained by photographing the road ahead of the vehicle while the vehicle is traveling on a curved road, and includes lane boundary lines 901, 902, 903, and 904 drawn on the road. In this embodiment, these lane boundary lines are extracted from the image by the filtering process of FIG. 8 performed by the filtering unit 504.

[0081] According to the second embodiment of the present invention described above, the filtering unit 504 extracts lane boundary lines based on the distance between multiple straight lines at the bottom of an ROI, which is a region that is preset in an image corresponding to a road. In this way, even if the vehicle is traveling on a curved road and therefore the distance between lane boundary lines does not match at the top of the ROI, the lane boundary lines can be extracted.

[0082] The present invention is not limited to the above-described embodiment, and can be implemented using any components without departing from the scope of the present invention. The above-described embodiments and modifications are merely examples, and the present invention is not limited to these contents as long as the features of the invention are not impaired. In addition, although various embodiments and modifications have been described above, the present invention is not limited to these contents. Other aspects conceivable within the scope of the technical idea of ​​the present invention are also included in the scope of the present invention. [Explanation of symbols]

[0083] 1...vehicle position estimation system, 10...communication network, 40...vehicle-mounted device, 50...vehicle position estimation device, 401...communication unit, 402...imaging unit, 403...position detection unit, 404...storage unit, 501...communication unit, 502...preprocessing unit, 503...straight line detection unit, 504...filtering unit, 505...own lane detection unit, 506...vehicle position estimation unit, 507...driving state extraction unit, 508...storage unit

Claims

1. A vehicle position estimation device that estimates a position on a road on which a vehicle is traveling, comprising: a straight line detection unit that detects a plurality of straight lines drawn on the road from an image obtained by photographing a predetermined photographing range including the road; a filtering unit that extracts lane boundary lines that separate lanes on the road from the plurality of straight lines detected by the straight line detection unit; a current lane detection unit that detects, based on the lane boundary lines extracted by the filtering unit, a lane on the road in which the vehicle is traveling as a current lane, The filtering unit extracts a plurality of the lane boundary lines from the plurality of straight lines based on intervals between the plurality of straight lines in the image, The vehicle position estimation device, wherein the lane detection unit identifies a pair of lane boundary lines located on the left and right sides of the vehicle from the multiple lane boundary lines based on the positive and negative inclinations of the multiple lane boundary lines in the image, and detects the vehicle's lane based on the pair of lane boundary lines.

2. 2. The vehicle position estimation device according to claim 1, A vehicle position estimation device, wherein the filtering unit extracts the lane boundary lines based on the intervals between the plurality of straight lines at the top or bottom of a region that is preset in the image corresponding to the road.

3. 3. The vehicle position estimation device according to claim 2, The filtering unit calculates the interval for each combination of the straight lines, and extracts a combination of a plurality of the straight lines having the same interval as the lane boundary line for each lane.

4. 4. The vehicle position estimation device according to claim 3, The filtering unit outputs the interval value calculated for each lane boundary line, The vehicle position estimation device, wherein the own lane detection unit calculates a width of the own lane based on the interval value output from the filtering unit.

5. 3. The vehicle position estimation device according to claim 2, The filtering unit calculates the interval by interpolating, among the plurality of straight lines, straight lines that do not reach the top or bottom in the image to the top or bottom.

6. 1. A method for estimating, by a computer, a position on a road on which a vehicle is traveling, comprising: The computer Detecting a plurality of straight lines drawn on the road from an image obtained by photographing a predetermined photographing range including the road; extracting lane boundary lines that separate lanes on the road from among the straight lines based on intervals between the straight lines in the image; identifying a pair of lane markings located on the left and right of the vehicle from the plurality of lane markings based on positive and negative inclinations of the plurality of lane markings in the image; A vehicle position estimation method, comprising: detecting a lane on the road in which the vehicle is traveling as the vehicle's own lane based on the identified pair of lane boundary lines.

Images