WO2015025555A1

WO2015025555A1 - Traffic volume measurement device and traffic volume measurement method

Info

Publication number: WO2015025555A1
Application number: PCT/JP2014/059862
Authority: WO
Inventors: 孝光渡辺; 渡辺　孝弘
Original assignee: 沖電気工業株式会社
Priority date: 2013-08-21
Filing date: 2014-04-03
Publication date: 2015-02-26
Also published as: JP5783211B2; JP2015041187A

Abstract

[Problem] To provide a technology whereby precision of traffic volume measurement is improved when measuring traffic volume on the basis of a captured image. [Solution] Provided is a traffic volume measurement device (10), comprising: an information acquisition unit (111) which acquires a captured image whereupon a road surface is image captured and an image capture time; a location detection unit (114) which detects a prescribed detection location on the basis of a vehicle region which is extracted from the captured image; and a measurement unit (115) which measures, as a traffic volume, a number of vote peaks which are obtained with regard to combinations of detection locations and image capture times.

Description

Traffic volume measuring apparatus and traffic volume measuring method

The present invention relates to a traffic volume measuring device and a traffic volume measuring method.

In recent years, a technique for measuring traffic volume from a captured image obtained by capturing a road plane with a camera has been developed. In such traffic measurement technology, the vehicle silhouette area is extracted from the captured image based on the background difference method, and the traffic volume is measured by counting the number of vehicle silhouette areas that have passed through the imaging range while tracking the vehicle silhouette area. It is common (see, for example, Patent Document 1).

JP-A-7-262489

However, if the vehicle silhouette area cannot be stably extracted from the captured image based on the background subtraction method, it may be difficult to accurately track the vehicle silhouette area. In such a situation, it is difficult to accurately count the number of vehicle silhouette regions that have passed through the imaging range, and it may be difficult to accurately measure the traffic volume.

For example, when the camera is installed outdoors, there is a possibility that the vehicle silhouette region cannot be stably extracted due to a change in weather or a change in sunlight. In addition, when a phenomenon occurs in which the vehicle on the back side is concealed in the vehicle on the near side, a situation may occur where the vehicle silhouette area overlaps between the vehicle on the near side and the vehicle on the back side. There is a possibility that the vehicle silhouette region cannot be extracted stably.

Therefore, the present invention has been made in view of the above problems, and an object of the present invention is to provide a technique for improving the accuracy of traffic volume measurement when measuring traffic volume based on a captured image. There is.

In order to solve the above problem, according to an aspect of the present invention, an information acquisition unit that acquires a captured image obtained by capturing a road plane and an imaging time, and a predetermined area based on a vehicle area extracted from the captured image. A traffic volume measuring device, comprising: a position detection section that detects a detection position of the camera; and a measurement section that measures traffic volume based on the number of voting peaks obtained for the combination of the detection position and the imaging time. Provided.

The measurement unit may perform a Hough transform on the combination of the detection position and the imaging time, and measure the number of voting peak points as a traffic volume in the Hough space.

The position detection unit may detect an edge feature from the captured image and extract the vehicle region based on the edge feature.

The measuring unit may specify the vehicle speed based on the position of the voting peak point.

The position detection unit may detect, as the detection position, coordinates on a vehicle travel axis in a real space set in accordance with an actual size.

The position detection unit detects a predetermined vertical plane orthogonal to the vehicle travel axis in real space based on the vehicle region, and detects an intersection coordinate between the vehicle travel axis and the vertical plane as the detection position. Good.

The position detection unit may detect a vehicle front surface or a vehicle back surface as the predetermined vertical plane.

The position detection unit detects the front surface of the vehicle as the vertical plane when the vehicle traveling direction in the real space is from the back to the front, and the vehicle traveling direction in the real space is the direction from the front to the back. Alternatively, the rear surface of the vehicle may be detected as the vertical plane.

Further, according to an aspect of the present invention, a step of acquiring a captured image obtained by capturing a road plane and a capturing time, a step of detecting a predetermined detection position based on a vehicle region extracted from the captured image, And measuring the traffic volume based on the peak point of voting obtained for the combination of the detection position and the imaging time, a traffic volume measuring method is provided.

As described above, according to the present invention, it is possible to improve the accuracy of traffic volume measurement when traffic volume is measured based on a captured image.

It is a figure for demonstrating the outline | summary of embodiment of this invention. It is a figure which shows the function structural example of the traffic measuring device which concerns on embodiment of this invention. It is a figure which shows the example of the parameter used by the setting part. It is a figure for demonstrating the example of the function of a setting part. It is a figure for demonstrating the example of the function of a position detection part. It is a figure which shows the example of the combination of a detection position and imaging time. It is a figure which shows the result obtained by performing a Hough transformation with respect to the example of the combination of the detection position and imaging time shown in FIG. It is a flowchart which shows the operation example of the traffic measuring device which concerns on embodiment of this invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the present specification and drawings, components having substantially the same functional configuration are denoted by the same reference numerals, and redundant description is omitted.

In the present specification and drawings, a plurality of constituent elements having substantially the same functional configuration may be distinguished by attaching different alphabets or numbers after the same reference numeral. However, when it is not necessary to particularly distinguish each of a plurality of constituent elements having substantially the same functional configuration, only the same reference numerals are given.

[Description of overview]
Subsequently, an outline of an embodiment of the present invention will be described. FIG. 1 is a diagram for explaining an outline of an embodiment of the present invention. As shown in FIG. 1, a traffic volume measuring device 10 incorporating an imaging unit and a road plane exist in real space. Further, the traffic volume measuring device 10 incorporating the imaging unit is installed in a state where the imaging direction is directed to the road plane. The captured image Img ′ captured by the traffic measuring device 10 shows the boundary line of the lane provided on the road. Further, as shown in FIG. 1, the center of the lens of the traffic measuring device 10 is set to the origin O.

FIG. 1 shows an example in which an imaging unit is incorporated in the traffic measurement device 10, but the imaging unit is not incorporated in the traffic measurement device 10 and is installed outside the traffic measurement device 10. It may be. In such a case, for example, the traffic measurement device 10 may acquire the captured image Img ′ by receiving the captured image Img ′ transmitted from the imaging unit. Further, for example, the traffic volume measuring device 10 may acquire the captured image Img ′ by reading the captured image Img ′ recorded on the recording medium.

Here, a technique for measuring traffic volume from a captured image Img ′ obtained by imaging a road plane by an imaging unit has been proposed. In such traffic measurement technology, a vehicle silhouette area is extracted from the captured image Img ′ based on the background difference method, and the number of vehicle silhouette areas that have passed through the imaging range is counted while tracking the vehicle silhouette area. Is generally measured.

However, when the vehicle silhouette region cannot be stably extracted from the captured image Img ′ based on the background subtraction method, a situation in which it is difficult to accurately track the vehicle silhouette region may occur. In such a situation, it is difficult to accurately count the number of vehicle silhouette regions that have passed through the imaging range, and it may be difficult to accurately measure the traffic volume.

For example, when the imaging unit is installed outdoors, there is a possibility that the vehicle silhouette region cannot be stably extracted due to a change in weather or a change in sunlight. In addition, when a phenomenon occurs in which the vehicle on the back side is concealed in the vehicle on the near side, a situation may occur where the vehicle silhouette area overlaps between the vehicle on the near side and the vehicle on the back side. There is a possibility that the vehicle silhouette region cannot be extracted stably.

Therefore, in this specification, a technique for improving the accuracy of traffic volume measurement when measuring traffic volume based on captured images is proposed.

The outline of the embodiment of the present invention has been described above.

[Details of the embodiment]
Next, details of the embodiment of the present invention will be described. First, the functional configuration of the traffic volume measuring device 10 according to the embodiment of the present invention will be described. FIG. 2 is a diagram illustrating a functional configuration example of the traffic volume measuring device 10 according to the embodiment of the present invention. As shown in FIG. 2, the traffic volume measuring device 10 according to the embodiment of the present invention includes a control unit 110, an imaging unit 170, a storage unit 180, and an output unit 190.

The control unit 110 has a function of controlling the entire operation of the traffic measuring device 10. The imaging unit 170 has a function of acquiring a captured image by imaging a real space, and is configured by, for example, a monocular camera. The storage unit 180 can store a program and data for operating the control unit 110. In addition, the storage unit 180 can temporarily store various data necessary in the course of the operation of the control unit 110. The output unit 190 has a function of performing output in accordance with control by the control unit 110. The type of the output unit 190 is not particularly limited, and may be a measurement result recording device, a device that transmits a measurement result to another device via a communication line, or a display device. An audio output device may be used.

In the example illustrated in FIG. 2, the imaging unit 170, the storage unit 180, and the output unit 190 exist inside the traffic measurement device 10, but all or one of the imaging unit 170, the storage unit 180, and the output unit 190. The unit may be provided outside the traffic measurement device 10. The control unit 110 includes an information acquisition unit 111, a setting unit 112, an output control unit 113, a position detection unit 114, and a measurement unit 115. Details of these functional units included in the control unit 110 will be described later.

The functional configuration example of the traffic measurement device 10 according to the embodiment of the present invention has been described above.

First, calibration can be performed by the traffic measuring device 10 according to the embodiment of the present invention. More specifically, a process for calculating a road plane formula (hereinafter also referred to as a “road plane type”) and a process for calculating the traveling direction of the vehicle may be performed as calibration. Hereinafter, calibration that can be performed by the setting unit 112 will be described with reference to FIGS. 3 and 4.

FIG. 3 is a diagram showing parameters used by the setting unit 112. The setting unit 112 first determines the size pix_dot of the captured image Img ′ per unit pixel of the image sensor based on the size of the image sensor that constitutes the image capturing unit 170 and the size of the captured image Img ′ provided to the control unit 110. As a parameter. The captured image Img ′ is generated based on the captured image Img captured on the imaging surface of the image sensor that is separated from the origin O by the focal length. Further, the captured image Img ′ provided to the control unit 110 can be acquired by the information acquisition unit 111 and used by the setting unit 112.

As shown in FIG. 3, here, a case where the imaging device is a CCD (Charge Coupled Device) will be described as an example, but the CCD is only an example of the imaging device. Therefore, the imaging element may be a CMOS (Complementary Metal Oxide Semiconductor) or the like.

Here, assuming that the CCD size is ccd_size and the size of the captured image Img ′ (horizontal: width × vertical: height) is img_size, the setting unit 112 can calculate pix_dot by the following (Equation 1). In general, since the CCD size is represented by the length of the diagonal line of the CCD, the CCD size is divided by the square root of the sum of squares in the vertical and horizontal directions of the captured image Img ′ as shown in (Equation 1). Is calculated by However, the calculation of the parameter pix_dot by such a method is only an example, and the parameter pix_dot may be calculated by another method. For example, the vertical or horizontal length of the CCD may be used instead of the diagonal line of the CCD.

Note that the CCD size is easily obtained from the imaging unit 170, for example. Further, the size of the captured image Img ′ is acquired from the storage unit 180, for example. Therefore, the control unit 110 determines, based on these sizes, the three-dimensional coordinates in the real space of the captured image Img captured on the imaging surface of the CCD and the two-dimensional coordinates of the captured image Img ′ provided to the control unit 110. The correspondence can be grasped. That is, the control unit 110 grasps the three-dimensional coordinates in the real space of the captured image Img imaged on the imaging surface of the CCD from the two-dimensional coordinates of the captured image Img ′ provided to the control unit 110 based on this correspondence. can do.

Calibration can be performed using the parameters calculated in this way. Hereinafter, the calibration performed using the parameters by the setting unit 112 will be described with reference to FIG.

FIG. 4 is a diagram for explaining the function of the setting unit 112. As shown in FIG. 4, an xyz coordinate system (real space) with the origin O as a reference is assumed. In this xyz coordinate system, the road plane expression is R1x + R2x + R3z + R4 = 0. Further, a traveling direction vector v that is a vector indicating the traveling direction of the vehicle is defined as (vx, vy, vz). In the following description, as shown in FIG. 4, a point (focal point) that is separated from the origin O by the focal length f is set on the y axis, and a plane that passes through this focal point and is perpendicular to the y axis is defined as an imaging surface. The description will be continued assuming that the captured image Img is captured on the imaging surface, but the setting of each coordinate axis is not limited to such an example.

2 Parallel lines are drawn in advance on the road plane. Therefore, the parallel two straight lines are shown in the captured image Img. On the road plane, two points Q1 and Q2 that are separated by a known size Q_dis are drawn in advance. In the captured image Img, two points Q1 and Q2 are displayed as Q1 '(xs1, f, zs1) and Q2' (xs2, f, zs2). In the example shown in FIG. 4, Q1 and Q2 are drawn as points on two parallel straight lines on the road plane. However, Q1 and Q2 are not particularly limited as long as they are points on the road plane. Not.

Of the two straight lines shown in the captured image Img, the two points through which the first straight line passes are T1 (x1, y1, z1) and T4 (x4, y4, z4), and the two points through which the second straight line passes are T2. Let (x2, y2, z2) and T3 (x3, y3, z3). Then, as shown in FIG. 4, the coordinates of the intersection of the straight line connecting each of T1, T2, T3 and T4 and the origin O and the road plane are t1 · T1, t2 · T2, t3 · T3 and t4 · T4. It is expressed. For example, the setting unit 112 can perform calibration based on the following (Precondition 1).

(Prerequisite 1)
(Condition 1) The direction vectors of two parallel straight lines on the road plane are the same.
(Condition 2) The roll of the imaging unit 170 is zero.
(Condition 3) The distance from the origin O to the road plane is the height H.
(Condition 4) Q1 and Q2 that are Q_dis apart exist on the road plane.
Note that the roll being 0 means that the imaging unit 170 is installed so that an object installed in a direction perpendicular to the road plane is reflected in the vertical direction on the captured image Img. .

The setting unit 112 can derive the relational expressions shown in the following (Equation 2) and (Equation 3) based on the various data acquired as described above and (Condition 1).

Also, the setting unit 112 can derive the relational expression shown in the following (Formula 4) based on the various data acquired as described above and (Condition 2). If the roll is in a state of 0, the perpendicular component to the road plane in the axial direction parallel to the road plane formula (in the example shown in FIG. 4, the x-axis direction) is 0, so the calculation formula is simplified. (For example, if the perpendicular component in the x-axis direction is 0, R1 = 0 can be calculated).

Also, the setting unit 112 can derive the relational expression shown in the following (Formula 5) based on the various data acquired as described above and (Condition 3).

Also, the setting unit 112 can derive the relational expressions shown in the following (Expression 6) and (Expression 7) based on the various data acquired as described above and (Condition 4).

Here, in K1, the distance from the origin O to Q1 (xr1, yr1, zr1) on the road plane is a multiple of the distance from the origin O to Q1 ′ (xs1, f, zs1) on the captured image Img. It is a value indicating whether or not Similarly, in K2, the distance from the origin O to Q2 (xr2, yr2, zr2) on the road plane is a multiple of the distance from the origin O to Q2 ′ (xs2, f, zs2) on the captured image Img. It is a value indicating whether or not Therefore, the relational expression shown in the following (Formula 8) can be derived.

The setting unit 112 can calculate the measured value Q_dis ′ of the distance between two points (Q1 and Q2) on the road plane from the relational expression shown in (Formula 8) by the following (Formula 9).

The setting unit 112 can calculate R1, R2, R3, and R4 when the difference between the measured value Q_dis ′ and the known magnitude Q_dis is the smallest based on (Formula 1) to (Formula 9). . By calculating R1, R2, R3 and R4 in this way, the road plane formula R1x + R2x + R3z + R4 = 0 is determined.

The road plane calculation method described above is only an example. Therefore, the setting unit 112 can also calculate the road plane equation by another method. For example, if the distance between two parallel straight lines on the road plane is known, the road plane formula is calculated without using (Condition 2) by using the distance between the two parallel straight lines on the road plane. be able to.

Further, the setting unit 112 can also calculate the traveling direction vector v (vx, vy, vz). More specifically, the setting unit 112 can calculate the traveling direction vector v by calculating the direction of at least one of two parallel straight lines on the road plane. For example, the setting unit 112 may calculate the difference between the coordinates t2 · T2 and the coordinates t3 · T3 as the traveling direction vector v, or the difference between the coordinates t1 · T1 and the coordinates t4 · T4 as the traveling direction vector v. It may be calculated.

The setting unit 112 can perform calibration by the method described above. The road plane equation R1x + R2x + R3z + R4 = 0 and the traveling direction vector v (vx, vy, vz) calculated by such calibration can be used for measuring the traffic volume and the vehicle speed. As illustrated in FIG. 4, the setting unit 112 may set a vehicle travel axis A that is parallel to the traveling direction vector v (vx, vy, vz). If the xyz coordinates are set according to the actual size of the real space as described above, the vehicle travel axis A can also be set according to the actual size in the real space.

Further, as shown in FIG. 4, the setting unit 112 may set the measurement range E1. If it does so, traffic volume may be measured based on the vehicle area | region extracted from the measurement range E1. For example, the setting unit 112 may set the measurement range E1 based on the input operation, or may automatically set the measurement range E1 based on the traveling direction vector v (vx, vy, vz). However, when the imaging range itself is the measurement range, the measurement range need not be set. Below, in order to demonstrate easily, the case where the imaging range itself is made into a measurement range is mainly demonstrated. The output control unit 113 may cause the output unit 190 to output various information set by the setting unit 112.

The calibration performed by the setting unit 112 has been described above.

Next, details of traffic volume measurement based on a captured image Img obtained by capturing a road plane will be described. As described above, the captured image Img is provided from the imaging unit 170 to the control unit 110. Furthermore, the imaging unit 170 has a timekeeping function. When the imaging time of the captured image Img is detected by the imaging unit 170, the detected imaging time is provided to the control unit 110. The information acquisition unit 111 acquires the captured image Img and the imaging time provided from the imaging unit 170 in this way. The timekeeping function may be included in the control unit 110, or the time obtained by the control unit 110 may be used.

Subsequently, the position detection unit 114 detects a predetermined detection position based on the vehicle area extracted from the captured image Img. The predetermined detection position may be a coordinate on the vehicle travel axis A. That is, the position detection unit 114 may detect coordinates on the vehicle travel axis A as detection positions. The detection by the position detection unit 114 will be described in more detail with reference to FIG. FIG. 5 is a diagram for explaining an example of functions of the position detection unit 114. Note that the example illustrated in FIG. 5 is merely an example of detection by the position detection unit 114, and thus the detection by the position detection unit 114 is not limited to the example illustrated in FIG.

Referring to FIG. 5, the vehicle V is traveling on the road plane. First, the position detection unit 114 extracts a vehicle region from the captured image Img. The vehicle area may be extracted in any way. For example, the vehicle region may be a region specified from a silhouette extracted by the difference between the captured images Img in the frames before and after the vehicle V is shown in FIG. Or the silhouette extracted by the difference of a background image and the captured image Img may be sufficient.

In addition, many edge features are detected on the vehicle V on the captured image by the vehicle contour, the windshield, and the like. Therefore, when the position detection unit 114 further detects an edge feature from the vehicle region based on the silhouette extracted by the above processing and detects the vehicle region based on the edge feature, a vehicle with higher accuracy against changes in the imaging environment. Region extraction may be performed.

Specifically, the position detection unit 114 may detect edge features based on the silhouette extracted by the above processing, and extract a collection of detected edge features as a vehicle region. However, if only the edge detection process is performed, there is a possibility that an area in which one vehicle V appears is divided into a plurality of edge areas and extracted. Therefore, the position detection unit 114 may combine edge features that are at a distance less than the threshold into one edge region. Specifically, the position detection unit 114 performs a labeling process on each edge feature, and if the labeled edge features are at a distance less than a threshold, the edge features are combined into one edge region. Good.

Subsequently, the position detection unit 114 detects the detection position based on the vehicle area as described above. For example, the position detection unit 114 may detect a predetermined vertical plane orthogonal to the vehicle travel axis A based on the vehicle region, and detect an intersection coordinate between the vehicle travel axis A and the predetermined vertical plane as a detection position. The predetermined vertical plane is not limited, but may be the front surface of the vehicle or the back surface of the vehicle. That is, the position detection unit 114 may detect the front surface or the rear surface of the vehicle as a predetermined vertical plane.

Alternatively, the position detection unit 114 may change whether to detect the front surface of the vehicle or the back surface of the vehicle as a predetermined vertical plane depending on the situation. For example, the position detection unit 114 detects the front surface of the vehicle as a vertical plane when the traveling direction vector v is a direction from the back to the front, and when the traveling direction vector v is a direction from the front to the back, The back surface may be detected as a vertical plane. In this way, it is expected that the detection accuracy will be further improved because there is a high possibility that the front side surface is more clearly reflected in the captured image Img than the back side surface.

Here, an example of a technique for detecting the vehicle front surface F1 as a predetermined vertical plane when the traveling direction vector v is a direction from the right back to the left front will be described. First, as illustrated in FIG. 5, the position detection unit 114 detects the vehicle front lowest point m1 ′ based on the vehicle region extracted from the captured image Img. The vehicle front lowest point m <b> 1 ′ is a point having the lowest height from the ground in the vehicle body of the vehicle V. The vehicle front lowest point m1 ′ may be detected in any way. For example, when the traveling direction vector v is a direction from the right back to the left front, The point (for example, the middle point) on the edge line (the lower left line segment of the vehicle area) that is the lowest point is detected as the vehicle front lowest point m1 ′.

Subsequently, the position detection unit 114 draws a plane perpendicular to the vehicle travel axis A passing through the intersection point m0 from the lowest point m1 in front of the vehicle in real space to the road plane with a vertical line having the minimum ground height h. It detects as vehicle front F1. Here, the minimum ground height h may be a predetermined value or a value determined based on the minimum ground height detected so far. When a predetermined value is used as the minimum ground height h, for example, an average value of the minimum ground heights of a plurality of vehicles may be used as the predetermined value. In FIG. 5, the distance between the ground contact point D0 where the vehicle V is in contact with the road plane and the vehicle body low plane is shown as the minimum ground height h.

Here, the method of detecting the vehicle front surface F1 as a predetermined vertical plane when the traveling direction vector v is in the direction from the right back to the left front has been described, but the traveling direction vector v is from the left front to the right back. Similarly, in the case of the direction, a point (for example, a middle point) on the lower edge line (the lower left line segment of the vehicle area) on the back of the vehicle is detected as the lowest ground point m1 ′ as the lowest ground point m1 ′. The rear surface of the vehicle can be detected by a similar method.

When the traveling direction vector v is a direction from the left back to the right front, the lowest ground point m1 ′ is a point on the lower edge line (the lower right line segment of the vehicle area) on the front of the vehicle (for example, middle Point) is detected as the lowest ground point m1 ′, and the front surface of the vehicle can be detected by the same method. Similarly, when the traveling direction vector v is in the direction from the right front to the left back, the lowest ground point m1 ′ is a point on the lower edge line (the lower right line segment of the vehicle area) on the back of the vehicle (for example, The middle point) is detected as the lowest ground point m1 ′, and the rear surface of the vehicle can be detected by the same method.

The detection position can be detected based on the vehicle area by the method described in the above example. If the detection position is detected by the position detection unit 114, a combination of the detection position and the imaging time is obtained. FIG. 6 is a diagram illustrating an example of a combination of a detection position and an imaging time. Subsequently, the measurement unit 115 measures the traffic volume from the combination of the detection position and the imaging time obtained in this way.

Since a passing vehicle can normally be regarded as moving at a constant linear velocity, the vehicle detection position and the imaging time satisfy a certain law. The measurement unit 115 performs voting on a combination in which the detection position and the imaging time are a constant rule, and if the vote count becomes a value (peak point) equal to or greater than a certain value, it is detected based on one vehicle area. It can be considered that the set of combinations is satisfied. Therefore, since it is possible to consider that the vehicle has passed the number of voting peaks, the measurement unit 115 may measure the number of voting peaks as the traffic volume. Note that the peak may be a case where the voting power exceeds a threshold and the voting power is maximized. Therefore, since it is possible to consider that the vehicle has passed the number of voting peaks, the measurement unit 115 may measure the number of voting peaks as the traffic volume. Note that the peak point may be when the vote count exceeds a threshold value or when the vote count becomes a maximum.

According to this method, since the number of voting peaks may be measured as the traffic volume, there is no need to track the vehicle silhouette area and count the number of vehicle silhouette areas that have been successfully tracked within the imaging range. Therefore, even when a situation in which it is difficult to accurately track the vehicle silhouette region occurs, the traffic volume can be accurately measured. Therefore, according to this method, it is possible to improve the accuracy of traffic volume measurement when measuring the traffic volume based on the captured image.

さらに Further explanation using specific examples. For example, when the vehicle travels on a substantially straight line as in the example shown in FIG. 5, it is estimated that the vehicle performs a substantially uniform linear motion. Therefore, as shown in FIG. 6, when each combination of the imaging time and the detection position is plotted on the two-dimensional coordinates, the detection position based on one vehicle area substantially follows a straight line as the imaging time changes. It is speculated that it will change.

Therefore, it is assumed that the combination of the detection position and the imaging time should change linearly, and the measurement unit 115 votes for a set of straight lines passing through each combination, and the vote count reaches a peak. What is necessary is just to consider that the set of the combinations detected based on one vehicle area satisfies the straight line. Therefore, the measurement part 115 should just measure the number of the straight lines from which a vote frequency becomes a peak as traffic volume. The measuring unit 115 can also specify the slope of the straight line as the vehicle speed.

Here, there is no particular limitation on how to detect a straight line having a peak vote count. For example, the measurement unit 115 may perform a Hough transform on the combination of the detection position and the imaging time in order to detect a straight line having a peak vote count. In this case, when the imaging time is X and the detection position is Y, the XY plane can be converted to the ρ-θ plane by the relationship shown in the following (formula 1).

Ρ = X · cos θ + Y · sin θ (Formula 1)

Here, ρ corresponds to the distance from the origin O to the straight line on the XY plane, and θ corresponds to the angle formed by the X axis and the straight line. One straight line in the XY plane is represented as an intersection of curves in the ρ-θ plane. FIG. 7 is a diagram illustrating a result obtained by performing Hough transform on the combination of the detection position and the imaging time illustrated in FIG. In the example illustrated in FIG. 7, voting peak points P <b> 1 and P <b> 2 are illustrated, and the measurement unit 115 may measure “2” corresponding to the number of voting peak points as the traffic volume.

Further, the measuring unit 115 may specify the vehicle speed based on the position of the voting peak point. More specifically, as described above, if θ is equivalent to the angle between the X axis and the straight line in the XY plane, and the vehicle speed is equivalent to the slope of the straight line in the XY plane, the measurement is performed. The unit 115 can specify the vehicle speed by tan θ using θ of the voting peak point. In the example illustrated in FIG. 7, the measurement unit 115 can specify the speeds of the two vehicles by tan θ using θ of each of the voting peak points P1 and P2.

The details of the traffic volume measurement based on the captured image Img obtained by capturing the road plane have been described above.

Subsequently, an operation example of the traffic volume measuring apparatus 10 according to the embodiment of the present invention will be described. FIG. 8 is a flowchart showing an operation example of the traffic volume measuring apparatus 10 according to the embodiment of the present invention. In addition, the flowchart shown in FIG. 8 only showed an example of operation | movement of the traffic volume measuring apparatus 10. FIG. Therefore, the operation of the traffic volume measuring device 10 is not limited to the operation example shown by the flowchart of FIG.

As illustrated in FIG. 8, first, in the traffic measurement device 10, the information acquisition unit 111 acquires a captured image in which a road plane is captured by the imaging unit 170 and an imaging time corresponding to the time at which the captured image is captured ( Step S11). Subsequently, the position detection unit 114 extracts a vehicle region from the captured image (step S12), and detects a predetermined detection position based on the vehicle region (step S13).

Subsequently, the measurement unit 115 performs Hough transform on the combination of the detection position and the imaging time, and measures the number of voting peak points on the ρ-θ plane as the traffic volume (step S15). Further, the measuring unit 115 identifies the vehicle speed based on the position of the voting peak point on the ρ-θ plane (step S16). More specifically, the measurement unit 115 can specify the vehicle speed by tan θ using θ of the voting peak point.

[Description of effects]
As described above, according to the embodiment of the present invention, the information acquisition unit 111 that acquires the captured image obtained by capturing the road plane and the imaging time, and a predetermined area based on the vehicle area extracted from the captured image. There is provided a traffic volume measuring device 10 including a position detection unit 114 that detects a detection position, and a measurement unit 115 that measures the number of voting peaks obtained for a combination of the detection position and an imaging time as a traffic volume. The

According to such a configuration, since the number of voting peaks may be measured as the traffic volume, it is not necessary to track the vehicle silhouette area that has passed through the imaging range while tracking the vehicle silhouette area. Therefore, even when a situation in which it is difficult to accurately track the vehicle silhouette region occurs, the traffic volume can be accurately measured. Therefore, according to this method, it is possible to improve the accuracy of traffic volume measurement when measuring the traffic volume based on the captured image.

[Description of modification]
The preferred embodiments of the present invention have been described in detail above with reference to the accompanying drawings, but the present invention is not limited to such examples. It is obvious that a person having ordinary knowledge in the technical field to which the present invention belongs can come up with various changes or modifications within the scope of the technical idea described in the claims. Of course, it is understood that these also belong to the technical scope of the present invention.

For example, in the above example, the case where the vehicle has a substantially constant linear motion has been described. However, the traveling state of the vehicle is not limited to this example, and the vehicle may travel on a curved road plane. Even in such a case, since it is estimated that the vehicle moves substantially at a constant speed, if the vehicle travel axis A is set along a curved road plane, the detection position based on one vehicle area is This is because, as in the above example, it is estimated that the image changes along a straight line with a change in imaging time. In such a case, the measurement range E1 may be set along a curved road plane.

Furthermore, the case where the speed of the vehicle changes may be used instead of the case where the vehicle moves at a substantially constant speed. For example, when each combination of the imaging time and the detection position is plotted on a two-dimensional coordinate, it is estimated that the detection position based on one vehicle area changes substantially along a curve as the imaging time changes. There may be cases.

In such a case, it is assumed that the combination of the detection position and the imaging time should change on the curve, and the measurement unit 115 performs voting on a set of curves that pass through each combination, and the vote count is What is necessary is just to consider that the set of combinations detected based on one vehicle area | region satisfy | fills the curve used as a peak. Therefore, the measurement unit 115 may measure the number of curves having a peak vote frequency as the traffic volume. Moreover, the measurement part 115 can also specify each vehicle speed by the differentiation of each curve where the vote frequency becomes a peak. Furthermore, the measurement unit 115 can also calculate the acceleration of each vehicle by the second-order differentiation of each curve having a peak vote frequency.

Each block configuring the control unit 110 includes, for example, a CPU (Central Processing Unit), a RAM (Random Access Memory), and the program stored in the storage unit 180 is expanded and executed by the CPU. Thus, the function can be realized. Or each block which comprises the control part 110 may be comprised by the hardware for exclusive use, and may be comprised by the combination of several hardware.

In this specification, the steps described in the flowcharts are executed in parallel or individually even if they are not necessarily processed in time series, as well as processes performed in time series in the described order. Including processing to be performed. Further, it goes without saying that the order can be appropriately changed even in the steps processed in time series.

DESCRIPTION OF SYMBOLS 10 Traffic measurement apparatus 110 Control part 111 Information acquisition part 112 Setting part 113 Output control part 114 Position detection part 115 Measurement part 170 Imaging part 180 Storage part 190 Output part E1 Measurement range F1 Vehicle front surface P1 Voting peak point A Vehicle travel axis V vehicle

Claims

An information acquisition unit for acquiring a captured image obtained by capturing a road plane and an imaging time;
A position detection unit that detects a predetermined detection position based on a vehicle region extracted from the captured image;
A measurement unit that measures traffic based on the number of voting peaks obtained for the combination of the detection position and the imaging time;
A traffic volume measuring device.
The measurement unit performs a Hough transform on the combination of the detection position and the imaging time, and measures the number of voting peak points on the Hough space as a traffic volume.
The traffic measuring device according to claim 1.
The position detection unit detects an edge feature from the captured image, and extracts the vehicle region based on the edge feature;
The traffic measuring device according to claim 1.
The measurement unit identifies a vehicle speed based on a position of the voting peak point;
The traffic measuring device according to claim 1.
The position detection unit detects, as the detection position, coordinates on a vehicle travel axis in a real space set in accordance with an actual size.
The traffic measuring device according to claim 1.
The position detection unit detects a predetermined vertical plane orthogonal to a vehicle travel axis in real space based on the vehicle region, and detects an intersection coordinate between the vehicle travel axis and the vertical plane as the detection position;
The traffic measuring device according to claim 5.
The position detection unit detects a vehicle front surface or a vehicle back surface as the predetermined vertical plane;
The traffic measuring device according to claim 6.
The position detection unit detects the front surface of the vehicle as the vertical plane when the vehicle traveling direction in the real space is from the back to the front, and the vehicle traveling direction in the real space is the direction from the front to the back. In detecting the back of the vehicle as the vertical plane,
The traffic measuring device according to claim 7.
Obtaining a captured image obtained by capturing a road plane and an imaging time;
Detecting a predetermined detection position based on a vehicle area extracted from the captured image;
Measuring traffic based on a voting peak point obtained for the combination of the detection position and the imaging time;
Including traffic volume measurement method.