JP2000163691A - Traffic flow measuring instrument - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a traffic flow measuring instrument which can accurately measure a traffic flow even when the flow contains the shadows of vehicles and a tall large-sized vehicles. SOLUTION: A traffic flow measuring instrument is provided with an input device 1 which picks up the image of a road, a vehicle detecting means 20 which sets a vehicle detecting area at a certain point on the road, detects the vehicles passed through the vehicle detecting area by processing the image data obtained from an image picking-up section, and outputs vehicle detecting signals, and a general purpose computer 3 which calculates the information on a traffic flow based on the information on the rising time and falling time of the vehicle detecting signals between two vehicle setting areas which are set at a fixed interval in the advancing direction of the vehicles.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a traffic flow measurement for measuring traffic information such as the number of vehicles passing on a road, the speed, the type of vehicle, and the like, by processing an image obtained by imaging the road. It concerns the device.

[0002]

2. Description of the Related Art Conventionally, as this kind of traffic flow measuring device, there has been one described in Japanese Patent Application Laid-Open No. 9-102097. In this method, one vehicle detection area is provided in the traveling direction of a vehicle, and the number of vehicles is measured by detecting a vehicle based on a change in luminance of the detection area (background difference method or time difference method).

[0003]

However, in vehicle detection characterized by a luminance change in a single vehicle detection area as in a conventional traffic flow measuring device, the shadow of a vehicle running in an adjacent lane may be erroneously detected. There is a problem that the vehicle may be measured. In addition, in general, in a video obtained from a camera installed so as to have a view in the depth direction from a road edge or above a road, a distant vehicle has a higher frequency of vehicle overlap on the image, and furthermore, Even before the vehicle is relatively unlikely to overlap, if a large vehicle with a high vehicle height is present as a following vehicle, it is inevitable that it will overlap with the vehicle in front, and two vehicles may be mistakenly measured as one. There was a problem that there is. In addition, if the measurement between two points (a method of obtaining the speed of each vehicle from the time required for the vehicle to travel a known distance between the two points) is performed under the condition where such an overlap occurs, the correctness is obtained. Since it is difficult to calculate the speed, there is a problem that accurate measurement of the traffic flow cannot be performed.

[0004]

According to the present invention, there is provided a traffic flow measuring device for capturing an image of a road, setting a vehicle detection area at a certain point on the road, and processing image data obtained from the imaging unit. A vehicle detection means for detecting a vehicle passing through the vehicle detection area and outputting a vehicle detection signal, and a rising time of the vehicle detection signal between two vehicle detection areas set at a constant interval in the traveling direction of the vehicle. A calculating means for calculating traffic flow information based on the fall time information.

[0005]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiment 1 FIG. 1 is a diagram showing a configuration of a traffic flow measuring device according to one embodiment of the present invention. In the figure, reference numeral 1 denotes an input device for obtaining a video signal of an area including a road serving as a measurement range by an imaging means such as a CCD camera, and 20 denotes image data 40 from the input device 1.
The vehicle detection means 31 detects the entry and exit of the vehicle based on the brightness change on a straight line (detection line) in the cross direction of the road, which is set in advance for each lane, and 31 is output from the vehicle detection means 20. Vehicle flag setting means for indicating whether a vehicle is present on the detection line based on the rising signal and the falling signal of the vehicle detection signal 21 for the line, and 32 is a vehicle flag for holding the state of the vehicle flag setting means 31 A storage memory, 33 is a time storage memory, which holds time, and 36 is a transit time for calculating and comparing the transit time of the front part of the vehicle and the transit time of the rear part of the vehicle passing between the detection lines to determine whether or not it is a shadow. The comparing means 38 is a number / speed calculating means for calculating the number of vehicles and the vehicle speed based on a signal from the detection time storage memory 33 and a signal from the passing time comparing means 36. .

In this embodiment, the vehicle detection means 20 is replaced by the general-purpose image processing device 2 and the vehicle flag setting means 3 is used.
1, vehicle flag storage memory 32, time storage memory 3
3. Passing time comparing means 36 and number / speed calculating means 38
Is performed by the general-purpose computer 3 which is an arithmetic means. The general-purpose image processing device 2 includes an A / D converter that converts an analog video signal into a digital image, a frame memory that stores the converted digital image,
The general-purpose image processing apparatus 2 and the general-purpose computer 3 are electrically connected by a general-purpose bus 4, and the vehicle detection signal 21 is transmitted to the general-purpose computer 3 through the general-purpose bus 4. It has become so.

Next, the operation of this embodiment will be described. As shown in FIG. 2, the input device 1 has a C
The CD camera is installed so as to have a field of view in the traveling direction and the depth direction of the vehicle from above a road serving as a measurement range, and captures an image as shown in FIG. Here, the detection lines (D1, D2, D3, D4) marked continuously in the transverse direction of each lane shown in FIG. 3 are used to measure the number of vehicles traveling on each lane and the average speed. The position to be set differs depending on the camera installation height, the double angle, the focal length of the lens, and the like, and the start and end point coordinate values of each line are given to the general-purpose image processing apparatus in advance as parameters.

FIG. 4 is a cross-sectional view of the right lane in FIG. 3 when the abscissa indicates the vehicle traveling direction and the ordinate indicates the height. The area shown by the dotted line is the imaging range. here,
Point A is when the tip of the vehicle is located behind the detection line D1 on the captured image, and Point B is when the tip of the vehicle just passes on the detection line D1 on the captured image. The point C is the time when the rear end of the vehicle passes on the detection line D1 on the captured image, and the point D is the position where the front end of the vehicle is just on the detection line D2 on the captured image. , And the point E is when the rear end of the vehicle on the captured image has just passed on the detection line D2.

FIG. 5 shows a state in which the vehicle passes the points A to E at a constant speed from the upstream of the visual field in FIG.
Vehicle detection signal 1 (vehicle detection line D1) and vehicle detection signal 2 (vehicle detection line D2) output by vehicle detection means 20
Represents the result. FIG. 6 shows the vehicle flag setting means 31 and the passing time comparing means 3 processed by the general-purpose computer 3.
6 is a flowchart showing a processing procedure of the number / speed calculation means 38.

First, a video signal picked up by a CCD camera as an input device 1
The analog image signal is converted into digital image data by the / D converter and supplied to the vehicle detection means 20. Then, the vehicle detection means 20 outputs a signal based on the image data when a vehicle exists on a preset detection line. As a method for determining whether or not a vehicle exists, a method disclosed in Japanese Patent Application Laid-Open No.
The technique adopted in the "speed measurement image processing apparatus" described in Japanese Patent Application Laid-Open No. 9-209 can be used. This is because the image data on the detection line where no vehicle exists is stored in the background memory in advance, compared with the image data on the detection line read in the next frame, It determines that a vehicle is present and outputs a signal (1). If it is less than the threshold value, the latest image data on the detection line is exchanged with the background memory, and it is determined that there is no vehicle, and a signal (0) is output.

In the flowchart of FIG. 6, the vehicle flag 1, the vehicle flag 2, and the vehicle passage flag are set to 0 as initial values.
Is given. In FIG. 4, when the vehicle tip is at the point A, the vehicle detection means 20 does not reach the detection line D1.
The general-purpose computer 3 is notified of 0 as the vehicle detection signal D2 via the general-purpose bus 4. At this time, in the general-purpose computer 3, the processing proceeds in the order of start → S1 → S4 → S6 → end. In this example, no processing is performed because the vehicle does not exist on the two detection lines.

When the leading end of the vehicle reaches the point B, the vehicle detecting means 20 notifies the general-purpose computer 3 via the general-purpose bus 4 of 1 as the vehicle detection signal D1 and 0 as the vehicle detection signal D2. At this time, in the general-purpose computer 3, the processing proceeds in the order of start → S1 → S2 → S3 → S6 → S7 → S10 → end. In this example, the vehicle flag 1 is set to 1, and the rising time T1s of the vehicle detection signal D1 is stored in the time storage memory 33.

Until the leading end of the vehicle exceeds the point B and reaches the point C, the vehicle detection means 20 outputs the vehicle detection signal D
1 as 1 and 0 as vehicle detection signal D2
To the general-purpose computer 3 via. At this time, in the general-purpose computer 3, start → S1 → S2 → S6 → S7
The process proceeds in the order of → S10 → End. In this example, nothing is processed while the vehicle flag 1 is kept at 1.

When the leading end of the vehicle reaches the point C, the vehicle detecting means 20 notifies the general-purpose computer 3 via the general-purpose bus 4 of 0 as the vehicle detection signal D1 and 0 as the vehicle detection signal D2. At this time, in the general-purpose computer 3, the processing proceeds in the order of start → S1 → S4 → S5 → S6 → S7 → S10 → end. In this example, the vehicle flag 1 is returned to 0, and the falling time T1e of the vehicle detection signal D1 is stored in the time storage memory 33.

When the leading end of the vehicle reaches the point D, the vehicle detecting means 20 notifies the general-purpose computer 3 via the general-purpose bus 4 of 0 as the vehicle detection signal D1 and 1 as the vehicle detection signal D2. At this time, in the general-purpose computer 3, S1
The processing proceeds in the order of → S4 → S6 → S7 → S8 → S9 → End. In this example, the vehicle flag 2 is set to 1, and the rising time T2s of the vehicle detection signal D2 is stored in the time storage memory 33.

Until the leading end of the vehicle exceeds the point D and reaches the point E, the vehicle detection means 20 outputs the vehicle detection signal D
0 is set as 1 and 1 is set as the vehicle detection signal D2.
To the general-purpose computer 3 via. At this time, in the general-purpose computer 3, start → S1 → S4 → S6 → S7
The process proceeds in the order of → S8 → End. In this example, nothing is processed while the vehicle flag 2 is kept at 1.

When the leading end of the vehicle reaches the point E, the vehicle detecting means 20 notifies the general-purpose computer 3 via the general-purpose bus 4 of 0 as the vehicle detection signal D1 and 0 as the vehicle detection signal D2. At this time, in the general-purpose computer 3, S
1 → S4 → S6 → S7 → S10 → S11 → S12 → S1
The process proceeds in the procedure of S3, and in the process of S13, S
14 → End or go to the end as it is. In this example, the vehicle flag 2 is returned to 0, the falling time T2e of the vehicle detection signal 2 is stored in the time storage memory 33, and the passing time (Tf = T2s-
T1s) and the transit time focusing on the rear part of the vehicle (Te = T2)
e-T1e) is calculated respectively.

Next, it is determined whether or not | Te−Tf |> α (Equation 1) is satisfied. If Expression 1 is satisfied, it is determined as a vehicle because it has a height component, the number of vehicles is increased by one, and the vehicle speed is calculated from the Tb value. If | Te
If −Tf |> α is not satisfied, it is determined as a shadow, and is not counted as the number of vehicles, and the speed is not calculated.

Here, the principle that the transit times at the front and rear portions of the vehicle are different will be described. FIG. 7 is an explanatory diagram for explaining the principle that the transit times at the front and rear portions of the vehicle are different.
FIG. 7A shows a case where attention is focused on the rear part of the vehicle, and the passing time of the rear part of the vehicle is determined by the vehicle detection signal D1 and the vehicle detection signal D.
2, the distance Le that the vehicle moves on the image is ideally equal to the distance L between the two detection lines.

FIG. 7B shows a case where attention is paid to the front part of the vehicle, and the passing time at the front part of the vehicle is calculated from the rising times of the vehicle detection signal D1 and the vehicle detection signal D2. However, since the vehicle has a height, when a bird's-eye view image is taken, the vehicle is detected before each detection line on the image. Therefore, the moving distance Lf detected on the image is
Since the distance between the two detection lines is shorter than L, the vehicle
When traveling between two lines at a constant speed, the transit time T focusing on the front part of the vehicle is shorter than the transit time Te focusing on the rear part of the vehicle.
f becomes shorter, and Te> Tf.

However, when the vehicle is erroneously detected as a vehicle due to the shadow, since the shadow has no height component, there is no difference in the moving distance between the front part and the rear part, and ideally Te = Tf. However, when | Te−Tf | <α is satisfied in consideration of the speed fluctuation of the vehicle and the illumination fluctuation on each detection line,
It is determined to be a shadow.

In this embodiment, a vehicle detection line is provided in a vehicle transverse direction having a certain known distance with respect to the vehicle traveling direction, and a passing time between the two detection lines, focusing on the front part and the rear part of the vehicle, is set. By comparing, it is possible to determine whether the detected area is a vehicle or a shadow formed by the vehicle, so that the number of vehicles can be measured with high accuracy. Thus, the average speed of the passing vehicle can be measured.

Embodiment 2 FIG. FIG. 8 is a diagram showing a configuration of a traffic flow measuring device according to another embodiment of the present invention. In the figure, the input device 1, vehicle detection means 20, vehicle flag setting means 31, flag storage memory 32, time storage memory 33, passing time comparison means 36, number / speed calculation means 3
8 is the same as that of the first embodiment, and when a change in brightness is given on the detection line due to some factor such as a change in sunshine or a momentary movement of a cloud, the vehicle detection unit 20 erroneously outputs the vehicle detection signal 21 as the vehicle detection signal 21. 1 and a noise removing means 34 for removing these noises and a single vehicle for preventing the single vehicle from being erroneously recognized as a plurality of vehicles when separated into a plurality of regions. Separation determination means 35
Is provided before the transit time comparing means 36.

Next, the operation of this embodiment will be described. FIG. 9 shows one vehicle detection line (D1 and D2).
Output of the vehicle detection signal 21 to the
The relationship between the vehicle flags is as follows: (1) noise, (2) single vehicle is separated, (3) single vehicle passes normally, and (4) two vehicles pass normally. FIG.
Are the vehicle flag setting means 31, the noise removing means 34, the single vehicle separation determining means 35, the passing time comparing means 36, and the number / speed calculating means 3 which are processed by the general-purpose computer 3.
8 is a flowchart showing the processing procedure of FIG.

Steps S1 to S15 correspond to the processing procedure of the vehicle detection signal D1, and steps S16 to S27 correspond to the vehicle detection signal D2.
The processing procedure of S28 to S31 represents the processing procedure of calculating the number of vehicles and the vehicle speed, and the processing procedure of S32 to S35 represents the processing procedure of canceling the contents stored in the time storage memory 33 relating to the vehicle detection signal D1. I have.

Here, for the sake of simplicity, when the vehicle detection signal D1 is in the state of FIG. 9 (1) to FIG. 9 (4), no vehicle exists on the vehicle detection line D2 and the vehicle It is assumed that the detection signal D2 always outputs 0. That is, as processing of the vehicle detection signal D2, S
That is, the processing is performed in the order of 16 → S19 → S28 → S32.

In FIG. 9 (1), at time A, the vehicle detection signal D1 is 0, so that in the flowchart of FIG.
Start → S1 → S4 → S15 → End
The vehicle flag 1 remains at 0. And at time B,
Since the vehicle detection signal D1 has risen and has output 1, in the flowchart of FIG. 10, start → S1 → S2 → S
3 → S13 → S15 → Processing of vehicle detection signal D2 → S32
→ Proceeding in the end procedure, the vehicle passage flag is set to 1 and the vehicle flag 1 is set to 1 in S3, and the rising time T1s of the vehicle detection signal D1 is stored in the time storage memory 33.

Then, at time C, the vehicle detection signal D1 outputs 1 without change, so in the flow of FIG. 10, start → S1 → S2 → S13 → S15 → processing of the vehicle detection signal D2 → S32 → end The vehicle flag 1 and the vehicle passage flag do not change. And
At time D, since the vehicle detection signal D1 falls and outputs 0, the start → S1 → S4 → S5 → S6 → S13
The procedure proceeds in the order of → S14 → S15 → End, and T1 is obtained from the minimum measurement time (maximum measurement speed) preset in S5.
If (time difference between T1s and the current time) is small, it is determined as noise, and the vehicle flag 1 is returned to 0 in S6. If there is no time data in the buffer of the vehicle flag and the detected time storage memory 33 in S13, the vehicle passing flag is reset to 0. Such a procedure makes it possible to remove noise.

In FIG. 9B, time A and time B are the same processing procedures as time A and B in (1).
Times C and D are the same processing procedure as time C in (1) above. At time E, since one vehicle has been separated into two, the vehicle detection signal L1 falls and outputs 0, so in the flowchart of FIG. 10, start → S1 → S
4 → S5 → S7 → S13 → S15 → Processing of the vehicle detection signal D2 → S32 → End, proceeding from the measurement minimum time (measurement maximum speed) preset in S5 to T1 (T1
If the time difference between s and the current time is large, it is determined that the vehicle is not noise, that is, it is determined that the vehicle is a vehicle.
And the current time T1e is stored in the time storage memory 33.

At time F, since the vehicle in the separated head area has passed, 0 is output as the vehicle detection signal D1. Therefore, in the flowchart of FIG.
The process proceeds in the order of → S4 → S8 → S13 → S15 → Processing of the vehicle detection signal D2 → S32 → End, and the vehicle flag 1 is maintained in the state of 2. If the time data is not registered in the buffer of the vehicle flag and the time storage memory 33 in S13, the vehicle passage flag is returned to 0.

At time G, the vehicle detection signal D1 rises again by the separated rear vehicle and outputs 1, so that in the flowchart of FIG.
S1 → S2 → S10 → S11 → S13 → S15 → Processing of vehicle detection signal D2 → S32 → End
If T2 is smaller than the preset vehicle time at 0, it is determined that the vehicle has separated, and the vehicle flag 1 is returned to 1. Times H, I, and J are the same processing as C in (1) above.

At time K, the vehicle detection signal D1 falls and outputs 0, so in the flow chart of FIG. 10, start → S1 → S4 → S5 → S7 → S13 →
The process proceeds in the order of S15 → processing of the vehicle detection signal D2 → S32 → end. If T1 (the time difference between T1s and the current time) is larger than the minimum measurement time (maximum measurement speed) preset in S5, the vehicle is not noise. And the vehicle flag 1 is set to 2 in S7, and the current time T1
e is stored in the time storage memory 33.

At times L and M (when the time from T1e to the current time is shorter than the preset inter-vehicle time [β2]), the rear part of the separated vehicle has finished passing, so that the vehicle detection signal D1 is set to 0. Is output. That is, in the flowchart of FIG. 10, start → S1 → S4 → S
8 → S13 → S15 → Processing of vehicle detection signal D2 → S32
→ Proceeding in the end procedure, the vehicle flag 1 is kept in the state of 2. In addition, if no time data exists in the buffer of the vehicle flag and the detected time storage memory 33 in the end procedure and S13, the vehicle passage flag is returned to 0.

At time N (the time from T1e to the current time exceeds the preset inter-vehicle time [β2]), 0 is output as the vehicle detection signal D1. That is, in the flowchart of FIG. 10, start → S1 → S4
→ S8 → S9 → S13 → S15 → Processing of vehicle detection signal D2 → S32 → End
Is returned to 0, and the values of T1s and T1e are registered in the buffer of the time storage memory 33. In this way, by comparing the elapsed time from the falling time of the vehicle detection signal with the preset inter-vehicle time [β2], it is possible to determine whether one vehicle has separated.

In FIG. 9 (3), times A and B correspond to times A and B in (2) and times C, D, E, F, G and
I and J are the time C of the above (2) and the times K, L, M, N,
Is the same processing as K, L, M, N in the above (2), and time L, M is the same processing as the time M, N in (2). FIG.
In (4), times A and B correspond to times A and B in (2) above.
And times C, D, E, F, G, and H correspond to time C in (2) above.
The time I is the same as the time E in the above (2), and the times J and K are the same as the time F in the above (2).

Then, at time L, the vehicle detection signal L1 rises again by the following vehicle, and 1 is output. That is, in the flowchart of FIG.
1 → S2 → S10 → S12 → S13 → S15 → Processing of vehicle detection signal D2 → S32 → End
If T2 is greater than the preset vehicle time, it is determined that the following vehicle has passed, and the vehicle flag 1 is returned to 1;
The values of T1s and T1e are registered in the buffer of the detection time storage memory, and the leading passage time of the succeeding vehicle is stored in T1s. Times M, N, O, P, Q, R, S, T are the same C, D, E, F, G, H, time U is the same time I, times V, W are the same time J, E and the time X are the time N in the above (2).
This is the same processing as.

Although the vehicle detection signal D1 has been described above, the same operation is performed with the vehicle detection signal D2. The operation of the vehicle detection signal D2 differs from the operation of the vehicle detection signal D1 in that the speed calculation flag is set to 1 after the rear portion of the vehicle has passed.

Next, a processing procedure for canceling the contents of the vehicle detection signal D1 stored in the time storage memory 33 will be described. This processing deletes the time data registered in the buffer of the time storage memory 33, for example, when the vehicle stops on the vehicle detection line D2 after passing on the vehicle detection line D1 before reaching the vehicle detection line D2. Process.

Specifically, if the condition "the vehicle passing flag is 1", "the vehicle flag 2 is 0" and "the number of time data registered in the buffer of the time storage memory is not 0" are satisfied in S32 , S33, and T4 (the time difference between the rising time data of the vehicle detection signal D1 registered first in the buffer and the current time) and Ψ (the time difference from the rising signal to the falling signal of the vehicle detection signal D1). The calculated two times are compared in S34, and if T4> Ψ, it is determined that the vehicle has stopped halfway, and the first registered time data in the corresponding buffer is deleted.

Here, the set values of the inter-vehicle times β2 and γ2 will be described with reference to FIG. Here, it is assumed that the vehicles 1 and 2 are traveling from left to right, and from time t1, t2,
The state of the vehicles 1 and 2 at the time of t3 (t1 <t2 <t3), and the black triangle represents the vehicle detection line.

Now, time t1 is the moment when the leading end of the vehicle 1 passes on the vehicle detection line on the image, and time t2 is the moment when the rear part of the vehicle 1 finishes passing on the vehicle detection line on the image.
Time t3 is the moment when the tip of the following vehicle 2 passes on the vehicle detection line on the image. It is also assumed that the traveling speeds of the vehicle 1 and the following vehicle 2 are the same, the height is h, and the length is Vw.

At this time, if the time when the vehicle 1 passes on the vehicle detection line is T1 [sec], T1 = t2−t1,
If the vehicle speed at that time is V1 [m / sec], V1
= L1 / T1 (Equation 2). Here, L1 [m] is the distance that the vehicle 1 has moved between time t1 and time t2.
At the time t1, the tip of the vehicle 1 is on the vehicle detection line on the image, but there is actually a distance m1 from the tip of the vehicle to the vehicle detection line. L1 = Vw
+ M1 (Equation 3).

From time t2 to time t3, vehicle 1
Is the moving distance of L2 [m], L2 = V1 * (t3
−t2) (Equation 4), and assuming that the inter-vehicle distance between the vehicle 1 and the following vehicle 2 is Lb [m], Lb = L1 + L2-Vw
(Equation 5) is obtained.

If the threshold value of the inter-vehicle distance is the distance to travel for one second of the traveling speed, β2 (γ2) = 1 * V1 (Equation 6), and if Lb> β2 (Equation 7), two vehicles are detected on the vehicle detection line. Is determined to have passed, and if Lb ≦ β2, it is determined that one vehicle has separated. In addition, the equation 7 is calculated at time T1,
When represented by T2 and m1, 1- (m1 / (Vw + m1)) *
T1> T2.

In this embodiment, noise is eliminated by setting the maximum measurement speed, and the inter-vehicle distance is set and determined according to the traveling speed, so that one vehicle passes through in a plurality of separated states. , Or whether two vehicles have passed in succession can be accurately distinguished.

Embodiment 3 In this embodiment, even if a plurality of vehicles can correctly detect the vehicle detection line D1 on the image, the vehicle detection line D2 (located farther from the vehicle detection line D1 when viewed from the CCD camera) is overhead-view photographing. Due to the nature of the above, there is a possibility that an image is taken so that a plurality of vehicles pass continuously. Therefore, the vehicle detection line D2
When there are a plurality of vehicles that have passed the vehicle detection line D1 when the vehicles have passed through the vehicle, a plurality of vehicle combination determination means 37 for determining whether or not the vehicles that have passed the vehicle detection line D2 have been combined are provided. Things.

FIG. 12 is a diagram showing a configuration of a traffic flow measuring device according to another embodiment of the present invention. In the figure, an input device 1, a vehicle detection means 20, a vehicle flag setting means 31,
The flag storing memory 32, the time storing memory 33, the noise removing unit 34, the single vehicle separation determining unit 35, the passing time comparing unit 36, and the number / speed calculating unit 38 are the same as in the first embodiment, and The determination means 37 is provided before the number / speed calculation means 38.

Next, the operation of this embodiment will be described. FIG. 13 shows vehicle detection signals D1 and D2 when two vehicles continuously pass vehicle detection lines D1 and D2.
And the states of the vehicle flags 1 and 2, the horizontal axis represents the time axis, and the vertical axis represents the respective output values. FIG. 14 is a flowchart showing the processing procedure of the plural-vehicle combination determining means 37 and the number / speed calculating means 38 which are processed by the general-purpose computer 3. Other processing is the same as that of the flowchart of FIG. The same is true.

In the example of FIG. 13, two vehicles are correctly separated and output as a vehicle detection signal D1 on a vehicle detection line D1 on a captured image. At this time, the rising and falling times of the first vehicle detection signal D1 are T1s1 and T1e1, and the falling and rising times of the vehicle detection signal D1 of the following vehicle are T1s2 and T1s.
1e2.

On the other hand, on the vehicle detection line D2, a signal as one vehicle has been output as the vehicle detection signal D2. At this time, the rising of the vehicle detection signal D2,
The falling edges are T2s and T2e. At time A, two time data (T
1s1, T1e1, T1s2, T1e2) are registered. Further, since the signal of the vehicle detection line D2 has not fallen and the speed calculation flag is 0, the processing procedure of S28 → S32 → S33 → S34 → end is sufficient in the flowchart of FIG.

Since the speed calculation flag is set to 1 at the time B, in the flowchart of FIG. 14, S28 → S29 → S30 → S31 → S37 → S38 →
The processing proceeds in the order of S39 → S37 → S38 → S40 → End. Then, in the first loop, in S38, | T6-T5
(= T1e1-T1s1) |> T5 (= T1e1-T1
Since the condition of s1) * 、 was satisfied, the loop goes around once more, and in S38, | T6-T5 (= T1e2-T1s1)
Since the condition of |> T5 (= T1e2−T1s1) * Ψ2 was not satisfied, the process proceeds to S40. In this case, since the loop has been performed twice, the number of vehicles is increased by two, and the vehicle speed is calculated from T1e2 and T2e.

In this embodiment, when a plurality of time data are registered in the buffer of the time storage memory when the vehicle detection signal D2 falls, the time data of the buffer and the rising of the vehicle detection signal D2 are used. By comparing the time until the fall, it is possible to determine whether or not a plurality of vehicles have passed the vehicle detection line D2 continuously, and the vehicles are overlapped and imaged on the vehicle detection line D2 on the image. In this case, it is possible to effectively separate the vehicles and calculate the number and speed of the vehicles.

Embodiment 4 In this embodiment, the vehicle width determination unit 22 is added, and the accuracy determined by the noise removal unit 34, the single vehicle separation determination unit 35, and the multiple vehicle combination determination unit 37 is increased. Various threshold values α and β used in the above three means
1, β2, γ1, γ2, ψ, and ε are set based on a vehicle height of 1.5 m and a vehicle length of 4 m. Although the types of vehicle height of the target vehicle are small, the time of passing on the detection line on the image differs depending on the vehicle height. In addition, the relationship between the vehicle height and the vehicle width is considered that a vehicle having a wide vehicle width is considered to have a high vehicle height. And more accurate noise removal, single vehicle separation, and multiple connection determination. FIG. 15 is a diagram showing a configuration of a traffic flow measuring device according to another embodiment of the present invention, and a vehicle width determining means 22 is newly provided in the general-purpose image processing device 2.

Next, the operation of this embodiment will be described. For example, the vehicle width determination unit 22 uses a threshold value used in the vehicle detection unit 20 as described in the first embodiment to determine whether a large vehicle (2.5 m in width) or a small vehicle (1.8 in vehicle width).
Classification to the extent of distinguishing m) is possible. For example, assuming that the detection line is 3 m, a small vehicle may be determined if the threshold value is 0.5 to 0.75, and a large vehicle may be determined if the threshold value is 0.75 or more. Further, the operation flow of this embodiment is the same as the flowchart shown in FIG. 10 of the second embodiment, but in the flowchart shown in FIG. 10, S5, S8, S10, S20, S23, S25,
In the process of S34, the respective threshold values α, β1, β2, γ1, γ2, ψ, and ε are determined based on the determination result of the vehicle width determination unit 22.

In this embodiment, when a vehicle is detected, a vehicle width is determined, and a vehicle height having a high correlation with the vehicle width is accurately estimated to set a threshold value. It is possible to improve the accuracy of the vehicle separation determination unit and the multiple vehicle combination determination unit.

Embodiment 5 When the installation position of the CCD camera is on the roadside and the installation height is low, the rear lane is concealed by a large vehicle traveling in the lane on the CCD camera side, for example, a height approximately the same as the vehicle height of a large vehicle. Would. That is, although the vehicle is not actually traveling in the back lane, the state may be the same as that of the vehicle existing on the vehicle detection line, and the vehicle detection signal may be output.
Therefore, in this embodiment, a vehicle presence detection area is installed above the vehicle detection line in the back lane, as viewed on the image,
Immediately after the vehicle has passed the vehicle detection line, the vehicle presence detection signal is used to confirm whether the vehicle detection signal is a false output from a vehicle in the front lane or whether a true vehicle has passed. Is to increase.

FIG. 16 is a view showing a configuration of a traffic flow measuring device according to another embodiment of the present invention. In the general-purpose image processing device 2, a vehicle presence detecting means 23 is newly provided. The vehicle presence detection means 23 performs processing in the same manner as the vehicle detection means 20.

Next, the operation of this embodiment will be described. FIG. 17 shows an image taken when a large vehicle passes through the front lane as viewed from the CCD camera, and the vehicle has passed through the vehicle detection line at a position farther than the vehicle detection line in the back lane as shown in the figure. There is provided a vehicle presence detection area 5 for determining whether or not the vehicle exists. FIG. 18 shows a vehicle detection signal D2 of the rear lane when a large vehicle passes through the front lane.
The state of the vehicle detection flag 2, the state of the vehicle presence detection signal E, and the state of the vehicle detection signal D2, the vehicle detection flag 2, and the state of the vehicle presence detection signal E when the vehicle passes through the back lane are shown.

As shown in FIG. 17, when a large vehicle passes through the front lane, the same effect as when the vehicle passes through the rear lane on the image has the same effect, and the vehicle detection signal D2 outputs an erroneous signal. Would. The vehicle presence detection signal rises before and after the rising of the vehicle detection line signal D2, and falls before the vehicle detection line signal D2 falls. That is, since no vehicle exists in the back lane, no vehicle exists in the vehicle presence detection area at the time when the vehicle detection line signal D2 falls. Therefore,
If the state of the vehicle presence detection signal E is 0 at the time of the fall of the vehicle detection line signal D2, it is determined that the signal is a fake signal from a large vehicle, and the processing corresponding to S24 and S27 in the flowchart shown in FIG. The section appropriately processes the vehicle flag 2, the speed calculation flag, the time data registered in the time storage memory, and the like.

On the other hand, when the vehicle passes through the back lane, the vehicle passing through the vehicle detection line D2 passes through the vehicle presence detection area for a certain period, and the vehicle presence detection signal E2
Falls shortly after the vehicle detection line signal D2 falls. In other words, if the state of the vehicle presence detection signal E is 1 at the time when the vehicle detection signal D2 falls, it is determined that the vehicle has passed the back lane, and no rejection processing is performed.

In this embodiment, a vehicle presence detection area for judging whether or not there is a vehicle passing through the vehicle detection line is provided at a position farther from the vehicle detection line in the back lane as viewed from the CCD camera. It is possible to determine whether the vehicle is a pseudo passing vehicle caused by a large vehicle passing in the front lane or a normal passing vehicle generated when a vehicle passes in the back lane, and the number of vehicles in the back lane can be accurately determined. It becomes possible to measure.

In the first to fifth embodiments, a case has been described in which two vehicle detection lines are set for one lane (that is, in the case of multiple lanes, two detection lines are set for each lane). Alternatively, a method of setting only two vehicle detection lines, detecting a vehicle, and then determining which lane the vehicle corresponds to may be applied. Further, a method other than the background difference may be applied as the vehicle detection means.

[0063]

As described above, according to the present invention, the road is imaged by the image pickup means, the vehicle detection area is set at a certain point on the road by the vehicle detection means, and the image data obtained from the image pickup section is obtained. And detects a vehicle that has passed through the vehicle detection area, outputs a vehicle detection signal, and calculates the vehicle detection signal between the two vehicle detection areas set at a constant interval in the traveling direction of the vehicle by the calculating means. Since the information of the traffic flow is calculated based on the information of the rising time and the falling time, the time from the rising of the vehicle detecting signal of one of the two vehicle detecting areas to the rising of the other vehicle detecting signal, By comparing the time from the fall of the vehicle detection signal to the fall of the other vehicle detection signal,
Since it is possible to determine whether the detected area is a vehicle or a shadow formed by the vehicle, the number of vehicles can be measured with high accuracy. This has the effect that the average speed can be measured.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of a traffic flow measuring device according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram for explaining an operation of the first embodiment.

FIG. 3 is an explanatory diagram for explaining the operation of the first embodiment.

FIG. 4 is an explanatory diagram for explaining an operation of the first embodiment.

FIG. 5 is an explanatory diagram for explaining the operation of the first embodiment.

FIG. 6 is a flowchart showing the operation of the first embodiment.

FIG. 7 is an explanatory diagram for explaining the principle that transit times at a front end and a rear portion of a vehicle are different.

FIG. 8 is a diagram showing a configuration of a traffic flow measuring device according to another embodiment of the present invention.

FIG. 9 is an explanatory diagram for explaining the operation of the second embodiment.

FIG. 10 is a flowchart showing the operation of the second embodiment.

FIG. 11 is an explanatory diagram for explaining set values of inter-vehicle times β2 and γ2.

FIG. 12 is a diagram showing a configuration of a traffic flow measuring device according to another embodiment of the present invention.

FIG. 13 is an explanatory diagram for explaining the operation of the third embodiment.

FIG. 14 is a flowchart showing the operation of the third embodiment.

FIG. 15 is a diagram showing a configuration of a traffic flow measuring device according to another embodiment of the present invention.

FIG. 16 is a diagram showing a configuration of a traffic flow measuring device according to another embodiment of the present invention.

FIG. 17 is an explanatory diagram for explaining the operation of the fifth embodiment.

FIG. 18 is an explanatory diagram for explaining the operation of the fifth embodiment.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 input device 2 general-purpose image processing device 20 vehicle detection means 21 vehicle detection signal 31 vehicle flag setting means 32 time storage memory 34 noise removal means 35 single vehicle separation determination means 36 transit time comparison means 37 multiple vehicle connection determination means 38 Speed calculation means

Claims (7)

[Claims] An image pickup means for picking up an image of a road, a vehicle detection area is set at a certain point on the road, image data obtained from an image pickup section is processed, and a vehicle passing through the vehicle detection area is detected to detect a vehicle. Vehicle detection means for outputting a signal; and information on traffic flow based on information on the rise time and fall time of the vehicle detection signal between two vehicle detection areas set at a constant interval in the traveling direction of the vehicle. And a calculating means for calculating the traffic flow. 2. The computing device according to claim 1, wherein a time from a rise of one vehicle detection signal to a rise of the other vehicle detection signal and a time from a fall of one vehicle detection signal to a fall of the other vehicle detection signal. 2. A traffic flow measuring device according to claim 1, further comprising a transit time comparing means for comparing the object in the vehicle detection area to determine whether the object in the vehicle detection area has a height or a shadow having no height. 3. The noise calculating means for comparing a time from a rise to a fall of a vehicle detection signal with a predetermined minimum passage time in a measurement target area and excluding information less than the minimum passage time. 3. The traffic flow measuring device according to claim 1, further comprising a removing unit. 4. The computing means compares the time from the rise to the fall of the vehicle detection signal with the time from the fall to the rise of the next vehicle detection signal,
The vehicle according to claim 1, further comprising a single vehicle separation determining unit configured to determine whether the single vehicle is information separated into a plurality of single vehicles.
The traffic flow measuring device according to 2 or 3. 5. The computing means compares a time from a rise to a fall of one vehicle detection signal with a time from a rise to a fall of the other vehicle detection signal,
3. A multi-vehicle combination determining means for determining whether the information is a combination of a plurality of vehicles.
The traffic flow measuring device according to 3 or 4. 6. The calculating means determines a width of the vehicle,
6. The traffic flow measuring device according to claim 2, further comprising a vehicle width determining means for changing a criterion based on the determined vehicle width. 7. When the installation position of the imaging unit is on the roadside and the installation height is low, the calculation unit sets the vehicle presence detection area farther than the vehicle detection area in the back lane when viewed from the imaging position. 7. The traffic flow measuring device according to claim 1, 2, 3, 4, 5, or 6.