JPS5939799B2 - Nighttime signal processing method for traffic flow measurement equipment - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides a nighttime signal processing method in a traffic flow measurement device, and a video signal of a traveling vehicle is obtained using an optical system, and various traffic flow information is obtained from the waveform pattern of this video signal. The present invention relates to a method for processing signals at night in a device that obtains signals.

The waveform of the vehicle video signal output from the optical system is significantly different between daytime and nighttime.

During the day, the video signal is based on the light reflected from the vehicle body, but at night, the reflected light from the vehicle body is not detected.
Detects only headlight light.

Since the light from the headlights has high brightness, the video signal has a steep waveform due to this light.

This invention provides a signal processing method at night that can measure traffic flow information and (vehicle speed) using the same method by just slightly changing the constants, even though the waveforms of video signals are significantly different between daytime and nighttime. It provides:

Hereinafter, the method of the present invention will be explained in detail with reference to the drawings.

The example shown below is a case where a traffic flow measuring device installed at a position overlooking a road and having a large number of light receiving elements is used.

This device has the advantage of being able to collect vehicle information from multiple points on the road at one location and being easy to install.

FIG. 1 shows how the camera of the traffic flow measuring device is installed.

The camera 1 is mounted at a required height position (for example, 6 m) above the road using a support 3 or the like, and within a required range (for example, 6 m) in the length direction of the road.
For example, it is set up to be an enemy of Loom).

As shown in FIG. 3, the camera 1 is equipped with an optical system 4 including a lens 5 and a large number of light receiving elements.

In this example, within the field of view of the camera 1, there are six detection points P1 to P6 along the lane in which traffic flow is to be measured.

On the imaging plane of the optical system 4 in the camera 1 and corresponding to these detection points, one pair of light receiving elements ds1.
dR1-ds6. dR6 is located.

These light receiving elements are composed of photodiodes, for example.

When measuring traffic flow in two lanes with one camera, a large number of light receiving elements are arranged in two rows on the imaging plane of the optical system 4.

At each detection point (represented by P), as shown in Figure 2,
A set area S and a reset area R are set apart from each other by a required interval l.

A pair of light receiving elements (represented by dSdR) correspond to these regions S and R, respectively.

When the vehicle CA passes through the detection point P from the set area S to the reset area R, a vehicle detection video signal shifted by the time t is output from the light receiving elements dS and dR, as will be described later.

Then, the traveling speed of the vehicle is calculated based on this time t and distance l.

The processing device 2 obtains traffic flow information such as vehicle running speed, headway distance, degree of congestion, number of passing vehicles, and others based on video signals from the camera 1 .

With reference to FIG. 3, each light receiving element dS, dR in the camera 1
The output signal of is amplified by an amplifier 6 equipped with an automatic gain control function, and then sent to a multiplexer channel device 7, where the outputs of each of the 12 light receiving elements are sequentially switched and output from A to D. It is sent to converter 8.

The A-D converter 8 operates at a predetermined period (4.5 m5 in this example).
) performs sum conversion on the input outputs of the light receiving elements, and sends the result to a central processing unit (referred to as CPU) 11.

The processing device 2 is equipped with two CPUs 11 and 12.

Based on the data sent from the A-D converter 8, the CPU 11 executes analysis processing of the output signal waveform of each light-receiving element, and also analyzes the output signal waveform of each light receiving element, based on the data sent from the A-D converter 8. D converter 8
control.

Based on the data obtained by the CPU 11, the CPU 12 calculates the traveling speed of the vehicle, detects the degree of congestion, and performs other traffic flow calculation processes, creates necessary traffic flow information, and sends the transmitter (not shown). Then, it is transmitted to the center.

Each of the CPUs 11 and 12 includes a program memory (not shown) that stores an execution program, and local memories 13 and 14 that store data, respectively.
The processing unit 2 also has a shared memory 15 that can be accessed by both CPUs 11,12.

4a to 4h, a light receiving element dS is shown.

An example of a vehicle video signal output from dR is shown.

The set signal is the output of the light receiving element dS, and the reset signal is the output of the light receiving element dR.

Figure 4a shows the most common video signal.

Since the vehicle reflects more light than the road surface, when the vehicle reaches the detection point P, the output of the light receiving element becomes thicker than the level at which the road surface is detected (road surface level LO).

Then, when the vehicle passes, the output of the light receiving element returns to the road surface level LO again.

Therefore, the time from the rising edge of the set signal waveform to the rising edge of the reset signal waveform is defined as the detection time t.

Since this detection time t is the time required for the vehicle to travel between the set area S and the reset area R (distance l),
The traveling speed V of this vehicle is determined by the following equation.

However, K is a constant.

The distance that a vehicle actually moves during the detection time (t) is influenced by the shape of the vehicle, vehicle height, etc., and varies slightly from vehicle to vehicle, and is actually not equal to the above-mentioned distance l.

However, by imagining a sensing surface at a position above the road surface by the required height, measuring the distance between both areas S and R on this sensing surface, and statistically correcting this distance, it is possible to It is possible to measure speed with high precision.

The signal waveform shown in FIG. 4b has a peak reaching a very high level.

Therefore, a cutoff level L1 is set at an appropriate level, and the set signal and reset signal are respectively set at this cutoff level L1.
The difference in time when the time exceeds 1 is defined as the detection time t.

Then, using this detection time t, the traveling speed V is calculated using the above-mentioned formula.

The signal waveform shown in FIG. 4c includes a signal component lower than the road surface level LO at the front.

This signal component A is due to a shadow appearing in front or behind the vehicle.

When there is a shadow, the amount of light from that area decreases, so the signal becomes lower than the road level LO.

In the case of such a signal waveform, the time points at which the set signal and the reset signal each exceed the road surface level LO are detected, and the time difference between these times is defined as the detection time t.

In addition, for the waveform in Fig. 4c, in order to confirm that there is a signal component due to the vehicle following component A due to the shadow,
A vehicle confirmation level L2 is provided, and the signal level is road surface level L.
It increases across O and checks whether this level L2 is exceeded.

The light receiving element may detect the shadow of a vehicle traveling in an adjacent lane.

The signal due to this shadow exhibits a level lower than the road surface level LO as shown in the component, but may reach a level slightly higher than the road surface level LO at the rear.

However, this slightly higher level signal component at the rear never becomes higher than the confirmation level L2, so the confirmation level L2 makes it possible to distinguish between a signal caused by a vehicle and a signal caused by the shadow of a vehicle traveling in another lane. I can do it.

The signal waveform shown in FIG. 4d also has signal components A and B due to shadows.
Contains.

However, the fall of the reset signal occurs earlier than the fall of the set signal.

The shadow component B of the reset signal appears when the shadow of the own vehicle overlaps with the shadow of an adjacent vehicle traveling in another lane.

Even in the case of such a signal waveform, the time point at which each signal exceeds the road surface level LO (more precisely, the level L2) is detected, and the time difference between these time points is defined as the detection time t.

In the signal waveform shown in FIG. 4e, the reset signal rises earlier than the set signal.

Since such a phenomenon is impossible, treat it as an error.
The detection time, which is the basis for calculating the traveling speed, is not measured. The signal waveform shown in Figure 4f is a waveform that appears when the bus passes the detection point P, rising from the road surface level LO, increasing, and reaching the peak. When this occurs, this peak level remains flat for a while without increasing or decreasing, and then decreases to the road surface level LO.

Regarding such a signal waveform, the time difference between the rising points of the set signal and the reset signal is defined as the detection time t, as in the waveform of FIG. 4a.

A state in which the signal level neither increases nor decreases is called an equilibrium state.

When no vehicle is detected, the output of the light receiving element is at the road surface level LO, maintaining an equilibrium state.

Also, the peak level of the signal shown in Figure 4f is also in equilibrium.

In this way, there are two levels in the equilibrium state, and in order to distinguish between them, the road surface identification level L3 is set at a position midway between the two levels.

The signal waveforms shown in FIG. 4g are those seen at night, with each signal having a small peak C at the front, followed by a high level peak.

The small peak C is due to light reflected from the headlights from the road surface, and the high level peak is due to light directly from the headlights.

In the case of such a signal waveform, the front small peak C
, and the time difference at which the signal component with the rear high-level peak exceeds the cutting level L1 is defined as the detection time t.

Of course, at night, the road surface level LO also changes.

The signal waveform shown in Figure 4h is similar to that shown in Figure 4c and includes a component A due to shadows appearing in front or behind the vehicle.

However, in the signal shown in FIG. 4c, the falling edge of the reset signal occurs earlier than the point at which the set signal crosses the road level LO, whereas in FIG. occurs later than the time when the road crosses the road surface level LO.

For such signal waveforms as well, the time difference between the times when the set signal and the reset signal cross the road surface level LO is defined as the detection time t.

In order to detect the rise or fall of the output signal waveform of the light receiving element, it is necessary to determine the change in the signal.

The output signal of the light receiving element is A-D at each sampling period.
Since it is AD-converted by the converter 8, the current AD-converted result is set as the current data Dt.

In order to determine whether the current data Dt is increasing, decreasing, or in equilibrium compared to the previous data, it is necessary to find the deviation Δω between the current data Dt and the previous data. .

The previous data to be used for determining this deviation Δω is assumed to be preprocessed data Do.

This pre-processed data DO is, for example, previously sampled data, and is updated every sampling period.

Furthermore, for the above determination, the reference amount to be compared with the deviation Δω between the current data Dt and the preprocessed data Do is defined as the deviation reference amount ω0.

Then, if Dt-DO≧ωO, it is in an increasing state, Dt-D
If O≦−ωO, the state is decreasing, lDt −DOI<ωO
If so, it is assumed to be in an equilibrium state.

FIG. 5 shows how the change in the signal is determined for each sampling period in this manner.

In the above, for convenience, the preprocessed data (1) is updated every sampling period regardless of the signal state, but by introducing the concept of a prescribed period, this is updated to multiple times (for example, four times) the sampling period. The deviation Δω is compared with the reference amount ωO every sampling period, and the deviation Δ
It is preferable to update the preprocessing data by using the current data Dt as the preprocessing data DO when ω is greater than or equal to the reference amount 00, or when the prescribed period T has elapsed.

If the rise or fall of the signal is slow, the amount of increase or decrease may not reach the reference amount ωO during the sampling period.

The specified period is introduced to detect such slow changes in the signal, and the amount of change within the specified period is equal to the reference amount ωO.
Check whether it has been reached.

The method for detecting such a change in waveform can be summarized as follows.

If Dt-DO≧ωO within a prescribed period, it is regarded as an increasing state, and if Dt-Do≦−ω0 within a prescribed period, it is regarded as a decreasing state.

Then, when these determinations are made, the current data Dt is adopted as the preprocessed data DO.

Further, even if the specified period has elapsed, if lDt - DOI<ω0, it is considered as an equilibrium state, and the current data Dt is adopted as the preprocessing data DO.

Various data are stored in the memories 13 and 15 in order for the CPU 11 to execute a waveform analysis process of the output video signal of the light receiving element.

The memory 13 has an area for storing current data Dt, preprocessed data Do, deviation Δω, and road surface level LO of the set signal and reset signal, and an area for storing the starting pattern of the set signal.

The starting pattern means that the set signal is in an equilibrium state and the road surface level LO is maintained for a certain period of time (this is the end confirmation time T).
This is a pattern that shows a change in the state (increase or decrease) after more than 1) have elapsed.

This is used to check whether the vehicle detection signal first rose or fell.

The end confirmation time T1 is a time provided to confirm that one vehicle has passed the detection point in order to clearly distinguish individual vehicles.

The memory 13 also has areas used as various flags.

Set signal flag F1 and reset signal flag F11
indicates that one vehicle is passing through the set area S and the reset area R and vehicle detection signals are output, and the output of each light receiving element in the set area S and reset area R is at the road surface level. It rises or falls from LO and makes the necessary changes, then reaches an equilibrium state at road surface level LO, and ends with confirmation time T1 while maintaining road surface level LO.
It remains on until the time period elapses.

The set signal temporary end flag F2 and the reset signal temporary end flag F12 indicate that the end confirmation time T1 of each set and reset signal is being measured, and when the signal returns to the equilibrium state at the road surface level LO. It is turned on, and turned off when the end confirmation time T1 has elapsed.

The set signal start flag F3 is the detection time t (vehicle speed)
It is turned on when the set signal rises or falls from the road surface level LO, or when it crosses the road surface level LO in an increasing state, and the reset signal is turned on when the road surface level LO is measured. It is turned off when the vehicle rises from the road surface level, falls from the road surface level LO, and crosses the road surface level LO in an increasing state.

The set signal restart flag F4 indicates that the signal components A and B due to the shadow of the vehicle are being measured, and is turned on when the set signal falls from the road surface level LO. It is turned off when the road surface level is increased.

The repeat start flag F5 indicates that the detection time t is being measured based on the cutting level L1, and is turned on when the set signal crosses the cutting level L1 in an increasing state, and the reset signal is in an increasing state. It is turned off when it crosses the cutting level L1.

The recollection flag F6 indicates that the detection time t has been measured based on the cutoff level L1, and is turned on when the repeat start flag F5 is on and the reset signal crosses the cutoff level L1 in an increasing state. The reset signal is turned off when the reset signal returns to the road surface level LO and the end confirmation time T1 has elapsed.

The road surface crossing flag F7 is set when the set signal first falls to a level lower than the road surface level LO and then exceeds the road surface level LO, and is used to remember that the set signal has exceeded the road surface level LO. be.

This flag F7 is turned off when the reset signal similarly exceeds the road surface level LO.

The data abnormality flag F13 is turned on when the rising and falling orders of the set signal and the reset signal are reversed.

The reset signal start flag F14 is turned on when the reset signal starts rising or falling for the first time,
It will be turned off immediately.

Further, a predetermined area of the memory 130 is used as a vehicle speed counter C1, a re-vehicle speed counter C2, and an after-end time counter C3.

The vehicle speed counter C1 counts the detection time t based on the road surface level LO, and its contents are as follows: When the set signal crosses the road surface level LO in a rising, falling or increasing state, Cleared and timekeeping starts from these points.

The re-vehicle speed counter C2 measures a detection time t based on the cutting level L1, and is cleared when the set signal crosses the cutting level L1 in an increasing state.

Although these two counters C1 and C2 are expressed as if two are provided for convenience of explanation, one area may be shared by both counters C1 and C2.

The post-completion time counter C3 measures the completion confirmation time T1.

This counter C3 is provided one each for a set signal and a reset signal, but is represented by 03 to simplify the explanation.

If the counter C3 is set signal flag F1 (reset signal flag F11) is on and set signal temporary end flag F2 (reset signal temporary end flag F12)
is off, the set signal (reset signal) is cleared when it reaches the road surface level LO in an equilibrium state, and the timing operation starts from this point.

The shared memory 15 stores a deviation reference amount ωO1 cutting level L1,
It includes areas for storing vehicle confirmation level L2, road surface discrimination level L3, end confirmation time T1, and abnormal speed count number CM.

For example, when the shadow of a vehicle traveling in an adjacent lane is detected, the output signal of the light receiving element falls from the road surface level LO, undergoes a predetermined change, and then returns to the road surface level LO. However, due to the characteristics of the circuit (especially The amplifier 6) may temporarily reach a level slightly higher than the road surface level LO.

The vehicle confirmation level L2 is set to a level that is slightly higher than the level of the signal caused by this shadow and is as close as possible to the road surface level LO.

When the output signal of the light receiving element exceeds this level L2 in an increasing state, it is determined that this signal is caused by the vehicle and has exceeded the road surface level LO.

If the vehicle counters C1 and C2 continue counting and are not cleared forever, it is abnormal.

This occurs when the reset signal is output earlier than the set signal, or when noise enters only on the set side and the vehicle counters C1 and C2 start counting due to this noise.

The abnormal speed count number CM is set to a value slightly larger than the count number of the vehicle counter C1°C2 when counting the vehicle with the slowest expected speed, and the counters C1 and C2
If the count value exceeds this count number CM, it is considered an abnormal situation and the measurement process is stopped as described later.

The memory 15 also has an area used as a collection flag F15 and an end confirmation time transfer flag F16.

The collection flag F15 indicates that the measurement of the vehicle speed has been completed, and is turned on when the data abnormality flag F13 is off and the end confirmation time T1 of the reset signal has elapsed.

When this flag F15 is turned on, the CPU 12
takes in the speed data stored in the speed data area of memory 15.

When the CPU 12 finishes taking in the speed data, this flag F15 is turned off.

The end confirmation time transfer flag F16 is turned off when the end confirmation time T1 of the reset signal has elapsed.

The CPU 12 has a function of determining the end confirmation time T1 according to the loaded speed data, and changes the end confirmation time T1 for each vehicle speed measurement.

When the end confirmation time T1 selected by the CPU 12 is transferred to the shared memory 15, this flag F16 is turned on.

As the completion confirmation time T1, a short time is selected when the measured vehicle speed is fast, and a long time is selected when the measured vehicle speed is slow.

The memory 15 includes a speed data area for storing speed data by transferring the contents of the vehicle counters C1 and C2;
A pattern storage area is also provided for storing the state of change in the waveform of the reset signal.

The data in this pattern storage area is used in various waveform pattern processing.

FIG. 6 shows the procedure of waveform analysis processing of the set signal by the CPU 11.

This process is executed every 4.8 mB.

First, the AD-converted current data Dt is read, and the preprocessed data Do is subtracted from the current data Dt to obtain the deviation Δω (step 21).

Then, the absolute value of this deviation Δω is compared with the reference amount ωO to determine whether the change in the set signal is increased, decreased, or balanced (step 22). If the change is increased or decreased, the set signal temporary end flag F2 is turned off (step 23).

This flag F2 is always off when the set signal is in an increasing or decreasing state, and is turned off if it was on when the signal moves from equilibrium to increasing or decreasing.

Next, by looking at the sign of the deviation Δω, it is determined whether it is in an increasing state or a decreasing state (step 24), and if it is an increasing state, it is checked whether the current data Dt is equal to or higher than the cutting level 51 (step 24).
Step 25).

If the current data Dt is at the cutting level 51 or higher, it is checked whether the recollection flag F6 is on (step 26).

If the re-collection flag F6 is off, it is further checked whether the re-start flag F5 is on (step 27).

If this time the data Dt is at the cutting level 51 or above (YES in step 25) and both flags F (i and F5 are off), the set signal rises and increases until it exceeds the cutting level L1. Therefore, the vehicle speed counter C2 is cleared again (step 28) by turning on the re-start flag F5 (see FIG. 4b) (step 29).

As a result, as will be described later, the re-vehicle speed counter C2 starts counting the detection time t.

If NO in step 24.25, step 26゜27
If YES in step 29 and after the processing in step 29, the process proceeds to step 30.

In step 30, it is checked whether the set signal flag F1 is off.

Since the change state of the set signal is increase or decrease (YES in step 22), the set signal flag F1
If is off, it means that the set signal has started rising or falling.

Therefore, in order to see whether it is rising or falling, it is again determined whether it is increasing or decreasing (step 31).

If it is a decrease, it means a fall, and the set signal restart flag F4 is turned on (step 32) (see FIGS. 4c and 4d).

In the case of an increase, the process moves to step 33 without performing this process.

In this example, the road surface level LO is updated every time a vehicle is detected.

The road surface level LO can be updated at any time as long as no vehicle is detected, but in this case, when vehicle detection starts, that is, when the set signal rises or falls after the road surface level LO is in an equilibrium state. To (
(YES in step 30).

In step 33, the preprocessed data Do at that time is updated as the road surface level LO.

Then, turn on the set signal flag F1 and step 3
4) Turn on the set signal start flag F3 (step 35), and clear the vehicle speed counter C1 to measure the detection time t (step 36).

This starts the time counting operation of the counter C1 as described later.

Furthermore, the change pattern (increase or decrease) at the start of the set signal is stored (step 37), the preprocessed data is updated using the current data Dt as the preprocessed data Do (step 38), and the process ends.

If the set signal flag F1 is already on (NO in step 30), check whether it increases or decreases (step 40), and if it increases, the waveforms in Figures 4c, 4d, and 4h. Since the road surface level LO may be crossed as shown in FIG.

If the pattern at the start is a decrease, at the start the signal components A and B due to shadows fall from the road surface level LO, and then increase (YES at step 40).
Therefore, in order to check whether the signal level exceeds the vehicle confirmation level L2, the road surface liquid level LO is subtracted from the current data Dt to calculate the deviation Δω from the road surface level LO.
1 (step 42).

Then, this deviation Δω1 is compared with the confirmation level L2 (step 43), and if the deviation Δω1 is equal to or higher than level 2, it is assumed that the set signal exceeds the road surface level LO and is a signal from the vehicle, and the set signal is started. The flag F3 is turned on (step 44), the road crossing flag F7 is turned on (step 45), and the vehicle speed counter C1 is cleared to start counting the detection time t (step 46).
.

After this and in steps 40, 41, and 43 respectively, NO
In this case, the process moves to step 38.

The counter C1 is cleared when the set signal starts rising or falling from the road surface level LO (step 36) and when the set signal crosses the road surface level LO (step 46).

If equilibrium is found in step 22, the process moves to step 47.

As described above, it may be determined whether the pattern is a balanced pattern for each sampling period, or it may be determined whether the pattern is a balanced pattern in a time period of a prescribed period.

When the prescribed period is used as a reference, if there is equilibrium for each period, the process moves to step 47.

In step 47, it is checked whether the set signal temporary end flag F2 is on.

When this flag F2 is off, the set signal is being detected and the peak and level are in an equilibrium state as shown in FIG. 4F, or the road surface level LO is in an equilibrium state where no vehicle is detected.

Therefore, check whether the set signal flag F1 is on (
Step 48).

If this flag F1 is off, the road surface level LO is in an equilibrium state, so the process proceeds to step 38 and the current data Dt is set as the pre-processed data DO and the process is completed (step 38).
.

When the set signal flag F1 is on, either the peak level is in an equilibrium state, or the moment it falls and just reaches the road surface level LO and shifts to an equilibrium state.

Therefore, the current data Dt is compared with the road surface discrimination level L3 (step 49).

If the data Dt is level 53 or higher this time, step 38
Move to.

If the data Dt is less than level 53 this time, it has just reached the road surface level LO, so the set signal temporary end flag F2 is turned on (step 50), and after the end, the time counter C3 is cleared, and the time counter C3 is Preparation is made to start measuring the elapsed time after the end (step 51).

Then, the process proceeds to step 38.

If the set signal temporary end flag F2 is on (YES in step 47), the contents of the after-end time counter C3 are incremented by 1 to measure the time (step 52), and the contents of this counter C3 and the end confirmation time T1 are are compared (step 53).

If the content of counter C3 is less than time T1, step 3
8, if the contents of the counter C3 have reached the time T1, the counter C3 is cleared (step 54), the set signal flag F1 is turned off (step 55), and the set signal temporary end flag F2 is turned off. (Step 56
), proceed to step 38.

FIG. 7 shows the waveform analysis processing procedure of the reset signal by the CPU 11.

This process is also executed every 4.8 mS.

First, check whether the set signal start flag F3 is on (
Step 60), if this flag F3 is on, the content of the vehicle speed counter C1 is incremented by 1 and the vehicle speed (detection time t
) is counted (step 61).

Then, the contents of the vehicle speed counter C1 and the abnormal speed count number CM are compared (step 62).

If the content of the counter C1 is equal to or greater than CM, it is determined that the data is abnormal (see FIG. 4e).

The set signal start flag F3 is turned on when the set signal rises or falls and when the set signal crosses the road surface level LO.

Further, the set signal restart flag F4 is turned on when the set signal falls from the road surface level LO.

And these flags F3. When F4 is turned on, counter C1 is cleared and speed measurement begins.

In the case of abnormal data, in order to stop the speed measurement, the set signal start flag F3 and the set signal restart flag F4 are respectively turned off (steps 63, 64), and the vehicle counter C1 is cleared (step 65).

It is also preferable to detect the width of the reset signal to determine whether it is abnormal.

Next, check whether the re-start flag F5 is on (step 66), and if this flag F5 is on, add 1 to the content of the re-vehicle speed counter C2 to determine the vehicle speed (detection time t).
is counted (step 67).

Then, as in step 62, if the content of the re-vehicle speed counter C2 is greater than or equal to the abnormal speed count number CM (
(YES in step 68), the re-start flag F5 is turned off (step 69), and the re-vehicle speed counter C2 is cleared (step 70).

After the above vehicle speed counting process, the preprocessed data DO is subtracted from the current data Dt to find the deviation Δω (step 7
1), compare the absolute value of this deviation Δω with the reference amount ωO (
Step 72).

If the absolute value of the deviation Δω is equal to or greater than the reference amount 00, it means an increase or a decrease, so the reset signal temporary end flag F12 is turned off (step 73), and it is determined whether it is an increase or a decrease (
Step 74).

If it has increased, it is further checked whether the current data Dt has reached the cutting level L1 (step 75), and if it has reached it, it is checked whether the repeat start flag F5 is on (step 76).

If the current data Dt is equal to or higher than the cutoff level L1 and the repeat start flag F5 is on, the reset signal has reached the cutoff level L1 at this time, so the repeat start flag F5 is set in order to finish counting the vehicle speed. is turned off (step 77), and the recollection flag F6 is turned on (step 78) (see FIG. 4b).

Then, the contents of the re-vehicle speed counter C2 at this time are transferred to the speed data area of the memory 15 (step 7
9).

If NO in any of steps 74 to 76, the process immediately moves to step 80.

In step 80, it is checked whether the reset signal flag F11 is off.

Since the reset signal is in an increasing or decreasing state, if the reset signal flag Fil is off, the road surface level L
This is the time when the signal rises or falls from O for the first time.

Therefore, in step 8, the reset signal flag F11 and the reset signal start flag F14 are respectively turned on.
1.82), the preprocessed data DO at that time is stored as the road surface level LO (step 83).

Next, it is determined whether there is an increase or not to see whether it is rising or falling (step 84).

If it increases, it is a rising trend.

In the case of an increase, the vehicle speed can be measured if the set signal start flag F3 is already on (as shown in Figure 4a).
, it is abnormal if the set signal start flag F3 is off (FIG. 4e).

To confirm this, the state of the set signal start flag F3 is checked (step 85), and if this flag F3 is on, it is determined to be normal.

Since the detection time t has ended, the set signal start flag F3 is turned off (step 86), and just in case, the set signal restart flag F4 and the road crossing flag F7 are turned off (steps 87 and 88), and the vehicle The contents of speed counter C1 are transferred to the speed data area of memory 15 (step 89).

Then, the reset signal start flag F14 is turned off (
In step 90), the change in waveform is stored in the pattern storage area of the memory 15 as basic data for waveform processing (step 91), and the preprocessing data Do is updated with the current data Dt (step 92), and the processing is completed.

If the set signal start flag F3 is off in step 85, it means that there is an abnormality, so the data abnormality flag F13 is turned on (step 93) (see FIG. 4e) and the process moves to step 90.

In the case of a decrease in step 84, there is a signal component due to shadows, as shown in FIGS. 4c, 4d and 4h.

For such a signal, first check whether the set signal restart flag F4 is on (step 94), and if it is off, check whether the set signal start flag F3 is on (step 85). If F3 is also off, the data abnormality flag F13 is turned on (step 93
) (see Figure 4d).

If the set signal restart flag F4 is on, it is checked whether the road crossing flag F7 is on (step 95).

This flag F7 distinguishes the signals shown in FIG. 4c and FIG. 4h.

If the road crossing flag F7 is off (FIG. 4c), the state of the reset signal start flag F14 is further checked (step 96), and if this flag F14 is on, step 8
5, the state of the set signal start flag F3 is checked.

If this flag F3 is on, flag F3.
F4. Turn off F7 (steps 86 to 88).
), the contents of the counter C1 are temporarily transferred to the speed data area of the memory 15.

The contents of the counter C1 transferred in this case are the time from the fall of the set signal to the fall of the reset signal, but this content will later be changed from the time when the set signal exceeds the road surface level LO to the time when the reset signal reaches the road surface level. It is rewritten by the time (detection time t) up to the time when level LO is exceeded.

If the road surface crossing flag F7 is on (YES in step 95) (Fig. 4h), the set signal has already crossed the road surface level LO and the detection time t is being measured, so steps 86 to 89 are performed. Step 9 without
Go to 0.

If the reset signal flag F11 is already on in step 80, the reset signal has already risen or fallen and continues.

Therefore, it is checked whether it is in an increasing state or a decreasing state depending on the sign of the deviation Δω (step 97), and if it is an increase, the difference between the current data Dt and the road surface level LO is determined to check whether it has crossed the road surface level LO. Δω1 is determined (step 98), and it is determined whether this difference Δω1 is equal to or higher than the vehicle confirmation level 52 (step 99).

If the difference Δω1 is level 52 or higher, it is further checked whether the set signal start flag F3 is on (step 100).
), if this flag F3 is on, the data abnormality flag F
13 off (step 101) (see Figure 4d)
, proceed to step 86.

Through the processing from steps 86 to 89, each flag F3.
F4. The detection time t (content of counter C1) that has been measured since F7 was turned off and the set signal exceeded the road surface level LO is transferred to the speed data area of the memory 15 (Figures 4c and 4d). (see figure and figure 4h)
.

If NO in steps 97, 99 and 100, the process moves to step 91.

In the case of an equilibrium state (NO in step 72), first check whether the reset signal temporary end flag F12 is on (
Step 102), if this flag F12 is off, it is checked whether the reset signal flag F11 is on (step 103).

If this flag F11 is off, it means that the road surface is being detected, and the process immediately moves to step 91.

If the reset signal flag F11 is on, there is a possibility that the road surface level LO has just been reached and equilibrium has been achieved. Therefore, the current data Dt is compared with the road surface identification level L3 (step 104), and the current data Dt is level 53. If it is less than that, it is assumed that the road surface level has returned to LO, the reset signal temporary end flag F12 is turned on (step 105), and after the end, the counter C3 is cleared to measure the time (step 106), and the process moves to step 91. .

If YES in step 104, the process is assumed to be in an equilibrium state at the peak level (see FIG. 4f), and the process directly proceeds to step 91.

If the reset signal temporary end flag F12 is on, (
YES in step 102), time counter C after completion
Add 1 to the contents of step 3 to measure the time (step 107),
It is checked whether the contents of the counter C3 have reached the end confirmation time T1 (step 108).

If the end confirmation time T1 has not been reached, the process moves to step 91.

When this time T1 is reached, the counter C3 is cleared (step 109), and the flag F11. F12 is turned off (steps 110, 111), and the end confirmation time transfer flag F16 is turned off (step 112).

Then, check whether the data abnormality flag F13 is on (step 113), and if this flag F13 is on,
The process moves to step 117, where the data abnormality flag F13 is turned off, and the process moves to step 91.

If the data abnormality flag F13 is off, it means that the data is normal. Next, it is checked whether the recollection flag F6 is on (step 114).

When this flag F6 is on, the speed is measured based on the cutting level L1.

If the recollection flag F6 is on, if necessary, perform the necessary processing on the data based on the cutting level L1, and then turn off the flag F6 (step 11).
5).

Then, the collection flag F15 is turned on (step 11).
6), proceed to step 91.

Even in the case of NO in step 114, the collection flag F15 is turned on (step 116), and the process moves to step 91.

The CPU 12 checks the state of the collection flag F15 every 19.2 mS, and if the flag F15 is on, the CPU 12 takes in the measured detection time data from the speed data area of the memory 16 and performs necessary processing, such as vehicle speed. calculation, and the degree of traffic congestion.

When the CPU 12 takes in the detection time data from the memory 15, it turns off the collection flag F15.

Also, when the end confirmation time transfer flag F16 turns off,
The CPU 12 transfers the completion confirmation time, which has already been calculated based on the traveling speed of the preceding vehicle, to the memory 15, and changes this time.

After this, the flag 16 is turned on.

Since the road surface level LO represents the amount of light reflected from the road surface, it varies depending on the weather such as clear, cloudy, rainy, etc., and it differs significantly between daytime and nighttime.

Therefore, a day/night discrimination reference level (this is level L4) is set near the middle between the average road surface level during the day and the average road surface level at night, and the road surface level LO and the discrimination reference level at that time are set. By comparing L4, daytime and nighttime can be distinguished.

In addition, as the above-mentioned deviation reference amount ω0, a deviation reference amount ω01 for daytime use and a deviation reference amount ω02 for nighttime use are prepared (
.

Then, the deviation reference amount ω01 for daytime is set to a relatively small value, and the deviation reference amount ω02 for nighttime is set to a value larger than the difference between the level of the small peak C of the nighttime waveform shown in Fig. 4g and the road surface level LO. shall be.

Then, by adopting ω02 as the deviation reference amount at night, the small peak C will be determined to be in an equilibrium state,
This peak C is ignored, and the detection time can be measured based on the cutting level L1.

The procedure for such daytime and nighttime discrimination processing is shown in FIG.

Day/night discrimination reference level L4 and day/night deviation reference amount ω01
．． ω02 is stored in advance in the memory 15 (as shown).

First, compare the road surface level LO and the reference level L4 (
Step 201), if the road surface level LO is the reference bell 54 or higher, the deviation reference amount is set as the deviation reference amount ω01 for daytime use.
is set in the deviation reference amount area of the memory 15 (step 202).

If the road surface level LO has not reached the reference level L4, a night deviation reference amount ω02 is set (step 203).
).

This process may be performed at appropriate time intervals.

As explained in detail above, according to the present invention, a set area and a reset area are provided at a required distance from a vehicle detection point, a video signal of a vehicle passing through each of these areas is obtained, and this signal is converted into a predetermined signal. The signal is increased by sampling at the sampling period and using that data as the current data, finding the deviation between the current data and the previous predetermined reference data for each sampling, and comparing this deviation with the predetermined reference amount. In contrast to the daytime signal processing method, which detects whether the signal is in a decreasing or balanced state, determines the waveform pattern of the video signal, and obtains the required traffic flow information based on this determination result, it is By setting the reference amount to a larger value than during the daytime, the range of deviation detected as being in an equilibrium state is widened.

Therefore, the signal component that changes due to the reflected light from the road surface of the headlight of the vehicle at night can be kept within the range of the reference amount, and the change in the signal component is determined to be in an equilibrium state.

As a result, the video signal of a vehicle at night is considered to be only a steep waveform caused by direct light from the headlights, and can be processed using the same procedure as the video signal of a vehicle during the day, making it possible to process traffic flow information. It's convenient.

[Brief explanation of the drawing]

Figure 1 is a diagram showing the installation status of the camera of the traffic flow measurement device.
Fig. 2 is an enlarged view of the detection point, Fig. 3 is a configuration and block diagram showing the inside of the camera and processing device, Fig. 4 is a time chart showing the video signal waveform and the status of each flag, and Fig. 5 is a time chart showing the state of each flag. Explanatory diagram showing waveform pattern detection, No. 6
FIG. 7 is a flow chart showing the procedure of set signal processing, FIG. 7 is a flow chart showing the procedure of reset signal processing, and FIG. 8 is a flow chart showing the procedure of discrimination processing between daytime and nighttime. P, PI~P6...Detection point, S...
Set range, R...Reset range, dS1 to dR6
°°°°°°Photodetector, Dt...This data,
DO... Preprocessing data, ω0... Deviation reference amount, ω01... Daytime deviation reference amount, ω02
...Difference reference amount for nighttime, L4... Day and night judgment reference level.

Claims (1)

[Claims] 1 Set a set area and a reset area at a required distance from the vehicle detection point, obtain a video signal of a vehicle passing through each of these areas, sample this signal at a predetermined sampling period, and use the data as the current data. For each sampling, the deviation between the current data and the previous predetermined reference data is determined, and by comparing this deviation with the predetermined reference amount, it is possible to determine whether the signal is in an increasing, decreasing, or balanced state. In contrast to the daytime signal processing method, in which the waveform pattern of the video signal is detected and the required traffic flow information is obtained based on the result of this discrimination, in the nighttime, the range of deviation detected from the equilibrium state is widened. A nighttime signal processing method in a traffic flow measuring device, characterized in that the reference amount is set to a larger value than during the daytime in order to obtain a traffic flow measurement device, and the same signal processing as the daytime signal processing method is performed.