JP2851697B2 - Image type vehicle detector - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention sets a cross section detection line extending in the width direction of a multi-lane road on a video screen of a traffic flow measurement area of the multi-lane road, and detects the cross section. By comparing the reference luminance pattern when the vehicle does not pass on the line and the vehicle passing luminance pattern when the vehicle passes, the position and traveling speed of the vehicle traveling on the multi-lane road are measured. The present invention relates to a high-precision image type vehicle sensor.

[Conventional technology and problems to be solved by the invention]

As is well known, many ultrasonic vehicle sensors are put to practical use in traffic flow measurement sensors. However, when trying to measure the traffic volume on a multi-lane road in a city with heavy traffic, this ultrasonic vehicle sensor installs an ultrasonic head corresponding to each lane as shown in Fig. 34. In addition to being unfavorable in terms of urban aesthetics, the sensing area of the ultrasonic head is limited, and if the driving lane is changed due to illegal parking or snow in winter, it will deviate from the sensing area. There was a disadvantage that the running vehicle could not be detected (see FIG. 35).

Therefore, in order to solve such a problem of the ultrasonic vehicle sensor, an image vehicle sensor using a new vehicle detection method by image processing has been conventionally proposed. This type of image-type vehicle sensor includes a shift unit that compares the brightness of the entire screen or the average brightness of the measurement area with reference to the brightness of the road surface, and determines that there is a vehicle difference when there is a brightness difference exceeding a certain level. Detection method, spatial difference method for detecting the edge and contour of the vehicle by applying differentiation processing to the measurement area or the entire screen of the sample screen to be measured for traffic flow, taking the brightness difference between two screens at a certain time interval There is a frame difference method for extracting the edge and contour of a vehicle by performing a typical luminance differentiation.

However, in the above-described shift detection method in the image-type vehicle sensor, the luminance difference between the road surface luminance and the luminance of the entire screen or the average luminance of the measurement area changes every moment depending on weather conditions, the presence or absence of a vehicle, and the like. It is necessary to automatically follow this threshold by using a logical method such as conditional exponential averaging in accordance with various conditions of environmental changes. In addition to the complicated structure, if only the method of comparing the threshold value in the dimension of the luminance difference level is adopted to realize the traffic flow measurement by the image processing, for example, at dusk, as in a black vehicle, When the road luminance and the vehicle luminance are very close to each other, or when the vehicle image cannot be clearly recognized due to external factors such as rain or fog, the vehicle cannot be detected with high accuracy. Such a disadvantage in vehicle detection accuracy is that a spatial difference type image type vehicle sensor that differentiates a sample screen and detects a vehicle in a video screen based on a difference in luminance difference level,
Alternatively, the same can be said for an image type vehicle sensor of a frame difference type which detects a vehicle based on a luminance level temporally changed by taking a luminance difference between two screens.

Also, in the spatial difference method, since the entire screen is differentiated, it is necessary to detect a white line other than the vehicle, a median strip, and the like, and it is necessary to remove those other than the vehicle later. There are drawbacks in that the hardware configuration of the image processing becomes complicated and the processing time required for vehicle detection becomes longer.
Even in the frame difference method, since the entire screen is differentiated to extract the edges and contours of the vehicle, the hardware configuration of the image processing is complicated, and the processing time required for vehicle detection is long. There was a drawback of becoming.

The present invention has been made in view of the above-described disadvantages of the conventional image-type vehicle sensor,
It can function effectively even at twilight or at night to measure the position of a running vehicle, and even if the vehicle stops during traffic jams, it can prevent cracking of vehicle detection,
Further, it is a technical object of the present invention to provide a high-precision image-type vehicle sensor capable of measuring the speed of a traveling vehicle.

[Means to solve the problem]

Means adopted by the present invention to solve the above technical problem will be described below with reference to the accompanying drawings.

That is, according to the present invention, in order to improve the position measurement accuracy of a traveling vehicle, a photographing means (1) for photographing a traffic flow measurement area of a multi-lane road and outputting a video signal of the traffic flow measurement area is provided. Each of the samples of the cross-section detection line (L) consisting of a number of sample points (SP1 to SPn) set in the width direction of the multi-diagonal road on the video screen of the traffic flow measurement area by inputting a video signal from this photographing means An image luminance extracting means (2) for extracting and outputting an image luminance signal corresponding to the points (SP1 to SPn); and an image luminance signal from the image luminance extracting means (2), on the cross section detection line (L). A road reference luminance pattern (Ri) when the vehicle does not pass and a vehicle passing luminance pattern (Ci) when the vehicle passes are stored, and a plurality of continuous sample points in the direction of the cross-section detection line (L) are grouped. Of the specified length The measurement group (D) is sequentially scanned in parallel in the direction of the cross-section detection line (L) to calculate a correlation value between the road reference luminance pattern (Ri) and the vehicle passage luminance pattern (Ci) corresponding to each measurement group. A correlation curve (CL) in the direction of the cross-section detection line (L) is obtained, and when the correlation curve (CL) exceeds a detection threshold level (Ts), the correlation curve (CL) corresponds to a peak luminance level position of the correlation curve (CL). This means employs an image processing determining means (3) for outputting a vehicle detection signal.

Further, in order to prevent double detection of the vehicle due to the shadow of the vehicle, which is inevitably caused when the weather is fine, the image processing determination means (3) is provided on a road when the vehicle does not pass on the cross-section detection line (L). A difference luminance level (Ci-Ri) of a correlation value between the reference luminance pattern (Ri) and a vehicle passing luminance pattern (Ci) when the vehicle passes is calculated, and the difference luminance level (Ci-Ri) is negative, and Same difference luminance level (Ci−
A method is adopted in which a band in which the rate of change of Ri) is substantially constant is excluded and a correlation curve is obtained in the direction of the cross-section detection line (L).

In order to prevent erroneous detection of the actual vehicle when the headlights of the traveling vehicle are reflected on the road surface in front of the traveling vehicle at night, the image processing determination means (3) is used to determine whether the vehicle passage luminance pattern (Ci) Detection threshold level (T
p), the vehicle's passing luminance pattern (Ci)
And a means for outputting a vehicle detection signal corresponding to the peak luminance level position of the correlation curve (CL) of the road reference luminance pattern (Ri).

Further, even if the vehicle stops in front of the cross-section detection line during traffic congestion, the image processing determination means (3) detects the vehicle on the cross-section detection line (L) in order to prevent the detection crack of the vehicle detection. Later, for example, a luminance extraction position signal of a succeeding vehicle detection line (L ') extending in the traveling direction of the vehicle is output to the image luminance extraction means (2), and an image luminance signal of the subsequent vehicle detection line (L') is input. In this way, the vehicle on the following vehicle detection line (L ') is detected, and after the vehicle detection output of the following vehicle detection line (L') is lost, the position detection signal of the traveling vehicle is output. Adopted.

Further, in order to measure the speed of the traveling vehicle, the image processing determination means (3) is provided with a plurality of parallel cross-section detection lines separated by a predetermined distance in the longitudinal direction of the multi-lane road on the video screen of the traffic flow measurement area. The speed of the traveling vehicle is calculated from the relationship between the time (tab) at which the same vehicle passes on (La, Lb) and the distance (Lab) the vehicle has traveled within the time (tab), and the traveling speed signal of the vehicle is processed. Is output.

〔Example〕

Hereinafter, the present invention will be described in more detail based on embodiments shown in the accompanying drawings. 1 to 8 are explanatory diagrams for explaining the principle of vehicle measurement by the image-type vehicle sensor of the present invention, and FIGS. 9 to 29 are explanatory diagrams showing other vehicle measurement modes by the vehicle sensor. FIGS. 30 to 33 are explanatory diagrams of the speed measurement of the traveling vehicle by the vehicle sensor.

In FIG. 1, the reference numeral (S) indicates an image type vehicle sensor according to the present invention, and the reference number (T) indicates initial setting / maintenance used only at the time of initial setting and maintenance of the image type vehicle sensor (S). Tool (T).

The image type vehicle sensor (S) includes a CCD camera (1) as a photographing unit, an image luminance extracting unit (2) as an image luminance extracting unit, an image processing determining unit (3) as an image processing determining unit, and vehicle detection. It comprises a display (4).

As shown in FIG. 2, the CCD camera (1) is installed on a support provided on the shoulder side of the multi-lane road so as to overlook the traffic flow measurement area of the multi-lane road to be measured via an arm, The traffic flow measurement area is photographed, and its video signal is output to the next-stage image luminance extraction device (2).

Incidentally, the initial setting / maintenance tool (T) monitors the position data of the cross-section detection line (L), which will be described later, on the video screen of the traffic flow measurement area. An image sensor (5) which supplies the image signal from the CCD camera (1) to the 3).
And a portable computer (6) for inputting position data of the cross-section detection line (L) by operating a keyboard.

The image sensor (5) converts a video signal from the CCD camera (1) into an image A / D converter (5) via a video amplifier (51).
2) and input to the synchronization signal extraction circuit (53). The synchronization signal extraction circuit (53) extracts a horizontal synchronization signal and a vertical synchronization signal from the video signal from the image amplifier (51) and outputs the same to the image memory (54). And the data synthesis circuit (55)
Image signal from image A / D converter (52), image memory (5
By synthesizing the horizontal and vertical synchronizing signals from 4) and the cursor display signal fetched from the computer (6) to be described later via the input interface (56), and outputting them to the image D / A converter (57) Then, the monitor TV (58) projects the cursor indicated by the computer (6) together with the video screen of the traffic flow measurement area.

The computer (6) interactively inputs the position data of the cross-section detection line (L) and the guide text displayed on the display screen of the monitor TV (58) by monitoring the video screen of the monitor TV (58). As shown in the figure, each measurement line (M) is marked continuously for each lane in the width direction of a multi-lane road.
By specifying the start point position (M1s, M2s, M3s, M4s) and end point position (M1e, M2e, M3e, M4e) of the (1, M2, M3, M4) with the cursor (see FIG. 3), these start position ( M1s, M2s, M3
s, M4s) and the end position (M1e, M2e, M3e, M4e) are set position data of the cross-section detection line (L) via the output interface (7) through the image processing determination device (3) of the image-type vehicle sensor (S). ). Incidentally, the position data is printed out by the printer (61) of the computer (6) and utilized as reference data at the time of maintenance.

The image processing determination device (3) includes a microprocessor (31)
And a signal processor (32), and the microprocessor (31) is connected to each measurement line (M1, M2, M3, M) from the computer (6) of the tool for initial setting and maintenance (T).
4) Set position data (M1s, M2s, M3s, M4s, M1e, M2e, M3e,
When M4e) is input through the input interface (33), the microprocessor (31) operates according to the program in the ROM (31a), and the position data (M1s, M2s, M3
s, M4s, M1e, M2e, M3e, M4e), a number of sample points (SP1 to SPn) of the section detection line (L) extending in the width direction of the multi-lane road as shown in FIG. The luminance extraction position address of each sample point (SP1 to SPn) is output to the input interface (22) of the image luminance extraction device (2).

Image brightness extraction device for image type vehicle sensor (S) (2)
Extracts an image luminance signal corresponding to each sample point (SP1 to SPn) of the section detection line (L) from a video signal from the CCD camera (1), and extracts a signal processor (32) of the image processing determination device (3). The video signal from the CCD camera (1) is input by the video amplifier (21), and the luminance extraction position address from the microprocessor (31) of the image processing determination device (3) is input. Input from the interface (22).

The video signal from the CCD camera (1) is supplied to the clamp level fixing circuit (23) via the video amplifier (21).
And a synchronization signal extraction circuit (24). The clamp level fixing circuit (23) fixes the DC level of the black level of the video signal component from the video amplifier (21) to a constant level and outputs the fixed level to the video A / D converter (25) at the next stage, The signal extracting circuit (24) extracts a horizontal synchronizing signal and a vertical synchronizing signal of the video signal from the video amplifier (21), and converts the horizontal synchronizing signal into a horizontal address counter (26a) and the vertical synchronizing signal into a vertical address counter ( Output to 26b). The horizontal address counter (26a) includes a synchronization signal extraction circuit (2
The horizontal synchronization signal from 4) is counted and the horizontal address for one frame is sequentially output to the digital comparator (27). The vertical address counter (26b) counts the vertical synchronization signal from the synchronization signal extraction circuit (24). Then, the vertical address for one frame is output to the digital comparator (27).

On the other hand, as for the luminance extraction position address from the microprocessor (31), the horizontal luminance extraction position address and the vertical luminance extraction position address are supplied to the preset counter (28) via the input interface (22). The preset counter (28) counts the horizontal luminance extraction position address and the vertical luminance extraction position address of the first to n-th sample points (SP1 to SPn) and outputs them to the digital comparator (27).

The digital comparator (27) controls the horizontal address from the horizontal address counter (26a) and the horizontal luminance extraction position address from the preset counter (28), and the vertical address and preset counter (28) from the vertical address counter (26b). Compare the vertical luminance extraction position address from the
Outputs A / D sampling pulse to A / D converter (25).
As a result, the video A / D converter (25) A / D converts the video signal from the clamp level fixing circuit (23) to correspond to each sample point (SP1 to SPn) of the cross-section detection line (L). An image luminance signal is output to a signal processor (32) of the image determination device (3).

Each time the microprocessor (31) of the image processing determining device (3) receives an interrupt signal of the vertical synchronizing signal from the synchronizing signal extracting circuit (24) of the image luminance extracting device (2), the microprocessor (31) described later. The road reference luminance data and vehicle passing luminance data from
Store in 1b). Further, the microprocessor (31)
Is to output a brightness control signal to the CCD camera (1) via the D / A converter when the brightness level of the road reference brightness data and the vehicle passing brightness data input from the signal processor (32) is extremely low. And the image luminance extracting device (2) supplies an image luminance signal having a stable luminance level to the signal processor (32).

As shown in FIG. 4, the signal processor (32) of the image processing determination device (3) receives an A / D sampling pulse interrupt signal from the digital comparator (27) of the image luminance extraction device (2). It operates according to the program of the ROM (32a), and obtains a cross section from the image luminance signal corresponding to the luminance extraction position address of each sample point (SP1 to SPn) of the cross section detection line (L) from the video A / D converter (25). Road reference luminance data of each sample point (SP1 to SPn) when the vehicle does not pass on the detection line (L) and vehicle passing luminance data of each sample point (SP1 to SPn) when the vehicle passes The data is sequentially acquired in correspondence with the luminance extraction position address of each sample point (SP1 to SPn) input from the preset counter (28), and after the acquisition of all data is completed, the road reference luminance data and the Stored in the RAM (32 b) seeking tanks passage correlation value of the luminance data (Dj),
The vehicle passing position coordinates (Pi) are calculated from the detection threshold level (Ts) and the correlation curve (CL) of the correlation value (Dj) in the direction of the section detection line (L), and then the road reference luminance data and the vehicle passing luminance are calculated. The data is output to the microprocessor (31), and the road reference luminance data is input by outputting a transfer interrupt signal to the microprocessor (31) and stored in the RAM (32b).

The vehicle detector (4) inputs the vehicle detection signal and the traveling speed signal output from the microprocessor (31) of the image processing determination device (3) via the input interface (41), and the input interface (41) The running position signal output from the driving circuit (42) is passed through a driver circuit (42) to display the corresponding lane corresponding to the vehicle detection signal (43a,
43b, 43c, 43d) are turned on, and the traveling speed signal output from the input interface (41) is transmitted to the driver circuit (4
By passing the vehicle through 4), the segment display (45) digitally displays the traveling speed of the vehicle.

Next, a basic vehicle detection processing operation by the sensor of the present embodiment having such a configuration will be described with reference to FIGS.

For example, as shown in FIG. 5, when a CCD camera (1) outputs a video signal of a state where one vehicle approaches a cross-section detection line (L) to an image luminance extraction device (2), the image luminance extraction is performed. Since the image luminance signal of only the road surface is input to the device (2), the luminance level corresponding to each sample point (SP1 to SPn) of the road reference luminance data is substantially constant as shown in FIG. It becomes the road reference luminance pattern (Ri).

In the state shown in FIG. 5, when the CCD camera (1) outputs a video signal when the vehicle passes through the cross-section detection line (L) to the image luminance extracting device (2), the image luminance extracting device (2) Since the image luminance signals of the vehicle and the road surface are input, the vehicle passage luminance level is such that the luminance level of the sample point of the passing vehicle portion is higher than the luminance level of the sample point of the road surface portion as shown in FIG. It becomes the pattern (Ci).

Then, as shown in FIG. 6, the signal processor (32) of the image processing determination device (3) has a predetermined length of a plurality of continuous sample points (SP1 to SP10) in the direction of the cross-section detection line (L). The measurement group (D) is sequentially scanned in parallel in the direction of the cross-section detection line (L), and the individual measurement groups (d1 to
A correlation value (Dj) between the road reference luminance pattern (Ri) corresponding to d16) and the vehicle passing luminance pattern (Ci) is obtained. This correlation value (Dj) is given by a = (j-1) Dτ. Is calculated by the following general formula. Here, (j) is the calculation processing number of the measurement group (D), and (Dτ) is the measurement group (D)
Is the shift amount.

The correlation value (Dj) obtained by the signal processor (32) for the individual measurement groups (d1 to d16) in this manner is the seventh value.
As shown in the figure, only the vehicle pass band in the cross-section detection line direction becomes a correlation curve (CL) exceeding the detection threshold level (Ts). Then, the signal processor (32) detects the detection threshold level (Ts) at which the correlation curve (CL) intersects with the detection threshold level (Ts).
From the position coordinates (S1, S2) of FIG.
The vehicle passing position coordinates (Pi) corresponding to the peak luminance level in the vehicle passing band of L) are obtained by the following general formula.

The vehicle passing position coordinates (Pi) are supplied in time series from the signal processor (32) to the microprocessor (31), and the microprocessor (31) checks whether or not the vehicle is of the same vehicle. Then, the microprocessor (31) subjects the vehicle detection signal corresponding to the vehicle passing position coordinates (Pi) to time-series processing as shown in FIG. 8, and then sends this vehicle detection signal to the vehicle detector (4). Output.

As a result, the vehicle detector (4) receives the vehicle detection signal from the microprocessor (31) by using the driver circuit (42).
Then, the display lamps (43a, 43b, 43c, 43d) of the corresponding lane corresponding to the vehicle detection signal are turned on.

The sensor of this embodiment, which operates as described above, performs a series of processing operations at the timing as shown in FIG. 8, and sets a group of measurement groups (D) at 10 sample points and sets three lanes. In order to obtain the correlation curve of the section detection line of
The microprocessor (31) must perform about 160 8-bit subtractions, about 160 divisions, and about 160 8-bit additions, but use a signal processor (32) having a hardware multiplication / division mechanism. This makes it possible to perform 16-bit multiplication / division in several hundred nanoseconds, and to obtain the above-mentioned correlation curve in approximately 160 × 0.2 μsec × 3 = 0.24 msec. Therefore, one frame (1/30 sec. )
Can be dealt with with sufficient margin.

FIGS. 9 to 11 show a vehicle passing luminance pattern (Ci) stored in the RAM (31b) of the microprocessor (31) when two vehicles pass on the cross-section detection line (L) at the same time. 7 shows a correlation curve (CL) of each measurement group (d1 to d16) calculated and processed by the signal processor (32).
In such a case, the vehicle passing position coordinates (P1) corresponding to the two peak luminance levels in the vehicle passing band are obtained from the position coordinates of the detection threshold level (Ts) on the vehicle passing band in the correlation curve (CL). , P2), Therefore, the vehicle can be detected in the total cross-section traffic flow measurement independent of the traveling position of the vehicle.

In addition, when installing a CCD camera (1) on a multi-lane road, it is considered that the installation height is limited to 10 m due to maintenance. Therefore, as shown in FIG. 12, an arm of about 2 m is required to prevent a vehicle image of a vehicle height of about 3 m traveling in the second lane from overlapping the third lane. Even so, the vehicle image protrudes into the adjacent lane and the vehicle image in the second lane and 3
Vehicle images in the lane may overlap. Figures 13 to 15
The figure shows a microprocessor (3) when two vehicles pass on the cross-section detection line (L) at the same time while approaching each other.
1) The vehicle passage luminance pattern (C stored in the RAM (31b))
i) and the correlation curve (CL) of each measurement group (d1 to d16) calculated by the signal processor (32).
If the vehicle passing band of the correlation curve (CL) exceeding the detection threshold level (Ts) due to the overlapping of the vehicle images is two or more, the signal processor (32) divides the band into two and divides the band into two. Position coordinates (P1, P
2) Therefore, it is possible to individually detect vehicles that have simultaneously passed through the cross-section detection line (L) in a close state.

However, even if the vehicle pass band of the correlation curve (CL) is divided into two and the peak luminance levels are individually calculated, one of the vehicle pass bands may be a shadow of a vehicle in fine weather (No. 16). As shown in FIG. 17, the road reference luminance pattern (Ri) is subtracted from the vehicle passing luminance pattern (Ci) to obtain difference luminance data (Ci−Ri), and the difference luminance data (Ci−Ri) is obtained. If the correlation curve (CL) is obtained excluding the band of the measurement group in which the change rate of the difference luminance data (Ci−Ri) is negative and the difference luminance data (Ci−Ri) is negative, the shadow data of the vehicle that occurs in fine weather is removed. Therefore, double detection of the vehicle can be prevented.

In the detection of a vehicle at twilight, for example, when one vehicle passes through the cross-section detection line (L), the vehicle passing luminance pattern (Ci) rises and falls according to the image luminance signal based on the vehicle and the shadow data of the vehicle. Although it fluctuates slightly (18th
Since the correlation curve (CL) between the reference luminance pattern (Ri) and the vehicle passing luminance pattern (Ci) rises significantly in the vehicle passing band as shown in FIG. 19, the vehicle passing position coordinates can be obtained without any trouble. it can.

In the vehicle detection at night, when one vehicle passes through the cross-section detection line (L) (see FIG. 20), the vehicle passing luminance pattern (Ci) is determined by an image luminance signal from the headlights of the vehicle. Since a large rise occurs at the headlight portion (see FIG. 21), as shown in FIG. 22, a correlation curve (C
L) can be obtained. Therefore, the vehicle passing position coordinates at the peak luminance level of the vehicle passing band may be obtained.

However, in vehicle detection at night, headlights may be reflected on the road surface in front of the vehicle and erroneously detected as the presence of a real vehicle. Therefore, as shown in FIG. When the peak luminance level (PK) of the pattern (Ci) exceeds the detection threshold level (Tp), the vehicle passing position coordinates of the peak luminance level in the vehicle passing band of the correlation curve (CL) shown in FIG. 24 are obtained. The above problem is solved by programming the signal processor (32).

In the vehicle detection at night, there is a possibility that a small light at the rear of the vehicle may erroneously detect the presence of a real vehicle, but when the vehicle passes on the cross-section detection line (L) (No.
25, the width (ΔL ′) of the peak luminance level of the small light portion of the vehicle passage luminance pattern (Ci) is the 20th.
Since the width (ΔL) of the peak luminance level of the headlight portion shown in the figure is 3/1 to 4/1 (see FIG. 26), it can be identified based on these peak luminance level widths (ΔL, ΔL ′). .

If the height of the CCD camera (1) is set at 10 m as shown in Fig. 27, even if a bus with a vehicle height of 3m and a vehicle length of 9m arrives, the camera will include points A and B in the same figure. If the field of view is set, the minimum inter-vehicle distance will be 2 m or less, and sufficient image resolution on vehicle images will be obtained. However, when the vehicle stops during the stagnation due to the trouble of the cross-section detection line (L), the detection cracking of the vehicle detection may occur only with the detection holding time shown in FIG. 8, so the image processing determination device (3) Is configured as described below to prevent the detection crack of the vehicle detection. That is, upon detecting a vehicle on the cross-section detection line (L), the image processing determination device (3) detects a following vehicle corresponding to about 2 m on an actual road in a direction opposite to a vehicle traveling direction with respect to a passing position of the vehicle. The horizontal luminance extraction position address and the vertical luminance extraction position address of the line (L ') (see FIG. 28) are output to the image luminance extraction device (2), and the luminance extraction position signal of the following vehicle detection line (L') is obtained. A correlation curve in the direction of the following vehicle detection line (L ') is obtained by the same method as described above, and when the correlation curve exceeds the detection threshold level, the vehicle on the following vehicle detection line (L') is detected. Is detected, and the vehicle detection signal is output to the vehicle detector (4) as the final detection output after the vehicle detection output of the following vehicle detection line (L ') disappears (see FIG. 29).

Next, the traveling speed detection processing operation of the sensor of this embodiment is
This will be described with reference to FIGS. 30 to 33.

For example, as shown in FIG. 30, two parallel cross-section detection lines (La, L) separated by a predetermined distance in the longitudinal direction of a multi-lane road on the video screen of the traffic flow measurement area of the monitor TV (58).
b) is set in the width direction of the multi-diagonal road, and vehicle detection is performed on these cross-section detection lines (La, Lb) in the same manner as described above. Then, the image processing determination device (3)
As shown in Figure 31, the same vehicle has both cross-section detection lines (La, L
b) The travel speed signal of the vehicle speed V obtained from the general formula V = Lab / tab is obtained from the vehicle detector based on the time (tab) passing through (tab) and the distance (Lab) traveled by the vehicle within the time (tab). Output to (4).

As a result, the vehicle sensor (4) passes the travel position signal from the microprocessor (31) to the driver circuit (44), so that the travel speed is digitally displayed on the segment display (45).

Therefore, it is possible to detect not only the speed of a vehicle traveling in a predetermined traveling lane, but also the speed of a vehicle whose traveling position is shifted due to snow in winter or illegal parking.

However, if the vehicle runs at 120 km / h, 120,0
00 (m) / 60 (minutes) x 60 (minutes) x 30 = 1.1m, which means that the video signal moves 1.1m in one frame (1 / 30sec), so the distance (Lab) is about 10m. In order to measure the speed of a traveling vehicle with an accuracy of 10% or less, the vehicle position at the distance (Lab) must be measured with an accuracy of ± 1 m or less. In such a case, the CCD camera (1) uses a shutter camera as a general rule to use a shutter speed of 120k.
It is necessary to make the image of a m / h high-speed running vehicle a blur-free image and perform position measurement with high accuracy. For this purpose, as shown in FIG. 32, two cross-section detection lines (La ′, Lb) corresponding to about 1 m on an actual road are arranged before and after the cross-section detection lines (La, Lb).
b ') and set the outer cross-section detection lines (La', L
b ′) is set to (Lab), and the shutter speed of the CCD camera (1) is set to 1/100 sec per field as shown in FIG. 33, and the vehicle passes the distance (Lab). Assuming that the time is (tab), it is possible to measure the position of the vehicle with an error of ± 1 m, and thus it is possible to detect the speed of the vehicle traveling at high speed with high accuracy.

〔The invention's effect〕

As described above, according to the present invention, since the line detection processing is based on the pattern comparison between the road reference luminance pattern and the vehicle passing luminance pattern on the cross-section detection line, the vehicle travels from day to dusk to night with the same algorithm using the correlation curve. The traveling position of the vehicle can be measured, and the traveling position of the traveling vehicle can be measured with high accuracy no matter where the vehicle on the multi-lane road passes.

Further, a difference luminance level between the road reference luminance pattern and the vehicle passing luminance pattern is calculated, and the difference luminance level is negative,
In addition, by excluding a band in which the change rate of the difference luminance level is substantially constant, the correlation curve is obtained by removing the shadow data of the vehicle appearing in the vehicle passage luminance pattern in fine weather. Only reliable detection is possible.

Further, by comparing the vehicle passing luminance pattern with the detection threshold level, the band of the headlight portion of the vehicle that appears in the vehicle passing luminance pattern at night can be specified, and therefore, the vehicle traveling at night can be reliably detected. Becomes possible.

Further, after detecting a vehicle on the cross-section detection line, a vehicle on the subsequent vehicle detection line is detected, and a vehicle detection signal is output after the detection output of the subsequent vehicle detection line is lost. The detection crack of the vehicle detection can be prevented.

Then, by performing line detection processing on a plurality of parallel cross-section detection lines separated by a predetermined distance in the longitudinal direction of the road, the time when the same vehicle passed through each cross-section detection line and the distance that the vehicle moved within the time , The speed of the traveling vehicle can be measured.

Therefore, the image-type vehicle sensor of the present invention, in fine weather,
Even at dusk or at night, the position of the running vehicle can be measured with high accuracy, and even if the vehicle stops in front of the cross-section detection line during a traffic jam, the detection crack of the vehicle detection can be prevented, and the speed of the running vehicle It can be used as a vehicle detector on multi-lane roads.

[Brief description of the drawings]

1 to 8 are explanatory views for explaining the principle of vehicle measurement by the image-type vehicle sensor of the present invention. FIG. 1 is a block diagram of the image-type vehicle sensor and a tool for initial setting and maintenance. FIG. 2 is a perspective view showing an installation state of the CCD camera, and FIG.
FIG. 4 is an explanatory view of a measurement line for setting position data of a section detection line, FIG. 4 is a flowchart showing the operation of a signal processor of the image processing apparatus, and FIG. 5 is a state in which one vehicle arrives at the section detection line. FIG. 6 is an explanatory view showing a road reference luminance pattern, a vehicle passing luminance pattern and a method of calculating a correlation value in the state of FIG. 5, and FIG. 7 is a road reference luminance pattern and a vehicle passing luminance shown in FIG. FIG. 8 is an explanatory diagram of a pattern correlation curve, and FIG. 8 is a time chart showing a sensing next-sequence process of the image processing determination device. 9 to 29 are explanatory views showing another mode of vehicle measurement by the vehicle sensor. FIG. 9 is an explanatory view showing a state where two vehicles are separated from each other and arrive at a cross-section detection line. FIG. 12 is an explanatory diagram of a road reference luminance pattern and a vehicle passing luminance pattern in the state of FIG. 9, FIG. 11 is an explanatory diagram of a correlation curve between the road reference luminance pattern and the vehicle passing luminance pattern shown in FIG. 10, and FIG. FIG. 13 is an explanatory view showing a specific installation state of a CCD camera, FIG. 13 is an explanatory view of a state in which two vehicles approach and arrive at a cross-section detection line, and FIG. 14 is a road reference luminance pattern in the state of FIG. And FIG. 15 is an explanatory diagram of a correlation curve between a road reference luminance pattern and a vehicle passing luminance pattern shown in FIG. 14, and FIG. 16 is an explanatory diagram of a vehicle arriving at a cross-section detection line in fine weather. , Figure 17 shows the road in the state of Figure 16 FIG. 18 is an explanatory view showing a quasi-luminance pattern, a vehicle passing luminance pattern, and difference luminance data between a road reference luminance pattern and a vehicle passing luminance pattern, FIG. 18 is an explanatory diagram of a road reference luminance pattern and a vehicle passing luminance pattern at dusk, FIG. Is an explanatory diagram of a correlation curve between a road reference luminance pattern and a vehicle passing luminance pattern shown in FIG. 18,
Figure 20 is an explanatory diagram of vehicles arriving at the cross-section detection line at night,
21 is an explanatory diagram of a road reference luminance pattern and a vehicle passing luminance pattern in the state of FIG. 20, FIG. 22 is an explanatory diagram of a correlation curve between the road reference luminance pattern and the vehicle passing luminance pattern shown in FIG. 21, and FIG. FIG. 24 and FIG. 24 show a method of preventing actual vehicle detection in vehicle detection at night, and FIG. 23 is an explanatory diagram of a road reference luminance pattern at night and a vehicle passing luminance pattern by headlights of the vehicle. No.
FIG. 24 is an explanatory diagram of a correlation curve between the road reference luminance pattern and the vehicle passing luminance pattern in FIG. 23, FIG. 25 is an explanatory diagram of a vehicle arriving at the cross-section detection line at night, and FIG. 26 is a road reference luminance pattern at night FIG. 27 is an explanatory diagram of a vehicle passing luminance pattern by a small light of the vehicle, FIG. 27 is an explanatory diagram showing a relationship between a CCD camera installation and an inter-vehicle image resolution at a stagnation, FIG.
FIG. 29 is an explanatory diagram of the succeeding vehicle detection line, and FIG. 29 is a time chart showing the sensing next-sequence processing of the image processing determination device in the succeeding vehicle detection line. 30 to 33 are explanatory diagrams of the speed measurement of the traveling vehicle by the vehicle sensor, FIG. 30 is an explanatory diagram of setting a cross-section detection line in measuring the speed of the traveling vehicle, and FIG. 31 is a cross-sectional view of the vehicle. Explanatory diagram showing a detection output when passing through a detection line,
FIG. 32 is an explanatory diagram of setting of a cross-section detection line in measuring the speed of a high-speed traveling vehicle, and FIG. 33 is a time chart showing the sensing sequence processing of the image processing determination device in measuring the speed of the high-speed traveling vehicle. FIG. 34 and FIG. 35 are explanatory views showing the installation state of a conventional ultrasonic vehicle sensor. (1): photographing means, (2): image luminance extracting means, (3): image processing determining means, (L): cross-section detecting line, (L '): succeeding vehicle detecting line, (SP1 to SPn): sample Point, (Ri): Road reference luminance pattern, (Ci): Vehicle passing luminance pattern, (D): Measurement group, (Ts): Detection threshold level, (Ci−Ri): Difference luminance level.

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁶ identification code FI G08G 1/04 G01V 9/04 S (58) Investigated field (Int.Cl. ⁶ , DB name) G01B 11/00-11 / 30 102 G01J 1/00-1 / 60,11 / 00 G01P 3/36 G08G 1/00-9/02 G06T 1/00-7/00

Claims (5)

(57) [Claims] An imaging means for photographing a traffic flow measurement area on a multi-lane road and outputting a video signal of the traffic flow measurement area.
A video signal is input from the photographing means, and each of the cross-section detection lines (L) formed of a number of sample points (SP1 to SPn) set in the width direction of the multi-diagonal road on the video screen of the traffic flow measurement area. Image luminance extracting means (2) for extracting and outputting an image luminance signal corresponding to the sample points (SP1 to SPn)
And a road reference luminance pattern (Ri) when the vehicle does not pass on the cross-section detection line (L) and a vehicle passing luminance when the vehicle passes from the image luminance signal from the image luminance extracting means (2). A measurement group (D) having a specified length, which stores a pattern (Ci) and has a plurality of continuous sample points in the direction of the cross-section detection line (L), is sequentially scanned in parallel in the direction of the cross-section detection line (L). Then, the correlation value between the road reference luminance pattern (Ri) and the vehicle passing luminance pattern (Ci) corresponding to each measurement group is calculated, and the cross-section detection line (L) is calculated.
Image processing for obtaining a direction correlation curve (CL) and outputting a vehicle detection signal corresponding to a peak luminance level position of the correlation curve (CL) when the correlation curve (CL) exceeds a detection threshold level (Ts). An image-type vehicle sensor comprising: a determination unit (3). 2. The image processing determining means (3) includes a road reference luminance pattern (Ri) when a vehicle does not pass and a vehicle passing luminance pattern (Ci) when a vehicle passes on a cross-section detection line (L). ) Correlation luminance difference level (Ci−
Ri) is calculated, and the difference luminance level (Ci−Ri) is negative,
2. The apparatus according to claim 1, wherein a correlation curve is obtained in the direction of the cross-section detection line (L), excluding a band in which a change rate of the difference luminance level (Ci-Ri) is substantially constant. The image type vehicle detector. 3. The image processing judging means (3), wherein a vehicle passing luminance pattern (Ci) is detected by a detection threshold level (Tp).
, A vehicle detection signal corresponding to a peak luminance level position of a correlation curve (CL) between the vehicle passing luminance pattern (Ci) and the road reference luminance pattern (Ri) is output. The image-type vehicle sensor according to claim 1, wherein 4. The image processing judging means (3) detects a vehicle on the cross-section detection line (L) and then outputs a luminance extraction position signal of a following vehicle detection line (L ') extending along the traveling direction of the vehicle. Is output to the image luminance extracting means (2), and the image luminance signal of the following vehicle detection line (L ') is inputted to detect the vehicle on the following vehicle detection line (L'), and this succeeding vehicle detection is performed. 2. The image type vehicle sensor according to claim 1, wherein a vehicle detection signal is output after the vehicle detection output of the line (L ') is lost. 5. The image processing judging means (3) comprises a plurality of parallel cross-section detection lines (La, L) separated by a predetermined distance in the longitudinal direction of a multi-lane road on a video screen of a traffic flow measurement area.
b) The speed of the traveling vehicle is calculated from the relationship between the time (tab) at which the same vehicle passes above and the distance (Lab) traveled by the vehicle within the time (tab), and the traveling speed signal of the vehicle is output. The image-type vehicle sensor according to claim 1, wherein the image-type vehicle sensor is configured as follows.