JP2732860B2 - Vehicle detection method and vehicle detection device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Summary] A vehicle is determined by a difference between density data of two images having gradations photographed at a fixed time interval of about several tens of milliseconds to several seconds and a difference image obtained from the difference. The moving direction is determined based on the positive and negative polarities.

In addition, the accuracy of the determination can be improved by adding that the widths of the polarities of the positive and negative differences generated as a result are equal to each other as a determination condition, or by increasing the density difference from the background of the vehicle.

[Industrial applications]

The present invention relates to a vehicle detection method that detects the movement of a vehicle in a parking lot or a predetermined area on a road, and is applicable to parking lot management, illegal parking monitoring, traffic congestion monitoring, and the like.

[Conventional technology]

Japanese Patent Application Laid-Open No. 60-27997 discloses a method of generating a binary image of a video signal of a road or the like taken by a television camera in order to measure a traffic flow, based on a certain threshold value. An apparatus for identifying a moving vehicle using an image is disclosed.

Further, a system has been put to practical use in which a sensor is provided at a dedicated entrance of a parking lot, the number of entering and exiting vehicles is counted, and the number of parked vehicles is calculated.

[Problems to be solved by the invention]

However, in the method of first performing binarization and performing the difference calculation, stable measurement cannot be performed when the density difference between the vehicle and the background is small.

In addition, in the case of outdoors, the brightness of the background greatly changes depending on the weather and the time zone, and thus the threshold value for binarizing the image taken by the television camera is set and maintained at an appropriate value. It was difficult to obtain an accurate image.

Furthermore, in the conventional parking lot management, since the number of vehicles entering and exiting is counted by the sensor at the parking lot entrance / exit section, if a counting error occurs, this error accumulates and the actual There was a problem of being different.

[Means for solving the problem]

Images are taken at regular time intervals of about several tens of milliseconds to several seconds into digital density data having gradations, and the difference between the density data at the same position of the two preceding and succeeding images is determined to extract the vehicle. Further, the moving direction of the vehicle is determined based on the positive and negative polarities of the difference in the difference image obtained thereby.

If the density difference between the background and the vehicle is small and it is difficult to identify the vehicle, the outline of the vehicle in the captured image is emphasized and the gradation conversion inside the outline is performed to perform the density difference between the background and the vehicle. Is increased, the accuracy of the determination can be increased.

In addition, the accuracy of the determination can be improved by adding to the determination conditions that the widths of the positive and negative polarities of the difference are equal.

As an example of an apparatus for embodying the vehicle detection method according to the present invention, a photographing means for photographing a predetermined area at predetermined time intervals of several tens of milliseconds to several seconds, and an image signal photographed by the photographing means are stored in a floor. An A / D converter for converting to digital density data having a tone, and a difference operation between digital density data at the same position of two images before and after taken at the above-mentioned fixed time interval, and a difference image obtained as a result is obtained. A vehicle detection device can be constituted by a microprocessor that extracts a vehicle based on the positive and negative polarities of the vehicle and determines the moving direction of the vehicle.

The parking condition, the number of parked vehicles and the occupancy of the parking lot can be determined by adding the functions of vehicle entry / exit to a predetermined area, parking / vacancy determination, and vehicle number calculation functions to the microprocessor. It monitors illegal parking on the road by adding a parking / stop judgment function.

Further, by adding a function of determining the speed and the distance between vehicles to the microprocessor, a function of monitoring traffic congestion on the road can be provided.

(Operation)

FIG. 1 is a view showing the principle of the present invention.
Shows the image position of the moving vehicle M at time T, and (b) shows the image position of the vehicle M at time T + t. When the image data of (b) is subtracted from the image data of (a), FIG. As shown in c), a region m having a positive polarity and a region m 'having a negative polarity are generated in the difference image, and only the moving vehicle is extracted.

In the conventional method, since the binarization processing of the image is performed first, there is a problem that threshold control of the binarization following the change in the luminance of the background is required, and only an unclear vehicle can be extracted.

On the other hand, in the present invention, since the difference between the two original images before and after at a short time interval t of about several tens of milliseconds to several seconds is obtained, the movement without being affected by the change in the brightness of the background. Since only the vehicle is stably extracted, the problems of the conventional method as described above are solved.

FIG. 9D shows the density value on the scanning line L in FIG. 9C, and the above-described case where the different direction of the vehicle M moves from the direction of FIG. 9B to the direction of FIG. If the movement is in the opposite direction, the difference between the above areas is reversed between the positive (+) and negative (-) areas indicated by n 'and n in FIG. The moving direction can be determined.

In the above description, the method using the forward difference operation of subtracting the subsequent image data from the previous image data has been described. However, even if the backward difference operation of subtracting the previous image from the subsequent image is performed, the polarity is positive or negative with respect to the moving direction. Is the same except that the reverse occurs.

When one vehicle moves, the widths of the areas m and m 'shown in FIG.
₁ is equal Habamo These differences indicated by x _2, if a plurality of vehicles such when the difference occurs is the width of these differential area x _1, x ₂ generally equally by the simultaneous movement Since there is no vehicle, it can be determined whether or not one vehicle has moved.

Since the density value of asphalt generally used for pavement such as roads and parking lots is generally smaller than the density value of the vehicle (close to black), when the vehicle moves on the asphalt, the above-mentioned FIG. As shown in (c), the difference between the density values at the portion m where the vehicle is no longer present becomes a value indicating that the vehicle has become brighter, that is, has become brighter.

If the density value of the vehicle is almost equal to the density value of asphalt, it is unlikely that the vehicle will be indistinguishable due to the presence of bright parts such as windowpanes and bumpers. If the density of the vehicle is lower than the density value of the asphalt, it is determined that the vehicle is moving in the opposite direction.

In order to solve such a problem, an image in which the outline of the vehicle is emphasized from the captured image and a gradation conversion is performed so as to increase the density difference between the outline and the background such as a road is generated, By determining the difference, the moving direction can be determined.

〔Example〕

FIGS. 2 to 4 show an example of a parking lot management system to which the vehicle detection method of the present invention is applied. The principle will be described first with reference to FIGS. 2 and 3. FIG.

FIG. 2 shows a parking area S ₁₁ , S ₁₂ ... S _1n , S ₂₁ , which is separated by a size such that a television camera C provided at a position overlooking the entire parking lot and one vehicle can be accommodated. S ₂₂ ... S _2n are shown. In this embodiment, management of parked vehicles is performed for each of these parking areas.

FIG. 3 shows an example of a screen on which an image taken by the television camera C is monitored for explanation, and a first vehicle entry / exit detection area S including parking areas S ₁₁ , S ₁₂ ... S _1n.
1, the parking area S _21, S ₂₂ ...... second vehicle and out detection region S ₂ containing the S _2n are displayed, enter the respective coordinates and the maximum available parking number of parking areas in advance in the management system Keep it.

It is preferable that the coordinates of the detection area be set as a rectangular area including each parking area as shown by a dotted line in FIG. 3 in consideration of ease of raster scanning. Even if it is not suitable for raster scanning due to the installation position of the camera, it is preferable that each detection area be a rectangle including the corresponding parking area.

Further, in the present invention, only a moving object such as a moving vehicle is extracted from the image by the difference processing. Therefore, when the image after this processing is displayed, the parking area shown by the solid line in FIG. It goes without saying that a stationary portion such as a line shown is not displayed on the screen.

The third in the first vehicle and out detection region S ₁ on the screen of the figure, the vehicle input and detection area S ₁ moves to the right of the vehicle had been in a certain parking area indicated by "→ A" in FIG. If the vehicle deviates, it can be determined that the parked vehicle has disappeared, so this parking area is stored as an empty vehicle, and conversely, the vehicle has moved to the left indicated by “← B” in the figure. Later, when the vehicle stops and the polarity reversal component disappears and disappears from the screen, it is assumed that the vehicle is parked,
The fact that this parking area is parked is stored.

Furthermore, even when a vehicle enters or exits a parking area that is shaded by a high vehicle or the like, the detection area corresponding to each parking area indicated by a dotted line in FIG. By taking longer, it is possible to determine whether the vehicle enters or exits the parking area before it becomes behind the vehicle on the near side, so that it is possible to estimate whether the parking area is vacant or empty.

By such processing, it is possible to know the presence or absence of vehicles in each parking area, and by adding the number of parked parking areas, the number of parked vehicles at that time, and the maximum number of parkable cars initially input By subtracting the number of parked vehicles from, the number of vehicles that can be parked can be detected and displayed.

FIG. 4 shows an embodiment of a processing system used for the above-mentioned parking lot management to which the vehicle detection method of the present invention is applied.

A low-pass filter 2 removes noise components from a video signal captured by a television camera 1 such as an ITV camera. The function of this low-pass filter is that noise generated by an amplifier or the like included in an imaging system generally exists almost uniformly from high frequency components to low frequency components.Therefore, by removing these high frequency components, random noise components can be reduced. At this time, in order to prevent generation of aliasing noise in the A / D converter, high-frequency components are removed within a range that satisfies the sampling theorem.

The output from which the noise has been removed is converted into digital density data having a gradation by an A / D converter 3, and a subsequent edge enhancement and gradation conversion processing section 4 applies a 4 to 8 neighborhood Laplacian filter to this density data. The contour component is emphasized by contour emphasis processing using, and the resulting contour is extracted by differential operation, and the inside of the contour is filled with a constant density value or given an offset value, so that the background and the Even in a vehicle having a small density difference, gradation conversion is performed so as to make the difference between the lightness and darkness of the vehicle and the background large so that the vehicle can be easily identified.

Note that the above-described contour enhancement and gradation conversion processing can be replaced by processing of the microprocessor 11 instead of providing the contour enhancement and gradation conversion processing unit 4 separately.

The digital density data having the gradation after the pre-processing as described above is processed in the following order.

(1) In the initial state, since no image data exists in any of the image memories 6 and 7, the memory control unit 5 outputs an image memory selection signal d to the image memory 6,
The output data a having been subjected to the above preprocessing is written into the image memory 6 through the image memory input bus b.

(2) After the writing, the memory control unit 5 immediately outputs the timer start signal f, and is set to about several tens of milliseconds to several seconds in the initial state, and a sampling timer for determining a time interval t for obtaining a difference. 8 is started.

(3) After the elapse of the time interval t, the sampling timer 8 outputs a timer up signal e to the memory control unit 5.

(4) The memory control unit 5 outputs the image memory selection signal c to the image memory 7 at the same time as receiving the timer up signal e, and outputs the output data a after a certain time interval t has elapsed.
Is written into the image memory 7 through the image memory input bus b.

(5) After the writing to the image memory 7 is completed, the memory control unit 5 outputs a difference processing signal i to the difference processing unit 9, and the difference processing unit 9 that has received this signal outputs the image memory output bus g, The contents of the image memories 6 and 7 are input from h, and the difference processing is performed, and the processing result is output to the output memory 10 through the output bus j.

Then, after the difference processing for the contents of the image memories 6 and 7 is completed, a difference processing end signal 1 is output to the microprocessor 11.

(6) When the difference processing end signal 1 is input, the microprocessor 11 performs the processing shown in the flowchart of FIG. 5 and described in detail later, and outputs the processing end signal m after the end of these processings.

(7) Upon receiving the processing end signal m, the memory control unit 5 transfers and writes the contents of the image memory 7 to the image memory 6 via the image memory bus b, and sends a timer start signal f to the sampling timer 8. Start this timer.

(8) (3) The subsequent processing is repeated.

FIG. 5 is a flowchart showing the processing of the microprocessor 11 after the difference processing end signal 1 is input, as described in (6) above, and the details of the processing procedure are as follows.

Step [1] is the respective vehicle parking area of the parking lot,
That is, for example, S ₁₁ , S ₁₂ ,... S _1n , S ₂₁ , S _22.
This is a step of clearing and initializing the number of times the detection areas corresponding to _2n have been processed to “0”, respectively, and as shown by the dotted lines in FIG. 3, the vehicle parking areas S ₁₁ , S ₁₂ ,
.. S _1n , S ₂₁ , S ₂₂ ,... By setting a value larger in the left-right direction than that of S _2n , for example, it is possible to estimate the parking and empty of a vehicle in a parking area that is a shadow area of another vehicle. Is as described above.

In the next step [2], the number of detection area processes is the number of vehicle parking areas, that is, in FIG. 3, the vehicle parking areas S ₁₁ , S ₁₂ ,.
S _1n , S ₂₁ , S ₂₂ ... It is determined whether or not the total number is larger than the total number of S _2n. If the number of times of processing is larger than the number of vehicle parking areas, one processing is completed. The process proceeds to step [11] and subsequent steps. If the number is equal to or less than the number of vehicle parking areas, the processing of step [3] is performed to perform processing of a detection area corresponding to the next vehicle parking area.

In the polarity determination process in step [3], the presence or absence of the positive / negative and negative / positive polarity reversal components in the detection area is detected by raster-scanning this area. To facilitate this raster scan, It is desirable that the range of the detection area be rectangular.

At this time, the number of scanning lines in which the polarity inversion is detected in this detection area is counted, and it is determined whether or not the vehicle is a vehicle by the counted value exceeding a threshold value set to, for example, 10. If it is determined that the flag is "1", the flag memory indicating the polarity inversion is set to "1" indicating the presence, otherwise the flag "0" indicating the absence is set.

In the next step [4], the flag is checked. If the flag is "1" indicating that the polarity is reversed, it is determined that the vehicle is moving, and the moving direction of the vehicle is stored in step [5]. That is, taking the detection region of the S ₁₁ of FIG. 3 shown as an example, a "+1" as when the polarity change of the density value changes from negative to positive is moving vehicle in approach direction, positive conversely When it changes from "-" to "negative", it is determined that it is moving in the exit direction, and "-1" is stored as the moving direction flag.

One sensing region if this storage is finished, for example, from so that the processing of S ₁₁ is completed, the next detection area back the detection area processing number in step [10] after +1 to step [2] , for example, performs processing for S _12.

If the flag is “0” indicating no polarity reversal in step [4], the state of the movement direction flag stored in step [5] during the previous process is checked in step [6]. The flag indicates that the vehicle is moving. If it was “± 1”, it was determined whether the previous flag set was due to the entry of the vehicle and then parked, or whether the previous flag set was due to the exit of the vehicle and became empty thereafter. Can be determined based on the moving direction of the vehicle when the flag is set.

If the moving direction flag at the time of the previous processing is also “0”, it means that there is no change in the state, so that the processing proceeds to the step [10] and the number of detection areas processed is “1”.
And returns to step [2].

In the next step [7], the movement direction flag is set to “+”.
Alternatively, the moving direction of the vehicle at the time of the previous flag setting is checked by the sign of "-", and if the moving direction is the approach direction, it is determined that the vehicle is parked, and the parked vehicle memory is stored in step [8]. If "1" is set, and if the vehicle is moving in the exit direction, it is determined that the vehicle has exited, and the parked vehicle memory is reset to "0" in step [9]. 10], the number of times of detection area processing is set to “+1”, and the process returns to step [2].

In this step [2], the number of times of detection area processing is the number of detection areas, that is, the vehicle parking areas S ₁₁ , S ₁₂ ,... S _1n , S ₂₁ , S ₂₂ .
When it is determined that the total number has become larger than the total number of S _2n, the processing for each detection area after step [3] is completed, and the process proceeds to the comprehensive management processing following step [11]. .

In step [11], the number of parked vehicles is calculated by adding the number of parked vehicle memories set to "1" in step [8] or the number stored in all parked vehicle memories.

Further, in step [12], the total number of parked vehicles is divided by the maximum number of parked parking lots to calculate the occupancy of the vehicles in this parking lot. After that, in step [13], FIG. Output the computation processing end signal m to the memory control unit 5 of FIG.

According to the parking lot management system shown and described in FIGS. 3 to 5, the presence or absence of a vehicle is updated for each vehicle parking area when the vehicle enters and exits. It is possible to prevent the number error from accumulating as compared with the conventional method of grasping the number,
There is an advantage that it is possible to determine whether a parked or empty vehicle is shadowed by a high vehicle.

Figure 6 through Figure 7 are those showing an embodiment of a traffic management system at intersections of roads to which the vehicle detection method according to the invention, shown as A ₁ to A ₈ in FIG. 6 (a) In areas where traffic flow is obstructed by parked vehicles, etc.
A1 is to monitor illegal parking in the area of ₁ , captures the movement of the vehicle and its movement direction by the positive and negative changes in the image density value, negative and positive, using the same method as in the traffic management system, Further, by tracking the change, it is determined whether the vehicle stopped in this traffic flow obstruction area is illegal parking or the stoppage of the traffic signal is legal based on the stop time. It is.

Figure 6 (b), the region, such as to interfere with the traffic flow the images taken by the television camera C in FIG provided (hereinafter, the interference region, hereinafter) for example, the position that can monitor the A ₁ (a) indicates the monitored screen, the portion surrounded by a broken line corresponding to the interference area a _1. FIG. 9C shows a difference image obtained by performing a difference process at a constant time interval t every several tens of milliseconds to several seconds by the same processing as that in the traffic management system.

FIG. 7 shows an example of this traffic management system. In addition to adding and modifying the function of the microprocessor 11 in the parking lot management system shown and described in FIG. A transmission / reception device 13 for transmitting control information to a parking / stopping timer 15, an external interface 12, and a traffic management system 14 of a central office, which are added to the configuration, The configuration and operation of the elements are the same as those of the system shown in FIG. 4, so that detailed description will be omitted.

FIG. 8 shows a flowchart of the processing of the microprocessor 11 after the difference processing end signal 1 as described above is inputted in the parking lot management system of FIG. 4, and the processing procedure will be described below. .

In the polarity determination processing in step [21], the presence or absence of positive / negative and negative / positive polarity reversal components in the disturbance area is detected by raster-scanning this area.
At this time, in order to accurately detect the vehicle in the obstruction area, the detection area is made wider than the obstruction area, and the range of the obstruction area is made rectangular to facilitate raster scanning. It is desirable to consider the number of scanning lines in which polarity reversal is detected within and determine whether or not the vehicle is a vehicle based on a count value exceeding a threshold value set to, for example, 10. The flowchart of FIG. 5 is as described above.

As a result, when it is determined that the polarity is reversed by the vehicle, a flag “1” indicating presence is stored in the flag memory indicating the polarity reversal,
Otherwise, a flag "0" indicating no case is set.

In the next step [22], the polarity reversal flag is checked. If the flag is "1" indicating that the polarity is reversed, it is determined that the vehicle is moving, and in step [23], the moving direction of the vehicle is changed. Remember. In other words, when it is detected that the vehicle has entered the obstruction area due to the change in the polarity of the density value, “+1” is set. When it is detected that the vehicle has moved out of the obstruction area, “−1” is set. Store as a flag.

When this storage is completed, in step [24], a reset signal n is sent to the parking / stopping timer 15 in FIG. 7 to set the time value of this timer to "0". A process end signal m indicating that the process has been completed is output to the memory control unit 5.

Upon receiving this processing end signal, the memory control unit 5 performs difference processing on the next set of images. When this processing ends, a difference processing end signal 1 is sent from the difference processing unit 9 to the microprocessor 11.

When the polarity reversal flag is checked in step [22], if the flag is "0" indicating no polarity reversal, the process proceeds to step [26], and is stored in step [23] in the previous process. Check the status of the set moving direction flag. If the moving direction flag is “± 1” indicating that the vehicle is moving, it is determined in step [26] whether the previous flag set was due to the approach of the vehicle and was subsequently parked. A check is made by the flag, and if not, it is similarly checked in step [27] whether the previous flag was set due to the exit of the vehicle and became empty.

If it is determined in step [26] that the moving direction of the vehicle when the previous flag was set was the approach direction, it is checked in step [28] whether the parking / stop timer 15 has already been started. If it has not been started, in step [29] the parking / stop timer start signal n is
Send to 15 to start this timer.

If it is determined in step [28] that this timer 15 has already been started, then in step [30] this timer 15
As a result of checking the timer up signal o of step [31], if the timer is not up, it is determined that the time during which the vehicle is stopped in the obstruction area is not sufficient for illegal parking, and the process proceeds to step [25]. Is sent to the memory control unit 5 as shown in the above. If the time-up signal o is present, it is determined that the vehicle is illegally parked in step [32], and the vehicle illegal parking signal q is generated in step [33]. It is sent to the external interface unit 12 in FIG.

If it is determined in steps [26] and [27] that the vehicle was exiting during the previous process, it is determined in step [34] that the vehicle has exited from the obstruction area after stopping.
In step [35], the parking / stop timer reset signal p is output, and if the vehicle illegal parking signal q has been transmitted, this is reset, and in step [25], the processing end signal m
To the memory control unit 5 to return to the initial state.

If it is determined in steps [26] and [27] that the vehicle has not entered or exited during the previous processing, the processing is performed on the assumption that no vehicle exists in this obstruction area or that the parking state has not changed. Finished and said step [25]
To return to the initial state.

In this embodiment, it is possible to not only perform a smooth running control of the traffic flow by monitoring the parked vehicles in the obstruction area at the entrance and exit of the intersection, but also to obtain the obstruction obtained by the above system from the external interface unit 12. The parking information in the area, and in addition to this parking information, the image information at that time from the television camera 1 is transmitted to the transmission device 13.
It can be transmitted to the central office 14 via the central office, and can be applied to automatic remote management of illegal parking by being combined with the traffic management system of the central office.

FIGS. 9 to 11 show an example of a traffic management system on a road to which the vehicle detection method according to the present invention is applied. FIG. 9 (a) shows a vehicle monitoring area A in which vehicle congestion management is performed by this system. And a television camera C provided at a position for monitoring the vehicle monitoring area.

FIG. 9 (b) shows a difference image in which the monitoring area A is photographed at a fixed time interval of about several tens of milliseconds to several seconds, a difference process is performed, a background component disappears, and only a moving vehicle is extracted. FIG. 3C shows the density value distribution on the scanning line L in FIG.

In FIG. 9 (c), the widths of the polarity inversion change components of the respective vehicles are substantially equal, and x ₁ ≒ x ₂ , x ₃ ≒ x ₄ , x ₅ ≒ x ₆
Further, the relationship between the moving speeds v ₁ , v ₂ , v _{3 of} each vehicle and the width of the polarity inversion change component can be expressed by the following equation.

v ₁ = x ₁ / t ≒ x ₂ / tv ₂ = x ₃ / t ≒ x ₄ / tv ₃ = x ₅ / t ≒ x ₆ / t where t is a predetermined number 10 as described above. known in the time interval for obtaining a difference of a few seconds milliseconds, of _{x 1, x 2, x 3} , x 4, x 5, each of the difference between the width concentration values in the difference image x ₆ 0 Since it can be known by detecting the changing coordinate point, the speed of each moving vehicle can be detected. At this time, it is needless to say that it is necessary to correct the difference width x depending on the installation position, the angle, and the like of the camera.

The speed of the moving vehicle in the monitoring area is compared with the legal speed, and if the speed is extremely low or if there is no movement for a certain period of time such as 10 minutes or more, it can be determined that the vehicle is congested. . Further, during traffic congestion, the state in which the inter-vehicle distance is shorter than that of normal vehicle running is maintained, and therefore, the normal distances D ₁ and D _{2 of the} same polarity difference shown in FIG. By detecting a state that is significantly shorter than the interval of the difference in the case of the inter-vehicle distance when the vehicle is traveling, the accuracy of congestion determination can be improved.

FIG. 10 shows an example of a traffic monitoring system to which the vehicle detection system according to the present invention is applied. The function of the microprocessor 11 in the parking lot management system shown and described in FIG. ,
And a transmission device 13 for transmitting traffic control information to the traffic management system 14 of the central office. The configuration and operation of other components are the same as those of the system of FIG. Omitted.

FIG. 11 is a flowchart showing the processing of the microprocessor 11 after the difference processing end signal 1 is input from the difference processing section 9, and the processing procedure will be described below.

In the polarity determination processing in step [41], the presence or absence of the positive-to-negative and negative-to-positive polarity inversion components in the monitoring area is detected by raster-scanning this area, and the vehicle in the obstruction area is detected. The detection area must be wider than the interference area to accurately detect the area, the area of the interference area must be rectangular to facilitate raster scanning, and polarity reversal has been detected in the interference area. It is desirable to take into account the fact that the number of scan lines is counted and that the count value exceeds a threshold value set to, for example, 10 to determine whether or not the vehicle is a vehicle.
This is as described above with reference to the flowchart of FIG.

As a result of the above processing, when it is determined that the polarity is inverted by the vehicle, a flag “1” indicating presence is set in a flag memory indicating polarity inversion, and otherwise a flag “0” indicating absence is set.

In the next step [42], the polarity reversal flag is checked. If the flag is "0" indicating no polarity reversal, the flow directly proceeds to step [51] to send the processing end signal m to the memory control unit 5. When the flag is “1” indicating that the polarity is reversed, it is determined that the vehicle is moving within the monitoring area, and the next step [4] is performed.
In the moving direction monitoring process in 3], it is determined whether the vehicle is moving in the same direction as the direction in which the traffic flow should move. The polarity reversal component changes from positive to negative as viewed from the raster scan direction, or vice versa. Monitor by doing. In the example of FIG. 9A, when the raster scan direction is from left to right,
When moving from positive to negative, it will match the direction of traffic flow.

In step [44], it is determined whether or not the traffic flow direction and the vehicle detection method match based on the above processing result. If these directions match, the next vehicle position detection processing step [45] Detects the coordinate position and the inter-vehicle distance of the moving vehicle. The vehicle position and the inter-vehicle distance are obtained from the vehicle positions x ₁ , x ₂ ,... X ₆ shown as positive and negative change points with respect to the density value in FIG. 9 (c) and the inter-vehicle distances D ₁ and D _2. Can be.

In the next vehicle speed calculation processing step [46], the above vehicle position _{x 1, x 2, ...... x} 6 and the vehicle speed v ₁ of the sampling time interval t by performing an arithmetic operation of substituting the expression of the, v ₂ , v ₃ , and in step [47], the obtained vehicle speed v is compared and determined with the speed prescribed by law. In the next step [48], the inter-vehicle distances D ₁ , D ₂ are calculated. It is compared and determined whether or not the inter-vehicle distance during normal traveling is smaller than a predetermined inter-vehicle distance.

From the processing result of step [47], if the moving speed of the vehicle is significantly lower than the legal speed, and if the inter-vehicle distances D ₁ and D ₂ are shorter than the inter-vehicle distance during normal traveling, Determines "vehicle congestion" in step [49], outputs a congestion signal indicated by n in FIG. 10 to the external interface 12 in step [50], and determines whether the moving speed of the vehicle is lower than the legal speed. Isn't particularly slow or the distance between cars
If D ₁ and D ₂ are not shorter than the inter-vehicle distance during normal running, the state of “vehicle congestion” is canceled in step [52], and the congestion signal n is reset in step [53].

When the processing in step [50] or [53] is completed, a processing end signal m is sent to the memory control unit 5 in step [51] to perform the next difference processing.

The external interface unit 12 and the transmission device 13
The traffic congestion status of the road and the image information at that time can be transmitted to the central office 14 through this, and in combination with the traffic congestion monitoring system to which the present invention is applied and the traffic management system of the central office, remote monitoring of traffic flow can be performed. Can be used for control.

(Effect)

In the vehicle detection method of the present invention, a difference in a time interval as short as about 10 milliseconds to several seconds is obtained, and processing is performed using this difference. Not affected.

Therefore, there is no need to control the threshold value and the reference value caused by the luminance change, which is the biggest problem in the prior art, and a special effect that stable processing and determination can be achieved is achieved.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining the principle of the present invention, FIG. 2 is a diagram showing an example of a parking lot in which a parking lot management system to which the present invention is applied is installed, and FIG. 3 is a screen monitoring this parking lot. FIG. 4, FIG. 4 is a block diagram showing an embodiment of a parking lot management system to which the present invention is applied, FIG. 5 is a flowchart showing the processing of the microprocessor, and FIG. 6 is an example in which the present invention is applied to traffic management. FIG. 7 is a block diagram showing an embodiment of a traffic management system to which the present invention is applied. FIG. 8 is a flowchart showing the processing of the microprocessor. FIG. 9 is an example in which the present invention is applied to traffic monitoring. FIG. 10 is a block diagram showing an embodiment of a traffic monitoring system to which the present invention is applied, and FIG. 11 is a flowchart showing processing of the microprocessor.

Continuation of front page (56) References JP-A-63-40862 (JP, A) JP-A-63-174198 (JP, A) JP-A-62-220884 (JP, A) JP-A-60-69786 (JP) , A) Japanese Utility Model Showa 50-108237 (JP, U) JP-B 43-24875 (JP, B1) JP-B 41-635 (JP, B1)

Claims (8)

(57) [Claims] 1. An image photographed at regular time intervals of about several tens of milliseconds to several seconds is converted into digital density data having gradations, and a difference between density data at the same position of the two preceding and succeeding images is obtained. Vehicle detection method characterized by extracting a vehicle according to the positive and negative polarities of the difference 2. The vehicle detecting method according to claim 1, wherein the moving direction of the vehicle is determined based on the positive and negative polarities of the difference. 3. The accuracy of determination is enhanced by enhancing a contour portion of the vehicle in the photographed image and performing tone conversion inside the contour to increase a density difference from a background. Item 4. The vehicle detection method according to Item 1. 4. The vehicle detecting method according to claim 1, wherein the accuracy of the determination is enhanced by adding to the determination condition that the widths of the positive and negative polarities of the difference are equal. 5. A photographing means for photographing a predetermined area at fixed time intervals of several tens of milliseconds to several seconds, and an A / D converter for converting an image signal photographed by the photographing means into digital density data having a gradation. And a difference calculation of the digital density data at the same position of the two images taken before and after the fixed time interval. A vehicle is extracted based on the positive and negative polarities in the resulting difference image, and the moving direction is determined. A vehicle detection device, comprising: a microprocessor for determining. 6. The vehicle detecting device according to claim 5, wherein the predetermined area is a parking lot, and a function of judging vehicle entrance / exit, judging parking / vacancy, and calculating the number of vehicles is added to the microprocessor. 7. The vehicle detection device according to claim 5, wherein the predetermined area is on a road near an intersection, and a function of determining whether the vehicle is parked or stopped is added to the microprocessor. 8. A vehicle detecting apparatus according to claim 5, wherein said predetermined area is a part on a road, and said microprocessor is provided with a function of judging a speed and an inter-vehicle distance of a vehicle passing through said area. .