JP2000149181A - Traffic stream measurement system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To quantify existence confirmation by comparing the finite difference between two preceding and subsequent changed luminance histograms with a preliminarily set threshold and discriminating the existence of a vehicle. SOLUTION: A vehicle detecting device 2 is provided with a luminance histogram preparing means which prepares a luminance histogram. When acquired data such as standard deviation value does not change for a fixed time, the average luminance, standard deviation and a luminance histogram for the time are road surface information and are recorded and stored as road surface data in a continuous image input device 1. Average luminance, luminance dispersion value and standard deviation are calculated based on the luminance histogram of a designated and set area, and the difference between an input image histogram and a road surface image histogram is calculated with calculation about the average luminance. The existence of a vehicle is confirmed by comparing four thresholds being specific thresholds 1 and 2, the dispersion threshold of road surface data and a threshold when a shadow takes place with a real-time image histogram.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a traffic flow measurement system, and more particularly to a method of calculating the passage and type of a vehicle by using an image processing technique for processing an image obtained by photographing an automobile passing a fixed point. The present invention relates to a traffic flow measurement system.

[0002]

2. Description of the Related Art There is a need to automatically determine the state of flow of vehicles traveling on an expressway. A traffic flow measurement system that responds to such a request is known. The image processing technology used by the known traffic flow measurement system includes:
It is a technology based on the fact that there are many edges on the front of the vehicle,
When the primary data obtained by vertically projecting and adding the designated area of the image subjected to the spatial differentiation filter processing is larger than a certain threshold value, it is determined that a vehicle exists in that area.

[0003] Known techniques of such a judgment format include those shown in FIG.
As shown in (a), (b) and (c), the large vehicle 6
When the rear part of the vehicle No. 01 passes through the designated area 604, the rear part is photographed as an edge 607 and appears.
There is a problem that the edge is detected as the edge of the following vehicle.

[0004] Vehicle detection is also performed for vehicle type determination. As a vehicle type determination method based on the vehicle length of a known traffic flow measurement system using image processing, when a vehicle 701 is detected in an area 702 set as shown in FIG. Is a method of determining whether the vehicle is large or small by determining whether or not the vehicle exists. Such a determination method only determines whether the vehicle is large or small, and cannot quantitatively determine the vehicle type. Furthermore, the occupancy rate, which is one of the characteristics of the road traffic flow and is calculated using the vehicle length, is not accurate because the vehicle length of each vehicle is not measured.

[0005] By measuring the length of a passing vehicle, it is required that the vehicle type can be quantitatively determined based on the length of the vehicle and the occupancy can be measured accurately.

[0006]

SUMMARY OF THE INVENTION The present invention has been made based on such a technical background, and achieves the following objects. An object of the present invention is to provide a traffic flow measurement system capable of quantifying existence confirmation. It is another object of the present invention to provide a traffic flow measurement system that can quantitatively determine the presence or absence of a vehicle with high accuracy. Still another object of the present invention is to provide a traffic flow measurement system capable of quantitatively calculating the speed of a vehicle whose existence has been confirmed. Still another object of the present invention is to provide a traffic flow measurement system capable of quantitatively calculating the length of a vehicle whose presence has been confirmed based on the measured speed. Still another object of the present invention is to provide
An object of the present invention is to provide a traffic flow measurement system that can accurately measure an occupancy of a vehicle whose existence has been confirmed based on a measurement speed. Still another object of the present invention is to provide a traffic flow measurement system capable of confirming the presence of a vehicle having a low luminance contrast with respect to a road surface by calculation with high accuracy.

[0007]

SUMMARY OF THE INVENTION A traffic flow measuring system according to the present invention comprises: a camera for obtaining a continuous image by continuously photographing a road surface and a vehicle passing through a set area on the road surface; A luminance histogram calculating means for calculating a luminance histogram by calculation, a difference calculating means for calculating a time-series difference of the luminance histogram, and the luminance histogram having no change during a set time as road surface data. Storage means, and difference calculation means for calculating a difference between two preceding and succeeding brightness histograms when the brightness histogram changes during the set time, and comparing the difference with a preset threshold value And vehicle detection means for determining the presence of the vehicle.

In the traffic flow measuring system according to the present invention, it is preferable that the vehicle detecting device includes a luminance histogram calculating means, a difference calculating means, and a road surface data storing means. The vehicle speed calculation means for calculating the vehicle speed can include a luminance histogram calculation means, a difference calculation means, a road surface data storage means, a vehicle existence time calculation means, and a conventional speed detection means.

Since the traffic flow measuring system according to the present invention detects the presence of a vehicle based on the difference between a road surface histogram recorded and stored and a real-time histogram, the presence / absence of the vehicle can be quantified with high accuracy. Then, since the continuous pixels of such a histogram are obtained in a time sequence, the vehicle length and the vehicle type can be determined by reusing them as pixel time-series information. Preferably, the luminance histogram includes a luminance average value and a luminance standard deviation. By using the luminance average value and the luminance standard deviation as judgment data, the numerical value of the vehicle presence can be further improved.

The traffic flow measuring system according to the present invention further comprises a speed measuring means for measuring the speed of the passing vehicle whose existence is confirmed by the histogram, and a vehicle length for calculating the vehicle length based on the existence and the speed. More preferably, calculation means are provided. Under the existence confirmation, vehicle information can be accurately obtained. It is also possible to determine the vehicle type based on the vehicle length and calculate the number of vehicles for each vehicle type based on the vehicle type and presence. Further, the headway can be calculated based on its presence and its speed. A warning can also be displayed based on the distance between the vehicles. By calculating the time occupancy based on the presence and the speed, the traffic flow can be accurately determined.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows, for each type of vehicle, the number of vehicles passing through a certain period of time, the average speed of the vehicle, and the occupancy, and the distance between the vehicles in front of the vehicle. The embodiment of the traffic flow measurement system according to the present invention, which can detect the passage of a dangerous vehicle on a warning display device when a short vehicle is detected, is shown.

A continuous image taken by an ITV camera installed at a fixed point is read into the continuous image input device 1. The continuous image is created by photographing a constant screen per fixed time. The vehicle information read by the continuous image input device 1 is transmitted to and input to the vehicle detection device 2 and the vehicle speed measurement device 3.

The vehicle detecting device 2 determines the presence or absence of a vehicle having various physical characteristics based on the vehicle information read by the continuous image input device 1. The vehicle length measurement device 4 receives the detection data regarding the vehicle passing at the time t obtained by the vehicle detection device 2 and the vehicle speed measured by the vehicle speed measurement device 3, and based on the detection data and the vehicle speed. , The length of the passing vehicle can be calculated.

The vehicle type determining device 5 can determine the vehicle type of the passing vehicle by comparing the vehicle length measured by the vehicle length measuring device 4 with a known vehicle length that has already been input. The vehicle number measuring device 6 can receive the vehicle type determination information determined for each passing vehicle by the vehicle type determining device 5 and count the vehicle number for each vehicle type.

The vehicle number measuring device 6 increments the number of vehicles for each vehicle type determined by the vehicle type determining device 5. The inter-vehicle distance counting device 7 receives information such as the passing time and the vehicle length of the vehicle that has passed through each lane from the vehicle number measuring device 6, and can calculate the inter-vehicle distance based on the information.
The warning display device 7 receives the inter-vehicle distance information of the inter-vehicle distance counting device 7 and displays a warning when the inter-vehicle distance is short. The occupancy measuring device 9 can calculate the time occupancy from the number of times of vehicle detection within a certain time obtained by the vehicle detecting device 2. Each device is described in further detail below.

Vehicle detecting device 2: As shown in FIG.
Detection areas 201, 202, 203, and 204 are set for each of the lanes 205, 206, 207, and 208. It is preferable that the detection areas 201, 202, 203, and 204 be set in position areas where front and rear vehicles do not overlap as much as possible. The vehicle detection device 2 includes a luminance histogram creating unit that creates a luminance histogram.

The brightness histogram creating means has a calculation function for graphing the relationship between the brightness of a plurality of set detection areas and the number of pixels corresponding to the brightness.
The total number of pixels per unit time is n. From the luminance histogram, the standard deviation of the luminance is calculated by the following equation.

S = {(1 / n) Σ (xi−x) 2｝ 1/2, x = (1 / n) Σxi. (1) where Σ is the subscript i It represents adding from 1 to n. The standard deviation s is a calculated value that quantitatively indicates how wide the luminance histogram is distributed from the luminance average value.

If the vehicle does not pass through the set area during the unit time, a road surface luminance distribution 410 having a single peak value (the number of pixels N) is obtained as shown in FIG. That is, reflected light reflected only from the road surface is detected. Even in the case where a moving cloud suddenly blocks sunlight, by shortening the photographing time of the unit continuous number of histograms, erroneous information due to the natural environment can be removed.

When a vehicle passes through a set area during a unit time, as shown in FIG. 4A, a vehicle-passing luminance distribution of a spectrum having two peak values is obtained in principle. Since the amount of light reflected from the vehicle is different from the amount of light reflected from the road surface, two peak values appear in the luminance distribution when the vehicle passes. The reason why the luminance histogram has a certain distribution width (dispersion value, standard deviation value) is mainly due to the change in the running direction of the vehicle body shape, which depends on the presence of moving objects such as clouds, vehicles, natural conditions such as rainfall and fog. The amount of light reflected from the fluctuating surrounding environment and reaching the set area is fluctuated.

If the acquired data such as the standard deviation (value) does not change during a certain time Δt, the average luminance, standard deviation, and luminance histogram during that time are road surface information. Record and save in the input device 1. The brightness histogram of the designated / set area is immediately created (almost in real time) by calculation in step S301 shown in FIG.

Hereinafter, the "histogram" is limited to that of the set area. Based on the histogram, immediately in step S302, the standard deviation s of Expression (1), which is the average luminance and the luminance variance (value), is calculated. Subsequently, with respect to the average luminance, in step S303, a difference between the input image histogram (an image irrelevant to the presence or absence of vehicle passage and which may coincide with the road surface image) and the road surface image histogram is calculated. Average brightness difference = Average brightness of input image−Average brightness of road surface image (2) Subsequently, for the entire brightness histogram, step S
An overall difference is determined by 304. Overall difference = input image luminance histogram−road image luminance histogram (3)

This overall difference is shown in FIG. The luminance spectrum 409 of the vehicle portion is a difference as it is. Since the number of pixels indicating light reflected from the road surface due to the passage of the vehicle decreases, the vertical axis amount of the histogram of the road surface portion is naturally negative as a difference as shown in FIG. The area difference 411 of the vertical axis amount of the vehicle portion is the sum of the positive components. When a white vehicle body that reflects light better than the road surface passes, the area difference 411 is substantially equal to that of FIG.
This corresponds to the vehicle portion 409 in FIG.

The vehicle portion 409 of the input image is distributed high (shifted to the right in the figure) in a white vehicle, and low (shifted to the left) in a black vehicle. When a passing vehicle is a small vehicle, a person makes a determination, and sets a histogram difference lower than the luminance distribution 411 of the small vehicle as the threshold value 1. A value slightly higher than the threshold value of the road surface data is set as a variance threshold value for variance (standard deviation).

FIG. 4D shows a luminance histogram of a vehicle having a reflective surface having low contrast with a road surface such as the upper surface of a gray truck. In this case as well, it is distributed lower due to sunlight reflected from the vehicle body, but compared to the road surface,
A high vehicle brightness distribution 412 is obtained. For such a vehicle, the histogram difference is determined by threshold value 2
Set as

The luminance histogram shown in FIG. 4 (e) is shown in the case where the shadow of the vehicle body, which occurs in relation to the position of the sun, occurs in the detection area. The luminance distribution 416 of the road surface portion where the shadow has occurred is distributed with a lower luminance (shifted to the left in the figure) than the previously saved road surface luminance distribution.
The threshold value of the average luminance difference is set to be equal to the average luminance value of 41 so as not to detect such a shadow portion in the area as a vehicle portion.
3 and a deviation difference 4 that is a difference between the deviation of the average luminance 414 of the road surface.
15 is set to a positive value. With this positive setting, the difference between the shadow and the vehicle body can be determined, and the presence of a vehicle with low contrast can be detected.

In step S304, the four threshold values of the two threshold values 1 and 2 described above, the variance threshold value of the road surface data, and the threshold value in the case where a shadow occurs are determined in real time. By comparing with the image histogram, the following judgment can be made by the conditional expression determined in step S305.

The conditional expression is represented by the following expression. Step S305-1: histogram difference (entire difference)
> = Threshold 1, and variance> = variance threshold

Or step S305-2: histogram difference >> = threshold 2, and average luminance difference >> = average luminance threshold, and variance >> = road surface variance

If such a conditional expression is satisfied, the existence of the vehicle is confirmed. By this confirmation, the vehicle detection flag at time t is turned ON. If condition 305-1 is satisfied,
The existence of a vehicle having a high contrast with respect to the road surface can be confirmed. If the condition 305-2 is satisfied, the presence of the vehicle can be confirmed by distinguishing from the shadow although the contrast with the road surface is low.

Vehicle speed measuring device 3: Vehicle speed measuring device 3
2. Description of the Related Art Known in the art is to detect the average speed of a passing vehicle by a vehicle tracking method based on template matching based on the vehicle speed of each lane passing through the length direction areas 205 to 208 in which a certain distance is set in the traveling direction shown in FIG. Speed detection method. Various other conventional methods are known as a method for detecting the speed of the passing vehicle, and any of the methods can be used in the present invention.

Vehicle length measuring device 4: FIG. 5 shows one large vehicle 501 and one small vehicle passing through one of the length regions 205 to 208 shown in FIG. 502, the presence of a vehicle is detected by passing through the detection area 202 for creating a histogram. A continuous time zone Δt in which the presence of the vehicle is continuously detected corresponds to the passing time of the vehicle.

The vertical axis in FIG. 5B indicates the presence of a vehicle.
The time zone 505 is the time 5 required for the large-sized vehicle 501 to pass the distance obtained by adding the length of the length direction area and the length of the vehicle body.
07 is shown. A time zone 506 indicates a time 509 required for the small vehicle 502 to pass a distance obtained by adding the length of the length direction area and the length of the vehicle body.

The vehicle length of the large vehicle 501 is obtained by calculation by multiplying the time 507 by the conversion speed of the large vehicle 501 measured by the vehicle speed measuring device 3 and the actual distance, and subtracting the distance in the length direction area. be able to. The conversion count is determined by actually measuring in advance according to the angle and size of the vehicle on the image.

Vehicle type determining device 5: The vehicle type and the vehicle length of the vehicle type have been input to the vehicle type determining device 5 in advance. The vehicle type is determined by comparing the reference vehicle length with the vehicle length of the passing vehicle obtained by the vehicle length measuring device 4. The vehicle type is, for example, large or small. Vehicle number measurement device 6:

By incrementing the number of vehicles for each vehicle type determined by the vehicle type determination device 5 (adding 1 in order), the total number of vehicles passing for each vehicle type can be known.

Inter-vehicle distance counting device 7: The inter-vehicle distance between the preceding vehicle and the succeeding vehicle in each lane can be calculated from the inter-vehicle time 508 in FIG. 5B. Inter-vehicle time 50
The actual inter-vehicle distance is calculated by multiplying 8 by the measurement speed and the conversion coefficient to the actual distance. When the speed is equal to or higher than a certain value and the inter-vehicle distance is equal to or less than a certain value, the passing of the vehicle is displayed on the following warning display device.

Warning display device 8: The warning display device 8 is installed forward in the traveling direction of the vehicle. If there is a warning, a warning such as caution of speed is displayed to notify the driver.

Occupancy measurement device 9: Calculates the time occupancy of vehicles in each lane within a certain time. Time 507 in FIG.
, The time when the vehicle 501 has passed the distance corresponding to the vehicle length is calculated. For all vehicles passing through, the time required to pass the distance corresponding to the vehicle length is calculated, and the total is calculated. The ratio of the total time to the measurement time is the time occupancy.

[0039]

The traffic flow measuring system according to the present invention can digitize the presence of a vehicle with high accuracy. Under the high-precision determination of the vehicle presence, the vehicle length, vehicle type, speed, inter-vehicle distance, traffic flow, and degree of congestion can be quantified one after another using the same pixel time-series information used in the presence determination. it can.

[Brief description of the drawings]

FIG. 1 is a circuit block diagram showing an embodiment of a traffic flow measurement system according to the present invention.

FIG. 2 is an oblique axis projection view showing a relationship between a lane and a detection area.

FIG. 3 is a flowchart of the traffic flow measurement system according to the present invention.

FIG. 4 (a), (b), (c), (d), (e)
Is a graph showing various histograms.

FIGS. 5A and 5B are conceptual diagrams of vehicle type determination, wherein FIG. 5A is an oblique projection view and FIG. 5B is a time chart.

6 (a), 6 (b), 6 (c) are oblique axis projection views showing conventional vehicle detection.

FIG. 7 is an oblique axis projection view showing conventional vehicle length detection.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1: Continuous image input device (camera, various calculation means) 2: Vehicle detection apparatus (various calculation means) 3: Vehicle speed measurement apparatus (various calculation means + speed calculation means) 4: Vehicle length measurement apparatus (various calculation means + vehicle) Length calculation means) 5: Vehicle type determination device (various calculation means + vehicle length comparison means) 6 ... Vehicle number measurement device (various calculation means + number calculation means) 7 ... Vehicle distance counting device (various calculation means + vehicle calculation means) 8 ... warning display device 9 ... occupancy measurement device (various calculation means + occupancy calculation means) 201 to 204 ... setting areas 408 and 409 ... input image histogram 410 ... road surface histogram (road surface data) 501 to 504 ... vehicle

Claims (8)

[Claims] 1. A camera for obtaining a continuous image by continuously photographing a road surface and a vehicle passing a set area on the road surface, and a luminance histogram calculation for calculating a luminance histogram for each of the continuous images by calculation. Means, difference calculating means for calculating a time-series difference of the luminance histogram, storage means for storing the luminance histogram that does not change during a set time as road surface data, and the luminance histogram during the set time A difference calculating means for calculating a difference between two luminance histograms before and after, and a vehicle detection for determining the presence of a vehicle by comparing the difference with a preset threshold value Traffic flow measurement system consisting of means. 2. The traffic flow measurement system according to claim 1, wherein the luminance histogram includes a luminance average value and a luminance standard deviation. 3. The vehicle according to claim 1, further comprising: speed measuring means for measuring a speed of the passing vehicle whose existence is confirmed by said histogram; and calculating a vehicle length based on said existence and said speed. Traffic flow measurement system consisting of vehicle length calculation means. 4. The vehicle according to claim 3, further comprising: vehicle type determining means for determining a vehicle type based on said vehicle length; and vehicle number calculating means for calculating the number of vehicles for each vehicle type based on said vehicle type and said presence. Traffic flow measurement system. 5. The traffic flow measuring system according to claim 3, further comprising: an inter-vehicle calculating means for calculating an inter-vehicle distance based on the presence and the speed; and a warning display means for displaying a warning based on the inter-vehicle distance. . 6. The traffic flow measurement system according to claim 3, further comprising: time occupancy calculating means for calculating a time occupancy based on the presence and the speed. 7. A traffic flow measurement system for producing a continuous image of a time series and comparing luminance histograms that are temporally different from each other in order to determine the presence or absence of a vehicle. 8. The traffic flow measurement system according to claim 7, wherein a standard deviation value of the luminance histogram is calculated.