JPH02122400A - Space density detection system - Google Patents

Abstract

PURPOSE:To grasp a traffic flow with automatically detecting space density by detecting the travelling speed of a vehicle from the location of the vehicle to be detected through a video signal and from the location of the vehicle to be detected through another video signal after the interval of a previously set time, detecting the number of vehicles in a measuring area with making correspondence the location of the vehicles, and operating it. CONSTITUTION:A video obtained at a video input part 1 takes the difference of the adjoining picture element value on a same scanning line in a characteristic extraction part 2 and binarizes the value corresponding to the respective picture elements. The vehicle speed is obtained in a vehicle speed detection part 4 with dividing by a detection time interval a distance between the location of the vehicle to be detected from a previous cycle and the location of the vehicle present in the vicinity of a location which is forecasted with adding a value multiplying the detection time interval with the travelling speed of the vehicle to the location of the vehicle to be detected from the previous cycle. A second storage part 7 successively receives the presence number of a detected one-frame from a presence number detection part 6, accumulates it, a space density detection part 9 divides the accumulated value of the presence number by the number of sample frame in a unit time, and obtains it as the space density. Thus, the space density can automatically be detected.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a method for detecting the spatial density of vehicles, which is useful information in road traffic control.

(Prior art and its issues) Detecting traffic conditions is very effective in maintaining smooth road traffic, and spatial density indicates how many vehicles are present per unit length. is important. Using this result, traffic controllers can grasp the road congestion situation and
Traffic information can be sent out or reflected in traffic signal control.

As devices for detecting traffic conditions, devices called ultrasonic vehicle detectors and loop vehicle detectors have been used. These devices apply ultrasonic waves or magnetism to a point on road-E, and measure the presence or absence of a vehicle based on changes in the waves. The number and speed of vehicles are detected based on the time of change.

However, in principle, these devices can only determine the traffic condition at one point, so they cannot measure traffic conditions over a fairly wide range, such as traffic density.

Even if there is a result of dividing the number of cars by the speed, 1
Although we have inferred the whole thing from the results of the points, it cannot be said that it is correct in terms of content.

On the other hand, a method for measuring traffic conditions by processing images obtained from television cameras was proposed in the same patent application 55 by Ikoshi.
184591 (vehicle presence position detection method). However, a method for detecting spatial density like this method has never been seen before.

As described above, an object of the present invention is to provide a method for accurately and automatically detecting spatial density, which is important in understanding traffic flow.

(Means for Solving the Problems) The spatial density detection method of the present invention photographs a road with a television camera and processes the video signal obtained from the television camera to distinguish between road sections and overlapping sections. Based on this distinction, the location of the vehicle traveling on the road is detected, and -
The running speed of the vehicle is detected from the position of the vehicle detected from the iron blood signal and the position of the vehicle detected from another video signal after a preset time interval, and the speed of the vehicle is detected from the position of the vehicle detected from the previous cycle. The current position of the vehicle is correlated with the current position of the vehicle based on the traveling speed of the vehicle, the number of vehicles within the measurement area is detected, and video frames sampled at regular intervals are used. It is characterized in that the cumulative value of the number of existing vehicles per unit time is divided by the number of video images 18 used for processing per unit time, and the obtained value is determined as the spatial density of vehicles.

(Example) Next, the present invention will be explained with reference to the drawings.

FIG. 1 is a front view showing one embodiment of the present invention.

In FIG. 1, a video input section 1 consists of a television camera.

The video obtained by the video input section 1 is sent to the feature extraction section 2.

'4. In the +i fine extraction unit 2, 1-
184591 (Vehicle location detection method)
As described in , by taking the difference between neighboring pixel values on the same line and taking 1" if the value of the difference is greater than a preset threshold, and taking 0° if it is smaller. The f value corresponding to the color pixel is binarized.

By summing up the number of 1's on the same search line within the measurement area,! T! fii7! get value The characteristic value is compared with the upper and lower threshold values of each scanning line obtained from the previous cycle, and if it is determined that a vehicle is present, it is 1'
, otherwise, a feature value of each scanning line (scanning line feature value) is obtained, which is set to 0°.

The upper threshold and lower threshold are determined by the following method.

That is, in this method, the total value of the number of "1's" on the same scanning line when the scanning line feature value is "0°" is calculated within a certain period of time, and the total value of the number of "1" with the highest frequency is calculated. A preset width is taken above and below with the center at the center, and the upper limit thereof is the upper threshold value, and the lower limit thereof is the lower threshold threshold value. The case where the scanning line feature value is 0' is used because the case of 1' is the state B where the vehicle is running, and the case of 0' is the state where the road surface is being photographed.For example, As the certain period of time,
You can use 1 minute or 30 seconds. Also, the vertical width can be about plus or minus 3. Generally, when the road surface is in a figurative state, the total value of the numbers of 1 on the scanning line is O, and if there are letters etc. displayed on the road surface, the total value of the numbers of 1 on the scanning line is O, and if there are letters etc. I get the value.

The vehicle detection unit 3 determines whether the signal is caused by a vehicle or noise from the longitudinal arrangement of the scanning line feature values obtained by the feature extraction unit 2. If the signal is caused by a vehicle, it is determined whether the signal is caused by a vehicle or not. The frontmost scanning line number and the last scanning line number are determined as the front and rear of the vehicle location. A method for detecting the front and rear portions of a vehicle from scanning line feature values is described in Japanese Patent Application No. 55-184591 (Vehicle Presence Position Detection Method) filed by the same applicant.

The vehicle speed detection unit 4 adds a value obtained by multiplying the traveling speed of the vehicle detected in the previous cycle by the detection time interval to the existing position of the vehicle detected in the previous cycle, and detects a vehicle existing in the vicinity of the predicted position. The vehicle speed is determined by dividing the distance between the presence position of the vehicle (1'l position) and the position of the vehicle detected in the previous cycle by the detection time interval. , 1st
It is stored in the storage unit 5.

Here, since the television camera that collects images of the road images the road from diagonally behind, vehicles in the distance may be hidden by vehicles in the foreground. In other words, since the position of the front of the vehicle determined by the vehicle detection unit 3 is affected by the height of the vehicle, it is detected as a position close to the front of the vehicle. Therefore, the front and/or rear of the vehicle may not exist in the device composed of the position and velocity of the previous cycle, hidden in the vehicle of the song.

In that case, it is determined whether or not the front and rear parts of the vehicle to be predicted are sandwiched between the front and rear parts of the main mountain of the pond, and the front and rear positions of the vehicle are estimated. Specifically, there are the following three cases.

■Is the rear of the vehicle only for vehicles coming from behind (if it is hidden by the front of →)?

The apparent front and rear positions of the previous cycle are Pf-Pb, respectively.
- cylinder circumference jlJ] If the detection speed is VO - the detection time interval and the apparent front position of the rear vehicle is Pl, then Pf - t - VO * t > PL > Pb + - VO * j, a) According to the description of b, a indicates the front of b (if the photograph is taken from the rear of the vehicle, it is far away).

■If the vehicle overlaps the front of the vehicle or the vehicle in front.

If the rear position of the vehicle in front is P2, then Pf-+-VOOL>P2> Pb+VO*t■■
If the image appears to overlap with the front and back of the vehicle, or with the vehicles in front and behind.

Pf-1-VO*t>P2>Pl,> Pb-+-
In these cases, for example, the position of the rear of the vehicle (
The place where speed is judged using H is ■ and ■
Otherwise, it becomes impossible to judge the speed. At this time, the speed of the nearby vehicle, that is, the vehicle that is affected by the shadow (for example, in case 6 of the above example, the speed of the rear vehicle, in case of ■, the speed of the vehicle in front,
In this case, use the average value of the speeds of the vehicles in front and behind. Also, the location of the hidden vehicle 1d is written in J! 1]
C) The γ position is substituted for the front position of the vehicle and the rear position of the vehicle. The front and rear positions of the vehicle obtained by the vehicle detection unit 3, the vehicle speed and prediction (11) obtained by the vehicle speed detection unit 4 are as follows:
It is stored in the first storage unit 5 and used for measuring the next cycle.

The existing vehicle number detection section 6 receives the front position and rear position of the vehicle obtained by the vehicle detection section 3 and the vehicle speed detection section t1, and counts the number of existing vehicles within the R1 measurement area. The vehicle is
It is moving upwards from the bottom of the TV camera footage.

(The image is taken from the rear of the vehicle), and the measurement area can be taken at the bottom and top of the TV camera image along the lane.

The number of vehicles existing in the corresponding frame can be counted by the number of combinations of the front and rear parts of the vehicle. If either the front or rear of the vehicle is outside the measurement area, consider the distance between the front and rear of the vehicle as the vehicle length, and calculate the ratio of the length of the vehicle that exists within both measurement areas to the vehicle length. , you can think of it as the number of machines in existence. For example, if 3 m of 5 m vehicles exist in the measurement area, the number of existing vehicles can be considered to be 06 vehicles. Also, the number of rear parts of the vehicle can be easily set to the number of existing vehicles.

The second storage unit 7 sequentially receives the number of machines present in one undetected frame from the number-of-present machine detection unit 6 and accumulates them. When it receives the 1ii time signal from the timer 8, it sends the cumulative Vt value to the spatial density detector 9 and sets the content to ○.

In Timer 8! It is running from 1 to 30 minutes after opening.

When performing spatial density detection every minute, every minute vehicle position,
In the two-space density detection section 9 which sends out the unit time signal to the second part 7, the fact is detected within the unit time.

The cumulative value V5 of the number of existing vehicles detected in each frame is
11 is received from boundary section 7 and divided by V 5 /'F by the number F of sample frames within the time, and the result is obtained as the spatial density. For example, if all frames of video are processed in one minute, the general NTS C
For the video fffj of the method, 1800 frames of video ('(i
is being processed, so F=1800.

]The average number of devices present per frame is 3, or V5=5.
400, 5400. '1800 will be calculated.

In this way, according to this embodiment, the spatial density, which is important in understanding traffic conditions, is automatically detected.

<Effects of the Invention> As explained above, the spatial density detection method of the present invention can automatically detect spatial density, which is important for understanding traffic flow, by using images from a television camera. The present invention has such effects.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing one embodiment of the present invention. 1... Image input section, 2... #Jin'i output section, 3
...Vehicle detection unit, 1...Vehicle speed and reaction detection unit, 5...
- First part - Membrane section, 6. Existence number detection section, 7. Second storage section, 8. Timer, 9. Spatial density detection section.

Claims (1)

[Claims] The road is photographed with a television camera, and the video signal obtained from the television camera is processed to distinguish between the road portion and the vehicle portion, and from this distinction, the location of the vehicle traveling on the road is detected, The running speed of the vehicle is detected from the position of the vehicle detected from one video signal and the position of the vehicle detected from another video signal after a preset time interval, and the speed of the vehicle is detected from the position of the vehicle detected from the previous cycle. The number of vehicles in the measurement area is detected by associating the current location of the vehicle with the vehicle location obtained from the vehicle's running speed, and is obtained using video frames sampled at regular intervals. A spatial density detection method characterized in that the cumulative value of the number of existing vehicles per unit time is divided by the number of video signals used for processing per unit time, and the obtained value is determined as the spatial density of vehicles.